Description
Many businesses have public relations departments dedicated to managing their reputation. In addition, many public relations firms describe their expertise in terms of reputation management. The public relations industry is growing due to the demand for companies to build corporate credibility and hence reputation.[citation needed] Incidents which damage a company's reputation for honesty or safety may cause serious damage to finances.
A Case study on corporate reputation
Abstract:The purpose of this study is to measure and compare the corporate reputation of four staffing and recruitment companies, from the perspective of graduating students. Corporate reputation is a concept that has interested researchers for the last four decades, and still the concept incorporates a number of varied definitions, contexts and measurement methods (Gotsi &Wilson, 2001). Along with definition ambiguity comes also measurement concerns such as whether to aggregate or disaggregate several perspectives into one general index of a company's reputation. Most researchers agree that corporate reputation is a suitable concept when describing the perceptions that stakeholders have of a company, and that it can only be ascribed to one company. This study takes the perspective to investigate one external stakeholder group of staffing and recruitment companies - graduating students. The group is of interest because it can possibly become an internal stakeholder group - employees. For this study a quantitative method has been applied and an investigation has been made through a survey. Methodology about measuring corporate reputation is discussed, and lay ground for the empirical data collection. An altered version of the Harris-Fombrun Reputation Quotient (RQ) is used, in order to achieve better fit of the instrument and to attain higher reliability and validity. The purpose of this study has been fulfilled by measuring corporate reputation from the perspective of graduating students. Empirical data from 125 students has been collected, and four corporate reputations of staffing and recruitment companies have been measured. The findings are not valid for generalisability but enoughto give a good estimation of the population of interest. In addition, a disaggregated analysis of the RQ components has confirmed the findings. Using the methodology of the theoretically based Harris- Fombrun Reputation Quotient further validates the findings. Graduating students perceive Academic Work in the most positive way, in respect to the other companies.
ii
Kandidatuppsats i företagsekonomi Titel: Författare: Handledare: Datum: Ämnesord: Hur avgångsstudenter uppfattar bemannings- och rekryteringsindustrin - En studie om företagets anseende Christopher Broman, Robin Cabander, Emilia Karlsson Börje Boers 200805-28 Företagets anseende, Corporate Reputation, Reputation Quotient, League table, Bemanningsföretag, Rekrytering
Sammanfattning
Syfte: Syftet med denna studie är att mäta och jämföra anseendet av fyra bemannings- och rekryteringsföretag, utifrån avgångsstudenters perspektiv. Företagets anseende är ett teoretiskt koncept som har intresserat forskare de senaste årtiondena och fortfarande råder delade meningar om definitioner, sammanhang och mätmetoder rörande konceptet (Gotsi & Wilson, 2001). Med tvetydiga definitioner kommer även oklarhet gällande mätningsmetod, som till exempel om man ska summera flera intressenters åsikter till ett övergripande index eller inte. Forskare är eniga om att företagets anseende är ett passande koncept när man ska beskriva de åsikter intressenter har om ett företag. Denna studie undersöker en specifik grupp av bemannings- och rekryteringsföretags externa intressenter, nämligen avgångsstudenter. Denna grupp är intressant då de kan komma att bli interna intressenter, via anställning av ett företag. Denna studie har en kvantitativ inriktning och undersökningen har utförts med hjälp av en enkät. Metodiken i att mäta företags anseende genom att använda sig av league tabeller och pilotstudier diskuteras och ligger till grund för empiriinsamling. Användandet av en anpassad version av Harris-Fombrun Reputation Quotient (RQ) ger studien ökad validitet och reliabilitet. Syftet med denna studie har blivit uppnått genom att mäta företagens anseende ur avgångsstudenters perspektiv. Empirisk data från 125 studenter har samlats ihop och fyra bemannings- och rekryteringsföretags företagsanseenden har blivit uppmätta. Resultatet kan inte generaliseras för en större population, men den kan användas för att ge en bra indikation hur åsikterna för den specifika populationen är. Vidare har en uppdelad analys av komponenterna i RQ gjorts vilket har kunnat säkerställa resultaten i den totala RQ. Användandet av den teoretiskt grundade RQ ger ytterligare validitet åt resultaten. Avgångsstudenter har mest positiva åsikter om Academic Work i jämförelse med de andra undersökta företagen.
Bakgrund:
Metod:
Slutsats:
iii
Table of Contents
1
1.1 1.2 1.3
Introduction........................................................................... 1
Background ............................................................................................1 Problem discussion ................................................................................2 Purpose ..................................................................................................3
2
2.1 2.2 2.2.1 2.2.2
Frame of Reference .............................................................. 4
Corporate reputation in a broader context ..............................................4 Corporate reputation as a concept .........................................................5 The importance of corporate reputation..................................................7 Reputation creation ................................................................................7
3
3.1 3.2 3.2.1 3.2.2 3.3 3.3.1 3.3.2 3.4 3.5 3.5.1 3.5.2 3.5.3 3.6 3.7 3.8 3.9
Method................................................................................... 9
Research approach ................................................................................9 How to measure corporate reputation ..................................................10 League table rating...............................................................................10 The Reputation Quotient ......................................................................11 Sample selection ..................................................................................13 Company sampling...............................................................................15 Sample characteristics .........................................................................15 Psychometrics ......................................................................................15 Choice of data collection ......................................................................16 Questionnaire design............................................................................17 Pilot Study ............................................................................................19 Translation bias ....................................................................................20 Generalisability .....................................................................................21 Reliability ..............................................................................................21 Validity..................................................................................................22 Analysing data......................................................................................23
4
4.1 4.1.1 4.1.2 4.1.3 4.1.4
Results and Analysis.......................................................... 26
The Reputation Quotient ......................................................................26 The Reputation Quotient separated by gender and school ..................28 Disaggregated Reputation Quotient .....................................................31 Analysis of "have no opinions"..............................................................33 The relation between corporate reputation and employment................35
5 6
6.1.1 6.1.2 6.1.3
Conclusion .......................................................................... 37 Discussion........................................................................... 39
Limitations ............................................................................................40 Implications ..........................................................................................41 Further Research..................................................................................41
References ................................................................................... 42 Appendices .................................................................................. 45
Appendix 1 - Most important factor when looking at a company. ......................45 Appendix 2 - Major Source in creating perceptions?.........................................46 Appendix 3 - Reliability testing statistics...........................................................48 Appendix 4 - Mean values and std. deviation ...................................................52
iv
Appendix 5 - Sample characteristics ................................................................54 Appendix 6 - Questionnaires ............................................................................56
List of Figures
Figure 1 - From identity to reputation ................................................................. 5 Figure 2 - Defining corporate reputation............................................................. 6 Figure 3 - How reputation is created .................................................................. 8 Figure 4 - The total Reputation Quotient .......................................................... 26 Figure 5 - Gender separated Reputation Quotient. .......................................... 28 Figure 6 - School separated Reputation Quotient. ........................................... 29 Figure 7 - The disaggregated Reputation Quotient. ......................................... 31 Figure 8 - Graphical overview of non-respondents........................................... 33 Figure 9 - The altered Reputation Quotient. ..................................................... 34 Figure 10 - Willingness to work for the companies........................................... 35
List of Tables
Table 3-1 - The Reputation Quotient................................................................ 12 Table 3-2 - The altered Reputation Quotient (RQ) ........................................... 13 Table 4-1 - Company league table ................................................................... 27 Table 4-2 - Company standard deviation ......................................................... 27 Table 4-3 - Table of non-response ................................................................... 33
v
1
Introduction
This section introduces the reader to the background and the problem discussion. Furthermore, it introduces the concept of corporate reputation, and presents the purpose and the delimitations for this thesis.
Students are graduating from Universities all over Sweden every semester. There are many different and important decisions that have to be made concerning job opportunities. The staffing and recruitment industry are closely monitoring the students' attempts for getting the necessary working experience they so desperately need in order to compete in the job market. This bachelor thesis focuses upon one external stakeholder group to the staffing and recruitment companies - graduating students. Corporate reputation is a concept that has interested researchers for the last four decades (Gotsi & Wilson, 2001), still the concept incorporates a number of varied definitions, contexts and measurement methods. A corporate reputation can be utilised by a firm to improve its competitive situation and to preserve a market position when under pressure (Mahon & Wartick, 2003). Having a positive reputation has been linked to several beneficial outcomes such as; greater intention to buy a product, higher perceived product quality, attracting investors to a larger extent, and enhanced organisation identification among employees (Caruana, 1997). Premium prices for products, lower costs of capital, and labour are some of the more concrete positive effects that can evolve from having a good reputation (Fombrun, 1996). Theory regarding corporate reputation is something that mostly has been established to suit the perspective of a company, in order to find out in which way its stakeholders view the organisation. Fombrun (1996, p. 78) states the following definition, which has been more widely used than most (Wartick, 2002). "A corporate reputation is a perceptual representation of a company's past actions andfuture prospects that describes thefirm's overall appeal to all of its key constituents when compared with the leading rivals." Corporate reputation research has had trouble reaching an unambiguous definition of the concept, despite the increasing interest in writing about it. Along with definition ambiguity comes also measurement concerns. Two major issues where opinions differ is whether to aggregate several perspectives taken into one grand reputation, and whether to make a general index of a company's reputation based on all attributes investigated to be able to rate it against others. However, most researchers agree that corporate reputation is a suitable concept describing the perceptions that stakeholders have of a company and that a corporate reputation is something that can be ascribed to one company.
1.1
Background
The staffing and recruitment industry in Sweden can be traced back to the 1950s. The development has, however, been moderate up until the last 10-15 years. Since then, the industry has expanded enormously. Before 1992, the development had been constrained by legal regulations forbidding any kind of private recruitment mediating. The first reason for the deregulating was the increase in demand for temporary labour in the early 1990s. Deregulations have also been made in several stages since then, with the intention to improve the function of the labour market, which have lead to the recognition of the importance of the staffing and recruitment industry (Bemanningsföretagen, 2005).
1
The structure of the labour market has changed because of deregulation, but also due to other factors. The demand of hiring skilled personnel and consultants has increased and companies today are in greater need of getting hold of and removing employees quickly and easily. The change can partly be explained by vast technological developments in combination with the globalisation process and the more service-focused society in developed countries. Increased competition from abroad, along with a more short-term oriented consumption, has shortened production life cycles. Therefore companies have to be flexible in order to survive the changing circumstances. This flexibility can partly be attained by having a flexible labour force (Bromander, 2007) and a solution to this is to hire services of a staffing firm. More and more companies have also decided to outsource human resource functions to recruitment firms, to be able to take advantage of their more efficient recruitment skills and wide networks (Armstrong, 2006). The staffing and recruitment industry has an important function in the Swedish economy today. Over 55 000 people are employed in the industry, which is about 0.9 percent of the labour force. The industry showed an increase of 18 percent in turnover in 2007 as compared to 2006 and it has nearly doubled in three years time, showing a clear expansionary trend (Almega, 2008). The rest of the EU has an average of 2.5-3 percent of the total employment in the industry, as compared to Sweden's 0.9 percent. In addition, the interest organisation of the recruitment and staffing companies (Bemanningsföretagen) has an expansionary viewpoint, with the goal of having 3 percent of the total employment in the industry (Bemanningsföretagen, 2005). The above stated arguments are all indications that the industry will continue to grow. There are four main service areas within the industry:
• • • •
Staffing or consulting services; renting out temporary personnel to a client firm Recruitment services include complete or partial outsourcing of the recruitment process. Contracting include complete or partial outsourcing of a functional unit within the client firm. Outplacement is a strategic service that includes consultation regarding changes in labour structure in the client firm.
The staffing services accounts for 90 percent and the recruitment services for 6 percent of the turnover in the industry. The remaining two services are relatively unexploited (Bemanningsföretagen, 2005). More and more companies are outsourcing the recruitment function to the staffing companies, since it is a costly process. The staffing companies have also become very competent in the matching process of employees and clients, due to the specialisation in this area and the close cooperation with their client firms. Therefore, the degree of recruitment services is increasing. Staffing and recruitment services are also often combined, providing clients with a type of trial employment for possible future employees (Johansson, 2006).
1.2
Problem discussion
It is not hard to make the assumption that a positive corporate reputation also has a positive effect on the ability to hire or recruit new skilled employees, which is one of the main challenges for recruitment and staffing firms. Following the rapid development of the staffing and recruitment industry, studies have been made to explain its characteristics.
2
Marketing research has been performed by, for example, Bemanningsföretagen, to study the public's opinion on staffing and recruitment companies. These, however, are not fully available to the public and lack a clear conceptual framework. Much of the research made on the Swedish market has focused on the staffing services with various areas of interest and perspectives taken. Some have taken the perspective of the consultants (Bolander &Nord, 1999; Olofsdotter, 2004), working for the staffing companies and explored how they experience their situation and the positive and negative factors belonging to it. Other research has taken the perspective of the customer companies and explored how they experience the consultants (Englund, Nordström & Ohlsson, 2000) or how they handle the staffing process (Strbac, 2006). This study takes a different perspective, as the aim is to investigate one external stakeholder group - graduating students. This group is of interest because it can possibly become an internal stakeholder group - employees. By measuring corporate reputation of firms, one can investigate the subjective perceptions of individual stakeholders towards companies. The perceptions of this group are of great interest to the staffing and recruitment companies. Since the staffing and recruitment industry is developing, more qualified services can be offered. The interests of the staffing and recruitment companies have therefore turned towards employees with higher education, making this an interesting group to investigate. The usage of the concept corporate reputation to assess how possible employees perceive companies that may have interest in hiring them is also appealing since this focus has not been used previously in corporate reputation studies. This study will not separately study the different services within the industry. Since major actors in the industry perform both staffing and recruitment, which also comprise a major part of the industry turnover, this is where focus will lie. It is also not possible to only investigate one service area when examining how the companies are perceived, since the perceptions of the companies are created through the information, communication and experiences exposing stakeholders concerning all services. This study aims to investigate how the stakeholder group, graduating students - potential employees - perceive four staffing and recruitment companies, as representatives of the industry. The corporate reputation measured on the companies sampled will be used to see whether the stakeholder group investigated perceives recruitment and staffing companies in similar or different ways. These implications are to be used to try to conclude how staffing and recruitment companies are perceived by this external stakeholder group. For this study, the population sampled will also be the definition of graduating students. Graduating students are, in this study, defined as currently enrolled students pursuing an academic degree at Jönköping University within business or engineering, who will apply for a job within twelve months.
1.3
Purpose
The purpose of this study is to measure and compare the corporate reputation of four staffing and recruitment companies, from the perspective of graduating students.
3
2
Frame of Reference
In this chapter the authors present an overview to the concept of corporate reputation, its context, and why thesubjectdeserves interest.
2.1
Corporate reputation in a broader context
Much of the research executed on the concept of corporate reputation is connected to that of corporate identity and corporate image. Therefore, the following discussion about corporate identity will be used to introduce and give explanation to the concept of corporate reputation. Many studies have been made on corporate identity but still thereexists ambiguous opinions regarding a proper definition of the concept (Moingeon &Soenen, 2002). Fombrun (1996, p. 36) states that corporate identity "captures the commonly understood features that employees themselves use to categorise how a company approaches the work it does, the products it makes, and the customers and investors it serves." Further, Fombrun (1996) discusses that the corporate identity derives from a company's success and failures during its existence and is recognised by everyone through the name and the many presentations it makes to describe its actions, plans and intentions. These presentations are interpreted by everyone, more or less favourably, who forms images of the company in question. This (corporate) image may or may not be in line with the corporate identity or the many presentations from the company. Stakeholders and the public evaluate, rate, and discuss the images of the company. Fombrun (1996, p. 37) concludes, based on the American Heritage Dictionary's definition of reputation, that corporate reputation is then the "overall estimation in which a company is held by its constituents." It is the affective or emotional interpretation of customers, investors, employees, and the general public to a company's name. Figure 1 describes the connection between corporate identity, image and reputation. In this study only the corporate reputation will be investigated.
4
Figure 1 - From identity to reputation, Fombrun (1996).
2.2
Corporate reputation as a concept
The area of corporate reputation has interested marketing researchers for the last four decades. Many researchers have very similar concepts and definitions, but still there is no unambiguous, generally accepted definition of the concept (Gotsi & Wilson, 2001). Abratt (1989) suggests that part of the confusion arises from the fact that some academics have used the term corporate image and corporate identity interchangeably. In addition, the term corporate reputation and corporate image have been used as identical (Gotsi & Wilson, 2001). The largest confusion in the research has been the issue of whether to separate corporate image and corporate reputation or to use the terms interchangeably. Another issue that differs among researchers is the perspective of whom the subject of the matter regards, meaning; whom it is that has the perception in question. Some researchers state that it is an external audience (Ind, 1997; Moingeon & Soenen, 2002) or outside members (Dutton, Dukerich, & Harquail, 1994) that are in focus. Additionally, some authors use more broad terms, such as stakeholders (Gotsi & Wilson, 2001; Davies, Chun, Silva,& Roper, 2003), implying that it may also include the perceptions of members inside the organisation. Dutton et al. (1994) argue that corporate reputation represents outside members' perception of corporate image, based on his own conceptual framework. Ind (1997) does not differ between corporate image and corporate reputation and suggests that it is "the picture that an audience has of an organisation through the accumulation of all received messages". In this study one external stakeholder group will be in focus, namely graduating students. As stated earlier, one of the major issues that differ in the research is the view of corporate reputation and corporate image (Gotsi & Wilson, 2001). These concepts have been described as identical, totally separated, or as interrelated phenomenon. Gotsi & Wilson (2001) merge marketing researchers broadly into two dominant schools of thought. The first views corporate reputation as synonymous with corporate image. The second school of thought, which is also more recent, considers the two concepts as different. A majority of the researchers, however, regard the concepts as interrelated.
5
Balmer & Stephen (2003) argues that image and reputation are the perception of an organisation by an individual, group or groups. Image, however, reflects perception at one point in time and reputation over time. Balmer & Stephen (2003) elaborates further, saying that reputation is judgements made of the organisation based on the organisation's behaviours, performance, and the collective experiences of the organisation. Reputations tend to be stable but can change significantly due to external or internal events such as industry disasters or ethical lapse by the company. Gotsi and Wilson, (2001) have done a thorough research on the theory available and concluded a definition from all previous definitions used, trying to find an unambiguous definition for future research. They have concluded that "A corporate reputation is a stakeholder's overall evaluation of a company over time. This evaluation is based on the stakeholder's direct experiences with the company, any other form of communication and symbolism that provides information about the firm's actions and/or a comparison with the actions of other leading rivals" (Gotsi & Wilson, 2001, p. 29). Gotsi and Wilson's article (2001) emphasises on the different opinions regarding corporate reputation versus corporate image and argues that the two concepts are not the same but they are highly interrelated. Corporate reputations are largely dependent on the everyday image that people form of an organisation while at the same time corporate reputation can influence stakeholder's everyday image of the firm. Figure 2 describe the interrelation visually.
Figure 2 - Defining corporate reputation, Gotsi & Wilson (2001).
As discussed in section 2.1, Fombrun (1996, p. 36) states that corporate identity "captures the commonly understood features that employees themselves use to categorise how a company approaches the work it does, the products it makes, and the customers and investors it serves." Fombrun's (1996) opinion is that the corporate identity is dependent on all possible presentations that are linked to a company. These presentations are interpreted by everyone who takes in images of the company in question. These images are interpreted by the public and stakeholders to shape the perceptions (i.e. reputation) towards the company. Further, Fombrun (1996, p. 78) also introduces what the author calls a working definition of corporate reputation; "A corporate reputation is a perceptual representation of a
6
company's past actions and future prospects that describes the firm's overall appeal to all of its key constituents when compared with the leading rivals." According to Wartick (2002) the working definition by Fombrun has been used more widely than most. However, in his reflection about the research made on corporate reputation, one distinct point of reflection is given on Fombrun's definition. Wartick (2002) finds the idea that all stakeholders' perceptions have to be aggregated into one assessment of reputation as exaggerated, foremost of measurement concerns. Also the fact that all constituents are included in the definition means problem because of the fact that all stakeholders may be of different importance and interest to the company. Wartick (2002) believes that the definition is designed in this way to suit the most common way of using measurement of corporate reputation, namely to rate and compare companies towards each other. Wartick (2002) argues that a disaggregated perspective would give a more truthful definition rather than saying that reputation is the overall appeal to all of its constituents. This study will take the critique appointed by Wartick (2002) to the definition into consideration. Therefore a disaggregated perspective will be used in this study to illuminate the perceptions about the companies. The corporate reputation will not be investigated for more than one stakeholder group, namely the graduating students. 2.2.1 The importance of corporate reputation
"To economists, reputations are traits that signal a company's likely behaviour. To strategists, reputation is a barrier to rivals, a source of competitive advantage. To marketers, reputations are perceptual assets with the power to attract loyal costumers." (Fombrun et al., 2000, p. 241) Trends are making stakeholders' perceptions more critical to companies and they invite managers to take a more active, centralised, focused, and scientific approach to communicating with those stakeholders. As the interest to value and manage reputation has grown, so has the academic focus towards the subject (Fombrun et al., 2000). Research and theory regarding corporate reputation is something that mostly has been established to suit the perspective of a company, to find out how its stakeholders view the organisation - in order to manage the reputation in ways that benefit of the company. A positive reputation can have several beneficial consequences. Some of the outcomes that reputation has been linked to are: the intention of buying a product or service, the attitude of buyers towards salespersons and products, perceived product quality, attracting investors, and organisation identification among employees (Caruana, 1997). Mahon &Wartick (2003) declare that most research has been conducted on the basis of two major ways to utilise reputation. The first is in assisting the organisation in improving its competitive position - an offensive use of reputation. The second is to use reputation in a defensive matter, to preserve a market position when being attacked. Fombrun (1996) develops these aspects when meaning that a number of tangible assets will evolve from having a good reputation; possibility to charge premium prices for products, lower costs of capital and labour, improved loyalty from costumers, and a cushion of goodwill when a crisis hits. These are consequences that should all contribute to performance difference between firms. Fombrun (1996) continues stating that in order to acquire a positive, enduring and resilient reputation; managers have to invest heavily in building and maintaining good relationships with the company's constituents. A firm with a good overall reputation owns a valuable asset (Caruana, 1997). 2.2.2 Reputation creation
Bevis (1967, cited in Davies et al., 2003, p. 63) describes reputation and also the creation of it in these words; "It is the net result of the interaction of all the experiences, impressions,
7
beliefs, feelings, and knowledge that people have about a company". Corporate reputation is likely created by more than one contact with the company. The perception of how to see a firm is built up of numerous communications (Davies et al., 2003). Bernstein (1967, cited in Davies et al., 2003) describes the complexity of these factors in his model (figure 3), and argues that all images of an organisation are seen through a number of filters, which implies certain preconceptions that influence the image received. The industry and country "filter" in particular, will influence all images received but will at the same time give the same effect to every company sharing the same filters. What differs in the creation of reputation between specific "filtersharing" companies are other factors, or communications, which are shown in the outer ring of Bernstein's model. This model (figure 3) can describe how different stakeholders get their perceptions that ground the corporate reputation for a specific company.
Figure 3 - How reputation is created, Bernstein (1967, cited in Davies et al., 2003 p. 63).
8
3
Method
The next chapter provides a presentation on howtomeasure corporate reputation. Moreover, population and the sample will be discussed as well as the composition of the survey used for collecting the empirical data. In addition, the later sections will touch upon generalisability, reliability, and validity.
3.1
Research approach
Most research starts with either of two different approaches. The deductive approach refers to the logical process of reaching a conclusion from something known to be true, and the inductive approach which involves the process of establishing a general proposition based on observation and particular facts (Zikmund, 2000). In research both approaches can be present. Experiences allow us to draw conclusions that also can be empirically verified by using scientific methods (Zikmund, 2000). This thesis aims to find out how graduating students perceive the staffing and recruitment industry. There have not been many studies conducted to assess one external stakeholder's view of a staffing and recruitment company, something which gives this study an inductive character because of the possibility of reaching conclusions about this specific subject. However, since the respondents' perceptions will be investigated and evaluated based on scientifically grounded facts concerning corporate reputation, the study also has a deductive character. Another aspect to consider, which should fit the purpose of the research, is what type of study the research conducted should have; whether it should be descriptive, explanatory or exploratory. According to Robson (2002), exploratory studies are viewed as finding out what is occurring in a situation; to find insights and to ask questions that can highlight a phenomenon in a new way (cited in Saunders et al., 2007). This study uses different theories concerning corporate reputation to assess the perceptions of the population sampled. The approach taken has not been applied in previous research, and can therefore shed new light on the situation of the graduating students and on the companies investigated. Robson (2002) also explains the descriptive research as a way of assessing a truthful profile to persons, events, or situations (cited in Saunders et al. 2007). For this study, the aim is to give an accurate picture of the perceptions of graduating students, and therefore the study is of descriptive character. Further, this study aims to compare the perceptions towards the four staffing and recruitment companies. This information is attained by using an altered version of the Harris-Fombrun Reputation Quotient (RQ), (see table 3-1 and table 32). This type of study also indicates explanatory characteristics. Saunders et al. (2007) presents explanatory studies as studies on circumstances or problems made in order to describe relationship variables. Therefore, all three of the different study types are present in different aspects of this study. Additionally, there are generally two different research strategies to collect data for empirical research. The quantitative methods are usually measured in numbers or words coded into numbers. Secondly, the qualitative approach put emphasis on words to give a rich picture of the situation and environment (Jacobsen, 2002). The qualitative data collection can come from observations, individual interviews and group interviews. The quantitative strategy can involve data collection through surveys that can be sent out in various ways or conveyed through an interviewer reading the questions for the respondent (Jacobsen, 2002). There are many pros and cons with both strategies; the approach and purpose of the research have to decide which strategy to choose. When conducting
9
exploratory research, qualitative data can give you deeper knowledge of the concept or the investigated problem rather than giving a greater amount of data. However, the quantitative research approach has a clear advantage considering the enhanced possibility for the researcher to make generalisations about the population investigated (Zikmund, 2000). This study is constructed with support by the theoretically grounded measurement, the Harris- Fombrun Reputation Quotient, which so far has been used foremost in quantitative studies. Thus, this study also has a quantitative approach when using questionnaires as method of data collection. Since the aim of the study has been to make some claims about the population investigated, a quantitative method was best suited because of the increased degree of generalisation. Perceptions are feelings and attitudes, which preferably are rated and analysed through a survey rather than through an interview (Fombrun et al., 2000).
3.2
How to measure corporate reputation
In the recent years much more academic research has been devoted to corporate reputation, which has contributed to an increased value of the concept to both practitioners and scholars. As a result, more models and measures for corporate reputation have emerged (Hillenbrand & Money, 2006). Many of the early methods have been developed for ranking companies in, for example, annual beauty contests such as "most admired companies" (Fombrun, 1996), wherein the most visible is probably Fortune's Most Admired Companies (MAC) (Fombrun et al. 2000). 3.2.1 League table rating
One of the approaches to measure corporate reputation is by using league tables (Bromley, 2002), which involves rating companies against each other according to a "reputation score" in a table. Fortune most admired companies (MAC) is one example of league table usage. For that study, the league tables are based on large samples of various kinds of stakeholders, however foremost executives, directors, and security analysts. The sample group then rates a selection of companies based on various attributes that are relevant to corporate success. Further, these companies can be divided into sub-samples representing different business sectors (Bromley, 2002). It can be argued that the most recurrently used and discussed set of data in the field of corporate reputation measurement is the Fortune MAC. It has been annually reported since the 1980s, and is based on the assessment of 300 companies in 40 different industries with a respondent group of thousands of senior executives, outside directors, and financial analysts (Wartick, 2002). The Fortune MAC has, however, been criticised by several researchers (Black, Carnes and Richardson, 2000, Brown and Perry 1994, Wartick, 2002, Fombrun, 1996) for several reasons. Brown and Perry (1994) found that the Fortune MAC is very dependent on the financial performance of the company creating a so called halo effect implying that the financial performance alone becomes too determinate in the overall impression. Brown and Perry (1994) argue that this effect has to be eliminated to make the study accurate. Other researchers have, however, argued that the solution does not come that easily. Harsh critique against the Fortune MAC is presented by Baucus (1995), who argues that data is so methodologically suspect that it should not be used at all. Wood (1995) also compares the findings from Fortune MAC as what the foxes think of the hen house (both cited in Wartick, 2002), implying that the outcome of the studies is rather obvious. Szwajkowski and Figelwicz (1997, cited in Wartick, 2002) are less harsh in their critique and think that the problems can be managed. This is also backed up by Wartick (2002), except for a few objections. Fombrun et al. (2000) present a table with the following methodological limitations of the Fortune ratings;
10
"Biased sampling frames, target firms selected by size of revenue, restriction to publicly traded companies, collusion because of sector membership of respondents, overrepresentation of senior managers, directors and financial analysts in samples, respondents may lack direct experience relevant to some attributes (items in rating scales), mainly pen and paper mail surveys." (cited in Bromley, 2002, p. 35). Bromley (2002) continues by adding to the above mentioned arguments that reports describing the surveys do not always state what instructions the respondents were given and how the questionnaires or rating scale were formed. Despite the critique, a number of industry followers within publications have adapted and based their own reputation rating measurements on the Fortune MAC, among them Financial Times and Asian Business (Fombrun et al, 2000). Moreover, Bromley (2002) presents an alternative method for constructing a league table of reputation using a standard rating scale. This method is called the free descriptive method. The method enables respondents to list various attributions (impressions of attributes) based on their experiences and interests. The respondents would be drawn from selected stakeholder groups depending on the purpose of the research. The more frequently mentioned attributions could then be used to construct a league table of merit using a methodology similar to that associated with the Fortune surveys. Attributions, regardless of how they are selected, could be assigned weights according to their frequency of occurrence or estimated importance. The free descriptive method tends to generate a wide range of attributions, most of which are shared by relatively small proportions of respondents. The differences in occurrence of attributions could indicate their relative importance, familiarity or salience at the time (Bromley, 2002). 3.2.2 The Reputation Quotient
In 1999, a number of researchers developed the Reputation Quotient (RQ), also sometimes referred to as the Harris-Fombrun Reputation Quotient (Davies et al., 2003). When developing the RQ, the starting point was 27 attributes (i.e. the items investigated) drawn from the most commonly used instruments to measure corporate reputation (among them Fortune MAC) along with additional factors suggested by academic literature on reputation and items provided by an commercial marketing firm, Harris Interactive. 32 attributes made up a first prototype model which was tested and refined through an initial pilot testing in the US air force and a following series of focus group interviews. A second prototype was tested in the PC hardware industry and, after a third study to validate the model, involving 8545 respondents, the final version of the RQ was presented. The findings demonstrated that people justified their feelings about companies on one of 20 attributes, which in turn was grouped into 6 dimensions (Fombrun et al., 2000), which are presented in table 3-1. Fombrun et al., (2000) suggest that reputation is a construct that combines two factors - emotional appeal and rational appeal, the latter representing a net assessment of the five dimensions, apart from emotional appeal. The 20 attributes are items used for measuring corporate reputation with the Harris-Fombrun Reputation Quotient. These items are easily applicable to use as perceptual questions about firms.
11
Table 3-1 - The Harris-Fombrun Reputation Quotient, Fombrun et al., (2000)
Emotional Appeal -good feeling about the company -admire and respect the company -trust the company Products and Services -stands behind products/services -offers high quality products/services -develops innovative products/services -offers products/services that are good value Vision and Leadership -has excellent leadership -has a clear vision for the future -recognizes/takes advantage of market opportunities Workplace Environment -is well managed -looks like a good company to work for -looks like it has good employees Financial Performance -record of profitability -looks like a low risk investment -strong prospects for future growth tends to outperform its competitors Social Responsibility -supports good causes -environmentally responsible -treats people well
The Harris-Fombrun Reputation Quotient is an index that sums up people's perceptions on these attributes, which can be used to rank companies in regard to their reputations. This aggregated usage of the RQ is, as when using league tables, a source of critique. Wartick (2002) has the opinion that to use one aggregated score to rank companies is not a credible and justified method. The reason for this frequent measurement approach is that the most common reason for measuring reputation is to rank companies against each other. Despite the criticism about lack of a rigorous conceptual definition in the study (Wartick, 2002; Groenland, 2002), Wartick claims that RQ is good in the sense that it is broad and can be applied on anyone, whether they are customers, employees, or the general public. This is, however, somewhat contradicting Fombrun's (1996) definition of corporate reputation which includes "all constituents". Wartick (2002) implies that the different stakeholders may have different importance in determining a grand aggregated reputation and therefore it is of higher interest to examine stakeholders separately. Groenland (2002) claims that the RQ shows satisfactory psychometric properties and that practical experience with the model, in many different commercial settings, increases confidence in usefulness of the method. According to Wartick (2002), the RQ meet most of the characteristics for a solid measure of reputation even though the entire data set is not available in literature. Groenland (2002) has performed a qualitative study to validate the dimensions of the RQ in which all six dimensions are supported as relevant for measuring corporate reputation. Groenland (2002), however, stresses the importance of being open to altering the dimensions depending on what is meant to be studied. For example, the cultural differences are raised as a factor that might affect the appropriateness of attributes being investigated. Groenland (2002) suggests that empirical analysis should be used to address the issue, in order to design an appropriate model with attributes suitable for the matter of interest. In his study, focus groups were used in an initial stage to validate the suitability of the RQ attributes when measuring corporate reputations in the Netherlands. The outcome of the study was that two additional dimensions (national origin and charismatic representatives) should be added in the Dutch case, and possibly in the original RQ-scale. For this study, inclusion of these dimensions has been considered. However, national origin and charismatic representatives were not found adequate for the setting of investigating the graduating students' perceptions of staffing and recruitment companies. This study has used the RQ to investigate how graduating students perceive the recruitment and staffing companies. The study has measured the attributes on all
12
companies investigated, and an aggregated assessment of the outcome of the RQ attributes are presented in a league table. Further, various analyses have been made to strengthen and validate the findings. The corporate reputations have only been determined through the investigation of one stakeholder group, namely graduating students. This disaggregated assessment of stakeholders' perceptions of corporate reputation is encouraged by Wartick (2002). Groenland (2002) stresses the importance of being open to altering the dimensions depending on what is meant to be studied. Therefore, the suitability of all the attributes from the RQ model have been reflected upon, whether to be included in the main survey, before the pilot study. Alteration has been made on the attributes and dimensions of the Harris-Fombrun Reputation Quotient, in order to suite the area and population of interest. Attributes that were not clearly applicable and understandable in the context of the study were removed to fit the sampled population and the investigated topic. This also led to a change in the dimensions of the Harris-Fombrun Reputation Quotient. The attributes and dimensions concluded appropriate for measuring the corporate reputation of staffing and recruitment companies, through the eyes of graduating students, have made the foundation on which to build the study. This is presented in table 3-2. When mentioning the Reputation Quotient or RQ from hereon in the report, this version is the one referred unless other is stated.
Table 3-2 - The altered Reputation Quotient (RQ)
Emotional Appeal -good feeling about the company -admire the company -trust the company Products and Services -the company offers high quality work opportunities -the company is an attractive employer -the company uses innovative marketing strategies -to work for the company will yield a valuable experience Financial Performance & Vision and Leadership -the company takes advantage of market opportunities -the company tends to outperform its competitors -the company has strong prospects of becoming more successful in the future Workplace Environment -the company seems to have a well functioning organisation -looks like a good company to work for Social Responsibility -the company has an important role in the society -the company is important for the job market
The quotient calculated, using RQ, is aggregated into one assessment, an index based on all attributes measured, to compare to other companies. Wartick (2002) does not support this aggregated measurement, claiming the extent to which the attributes used affect the overall impression of the reputation is not apparent. The concern could be solved partly by assigning different weightings to the attributes but Wartick (2002) advocates that this is hard to accurately determine. Wartick's critique has been addressed in the sense that RQ used in this study have also separated and presented by dimension and question to be able to se what relative differences are apparent, between the respective companies.
3.3
Sample selection
In this study the perceptions of one stakeholder group, namely the graduating students are in focus. This group was selected in order to have a more specified population that is clearly defined. It can, however, be argued that this population is still relatively unspecified to examine. In this study graduating students are defined as; currently enrolled students pursuing an academic degree at Jönköping University within business or engineering who will apply for a job within twelve months. Graduating students are, as a population, rather
13
un-investigated as a stakeholder group to the staffing and recruitment industry. It is, however, an interesting and important group to investigate. As the choice of education and the companies marketing efforts towards graduating students can be significantly different between different geographical areas, it might not be accurate to place all graduating students in the same population. Therefore the population is narrowed down to solely graduating students from Jönköping University in the fields of engineering and business. This delimitation has added validity to the study and enhanced the degree of generalisability. Saunders et al. (2007) discusses the generalisability issue when considering the applicability of findings on more than one organisation, and says that in such cases it should not be the aim to produce theory for all populations but to explain the particular setting of interest. This reasoning is also applicable on this study. Considering all graduating students as a part of one population would reduce the possibilities to accurately describe the perceptions of the students towards the companies investigated. The authors have found another aspect to highlight. Most companies in the staffing and recruitment industry have a common denominator, namely to have more job offers to people with an engineering or a business administration education. Therefore, only students enrolled in such programs were sampled. In this study the sample were picked from graduating students at Jönköping University within the fields of business and engineering. This is because all of the companies investigated are represented in the area of Jönköping, and all the respondents are as likely to have been exposed to same marketing efforts of the companies by various fairs and marketing campaigns on the campus area. Lists of graduating students are easily obtained and access to the research objects through classes is possible. This gives the opportunity of convenience sampling usage. This choice of sampling method has also provided the possibility to investigate a large proportion of the population. Saunders et al. (2007) describes the method as haphazardly picking people that are easily obtained. The method is also called Haphazard sampling. The sampling process is thereafter continued until a quota is reached (Saunders et al. 2007). As it was hard to estimate how many respondents that would be willing to participate in the study, a quota were not set at a specific number. The quota was instead decided upon the amount of responses that could be collected and were valid in regard to the required population characteristics. It is also important to note that it is not the study areas of the students, being business and engineering, that are in focus, but the fact that the companies sampled in the study are hiring people in these fields. 125 respondents were investigated in total, divided as; 31 respondents responded to questions on Poolia, Academic Work, and Adecco, while 32.responded to questions regarding Manpower questions. All methods of sampling are imperfect, and there will be some bias present. In convenience sampling, the respondents are picked due to the accessibility. Convenience sampling does, however, jeopardise the representativeness of the sample (Ruane, 2006). This is a source of bias making it possible for people that are part of the population but not being available, being underrepresented. Saunders et al. (2007) exemplify this as instead of investigating all managers you are investigating managers taking an MBA course due to the convenience of obtaining responses. Further, Saunders et al. (2007) argue that this bias is of minor relevance when there are small variations in the population. It can be argued that the differences between the persons not present for sampling and the persons that are is quite small. This, because all of the sampled students are enrolled in an education program, and are likely to attend at least some lectures. All respondents sampled therefore have been present in the classes and everyone has been asked to complete the survey. However, since
14
only the graduating students in engineering and business should be sampled a set of control questions helped limit the occurrence of the wrong persons being sampled. The respondents not part of the population were excluded from the study. However, it is hard to eliminate the fact that only the students participating in the lectures are sampled. This is further elaborated in section 3.8. 3.3.1 Company sampling
The companies selected for this research have had four prerequisites. They should all be targeting the investigated population through both marketing and possible employment opportunities, they should all be represented in Jönköping, and they should all be engagedin both staffing and recruitment services. To get this information, randomly picked companies have been investigated through various internet sources to see if they match the prerequisites. Again convenience sampling as presented by Saunders et al. (2007) has been used; the first four companies that matched our criteria were sampled for investigation. The quota of four companies was determined because it was the largest number considered manageable in this study. It was thought to be manageable in the sense that each company would be assigned an adequate number of respondents. Fewer respondents on each company would decrease the degree of generalisability. The reason why we are investigating firms active in both staffing and recruitment services is because the staffing service is very dominant in the industry. Recruitment services are, however, still influential on how the companies are perceived. The contracting and outplacement services are still unexploited and therefore not regarded in this study. 3.3.2 Sample characteristics
The four companies that are subjects of investigation for this study are Adecco, Academic Work, Manpower, and Poolia. The number of respondents was 31 across three out of four companies and Manpower had one more respondent. When distributing the surveys, an effort was made to attain equal number of respondents between the four sampled companies. Even though the number of respondents has been equally allocated the share of respondents in regard to gender and school diverge between the companies sampled. Adecco and Poolia have close to a 50 percent share between male and female respondents. Manpower together with Academic Work have a sample that is more male dominated as close to two thirds of the participants were males. Further, no particular difference can be found between the schools, where the sample was evenly allocated between the engineering (JTH) and business (JIBS) students. The mean age of the respondents were fairly similar between the companies, with mean age varying from 23.37 to 25.06 (See appendix 5 for statistics).
3.4
Psychometrics
The word psychometrics refers to the measurement of subjective estimates, such as judgements of a brand's attractiveness or a company's resolve (Bromley, 2002). The traditional rating procedure in social science is based on the following assumptions; "That the various attributes share a common factor, namely a company's ability to provide valued outcomes, and that the attributions (impressions of attributes) are quantifiable, and ratings constitute valid and reliable measures or indicators of these quantities." (Bromley, 2002, p. 40).
15
According to Bromley (2002), Fombrun demonstrates such a common factor in the RQ scale. However, Michell (1990, 2000) and Kline (1998) have reservations against the way psychometrics is used in the social and behavioural sciences (cited in Bromley, 2002). Bromley (2002) implies that when constructing a psychometric scale for assessing corporate reputation, the ideal procedure is to identify a representative set of attributes or a set of attributes within categories that are relatively independent of each other but still correlated with the overall score or criterion. Again, the Fortune surveys are taken as example. It is likely that the usual attributes for corporate reputation used in the Fortune surveys are interrelated, especially when investigating firms within a particular business sector. Further, it is stated that with reference to corporate reputation, a true score consists of the aggregated score a company would achieve on all the quantitatively scaled attributes related to corporate reputation. Bromley (2002) continues by saying that a test of corporate reputation contains a random set of categories. As the measurement is an estimate of a company's true score, it is the sum of the sub-scores from the items in each of the descriptive categories. The number of categories required and the number of items in each category are issues to be settled by empirical inquiry and practical convenience, within the legitimate psychometric framework. From a reliability standpoint, assuming the test is within the respondents' performance capacity, the larger number of categories and the larger the number of items within the categories the better. Bromley (2002) also states the importance of remembering that subjective attributions represent real and supposed attributes. Real attributes have an independent existence; however attributes may or may not be truly presented by a psychometric assessment. The stakeholder group investigated, that are asked to express attributions about a company, should contain either the entire group or a representative sample of it. The procedure for eliciting the attributions, such as a rating scale, checklist or freedescription should be standardised as should the instructions to respondents and the method of recording and analysing the data. The attributions should be reliable and valid, and a company's reputation score only makes sense when in relation to scores of other companies, and sensibly compared (Bromley, 2002). Kline (1998) and Michell (1990, 2000) attribute a problem to the psychometric exercise because it seems that it is not quantitative in the strict sense of the word. This implies that the numbers assigned to corporate attributes are not quantitative enough. The argument continues by exemplifying that there are no real measurement units for subjective attributions. The meaning of the numbers depends on understanding how corporate attributions have been defined and assessed (cited in Bromley 2002). Bromley (2002) also states that some stakeholder groups find certain corporate attributes appealing while others find other attributes appealing. However, the company success is dependent on theproduction of valued outcomes for all stakeholder groups. A solution to the problem would be to devise separate measures of corporate reputation for each stakeholder groups. Further, the companies could then be compared across the profile of stakeholder assessment (Bromley, 2002).
3.5
Choice of data collection
There are several different ways to use quantitative questionnaires, but overall they can be divided into two categories; interview-administrated and self-completion questionnaires. Interview administrated methods implies that the researcher is present during the data collection, whereas the self-completion refers to methods answered without researcher interference (Brace, 2004). Further, there are also mixed methods of the two mentioned,
16
such as interviewer supervised self-completion and interview administrated methods which can contain elements of self completion for the respondents. Brace (2004) has identified three key benefits of interviewer-administrated interviews; queries about the meaning of the question can be dealt with, misunderstood questions can be corrected, and the respondents can be encouraged to give deeper responses to open- ended questions. Saunders et al. (2007), emphasise that quantitative interviews usually are structured, where the interviewer has a set of questions and asks the respondent face-to- face. Brace (2004) points out that there is an interviewer/interviewee bias present when the interviewer is active in the interview, which means that there is a risk that the interviewer cause bias if, for example, asking the questions in different ways. Also respondents can try to impress the interviewer in his/her answers or try to avoid loss of face. Further Brace (2004) implies that these problems are slightly reduced in telephone-interviews, which is one method of interviewer-administrated interviews. For this study it is important to eliminate such bias in order to give higher validity to the perceptions of the respondents, which this research aims to analyse. Interviewer/interviewee bias is eliminated in self- completion surveys since the interviewer is not active (Brace 2004). Another problem with interviewer administrated interviews is the cost of obtaining a sufficient representative sample (Brace, 2004). In this regard, the self administrated surveys allow that a larger sample being collected for a lower cost since the interviewer does not have to be present for each respondent. This increases possibility to generalise for the targeted population. Saunders et al. (2007), however, point out that the response rate tends to be higher in cases where the interviewer is active. Saunders et al. (2007) also present a version of self completion surveys that tends to have a rather high response rate, and significantly higher than internet based and postal self completion surveys - the delivery and collection surveys. As the name implies this type of survey is in paper form and is handed out and later collected by the researchers. This is a method which brings an important advantage for this study. The sampled population in this study, graduating students, is still active in the university environment and attends classes. Therefore, this method can be used and the surveys can be given the students as they attend classes in auditoriums. This will reduce the constraints of administrating costs and time consumption significantly. Self-completion questionnaires have been chosen for data collection in this study. This is due to the possibility to investigate a large sample combined with low costs of obtaining it, as well as because of the reduction of the interviewer bias. There will also be room for a short presentation of the study if the surveys are handed out in an auditorium, which can give a deeper understanding of the research conducted. 3.5.1 Questionnaire design
As this study aims to investigate the corporate reputation for four staffing and recruitment companies from the perspective of graduating students, the investigation will be based on the already grounded theories on corporate reputation measurement. To examine the concept of corporate reputation, this survey has used the twenty items of the Harris- Fombrun Reputation Quotient (table 3-1) as a starting point. This study uses screening questions (Q17, Q19) in order not to include non-representative respondents in the sample. The screening questions are used to be able to only include the people that are representative for the population. When researching attitudes and behaviours, researchers tend to want to study people with certain characteristics (Brace,
17
2004). For this study it is of value to use screening questions to validate that the students attending classes are graduating students, and not students only taking a course or students that will not graduate and apply for a job within a year. The questions regarding the actual topic of interest are called the main questionnaire (Brace, 2004). When setting up the main questionnaire, it should normally start with the most general topics, before getting into detail (Brace, 2004). In this survey there will be an initial question added, where the respondents are asked to pick the most important factor when looking at a company. This introduced the respondents to the following questions and has been subject for analysis. In the response options, there has also be left a space where the respondents will be able to fill in an alternative answer if they would feel that the RQ factors are not sufficient enough for answering this question. This question has also provided information on what dimension in the RQ that is the most important when looking at a company. However, as it is corporate reputation that is investigated, this alternative has also been included to conclude whether or not overall reputation is most important in determining how to perceive a company. After the initial question, the factors of RQ will be investigated. There are many different response scales to choose between. However, since the RQ table (table 3-1) is built up mainly of different claims, a Likert scale seems to fit this situation best. The Likert scale is one type of attitudinal rating scale, which is sometimes called an agree-disagree scale (Brace, 2004). Most of the times the scale ranges between 1-5, where 1 represents strongly agree and 5 strongly disagree (Brace, 2004; Buglear, 2005). When setting up Likert scale type of surveys it is important that the statements are clear, so that the respondents are reluctant to pick the middle answer, neither agree nor disagree, due to the reason that they do not fully understand the statement (Buglear, 2005). Saunders et al. (2007) suggests that one could use an even scale (1-4) to force the respondents to take a stance. However, this can also make the respondents pick alternative that they do not fully agree with, which gives the answers less accuracy. Brace (2004) reflects upon the issue of what the author calls don't knows. People are reluctant o leave a scale blank even if they do not really have an opinion or do not know what to answer. Therefore, it can be good to include the don't know as one option in the scale. Further, Brace (2004) suggests that researchers should consider whether they want to be able to distinguish between mid-point responses, such as neither agree nor disagree, and responses made from respondents stating that they genuinely did not know what to answer or could not answer. This type of non-response can raise issues when using certain data analysis techniques. Taken Bromley's (2004) suggestions into consideration, this study the above mentioned has been taken into consideration, and an uneven scale (1-5) will be used. Besides that there will also be a sixth alternative included, in form of a have no opinion option. The have no opinion is included because the desire to distinguish the ones who may not consider themselves to have sufficient knowledge about the companies to state their perception from the ones knowing, but having a neutral opinion. These non-responses can also be further compared between the companies. Buglear (2005) points out that it is of significance not to include too many choices on the scale and beyond six can make the respondents indifferent between the options. Another important issue to regard is that if you use a series of statements you should keep the same order of response categories to avoid confusing the respondents (Dillman, 2000 cited in Saunders et al., 2007). As the RQ factors are grouped into six dimensions; emotional appeal, products and services, vision and leadership, workplace environment, financial performance and social responsibility.
18
The dimensional structure in the Harris-Fombrun Reputation Quotient has also been implemented in the questions used in this study, to ensure a logical flow of the questions. Another issue that can be of value to investigate in this study is how reputation is created. Bernstein presents a model (figure 3) for how corporate reputation is created, and the factors listed in this model will constitute the base of response options to a question that will be added into the questionnaire. There will, however, be slight modifications on some of the factors. Price points will be excluded from the response options due to the fact that it is not of interest for the stakeholder group studied in this research. Price points are more relevant for firms that are producing a product or service for its customers and customers of the staffing and recruitment companies are the firms that hire them to get staffing labour or recruitment services. Design will also be excluded as response option, since this is related to products or services as is Price points. Direct mail will also include e-mail, and media comment will also include internet sources. There will also be given room for the participants to express another response than the ones provided from the Bernstein model. The inclusion of reputation creation can give the study more credibility as a whole since it gives information on where the perceptions come from. For introduction purposes, some minor instructions prior to the questions are going to be presented in order to help minimise misinterpretations of how to fill in the questionnaire. The survey questions must have a constant and logical flow to help ensure that the questionnaire works as a conversation with the respondent. Most of the time this helps to get the candidate more involved and eager to devote attention and to answer as truthfully as possible. As with all methods there are always some flaws to be aware of when using Likert scales. Brace (2004) has identified four main problems; order effect, acquiescence, central tendency and pattern answering. The order effect arises from the tendency of respondents to pick alternatives to the left (Arlingstal, 1978, cited in Brace, 2004). Acquiescence means that respondents are more reluctant to pick disagree answers than agree (Karlton & Schuman 1982 cited in Brace 2004). Brace (2004) suggests that a possible solution is to place the disagree options to the left of the scale. Central tendency refers to the respondents as more willing to choose a middle option than the extreme ends of the scale. Albaum (1997) suggests the usage of a two-stage-question, which implies that you should divide your 1-5 scale into first asking the respondent if they agree or disagree and then in a second stage ask how strongly you would agree to your just given answer (cited in Brace, 2004). However, since the scale used in this survey will be made a bit differently with the choice of no opinion, it makes the extreme ends look differently than the traditional scale (1-5) and this segmentation was not needed. Brace (2004) further discusses that this type of two- stage-questions can be very time consuming and are more appropriate for telephone interviews. Pattern answering occurs when a respondent starts ticking boxes as a routine without considering the statements. This is often due to boredom or fatigue. Brace (2004) suggests a solution to mix both positive and negative statements, so that the respondents really have to read the questions, which have been considered for this survey. There has also been left some "white space" between the groups of questions representing the six dimensions of the RQ scale, in order to possibly prevent patterned answering. 3.5.2 Pilot Study
Quantitative research methods such as questionnaires almost always have identifiable errors. The errors range from typographical mistakes, form errors, presentation flaws to ambiguous survey instructions. Most of the time this occurs since the researchers are
19
working too close to the project and thus may overlook even the most obvious errors (Litwin, 1995). Researchers decide to conduct pilot studies to measure the range of ideas or opinions that people have or to find out the way that variables are related to each other (Fowler, 2002). For a pilot test to be effective, it is important that the respondents are similar to those who will be asked to participate in the main survey (Fink, 1995). In this study the target group is graduating students from Jönköping University within business or engineering. In order to use a sample as similar to the population as possible, the actual population was sampled. Therefore engineering and business students graduating from Jönköping University constitutes the pilot test sample. There are some key factors in the pilot survey that have been set under specific attention. Since the aim of the survey is to examine corporate reputation based on the already grounded theories on the matter, it is important that the respondents can locate and differentiate their response according to the ranking figures. Pilot surveys are an essential part of that trial process (Brace, 2004). The pilot testing allows the author a chance to correct these flaws before the main survey is handed out and used to gather real data (Litwin, 1995). Brace (2004) argues that failing to pilot the questionnaire thoroughly risk leading to failure with the project. Brace (2004) has pointed out some reasons why a questionnaire should be piloted and in this pilot survey the following is taken into consideration. First of all, the questions have to sound right. Most of the time questions can look acceptable on paper, but errors and misinterpretations are revealed when read out loud. This pilot survey has been both read out to its respondents and after the study was filled in, a complementary interview regarding the questionnaire design and content has been performed. Another thing to keep in mind is if the respondents understand the survey questions (Brace, 2004). Simplified language should be used to avoid complicated words and ambiguous questions which might lead to incomprehensible answers. No leading or loaded questions and no right and wrong should be promoted throughout the survey. Further, the researcher should ask the participants if the survey allowed them to say everything they wanted on the subject. This can help locate if some questions are imperfectly formulated or confusingly explained (Brace, 2004). The performed pilot study along with the following discussion of the questionnaire provided many valuable insights to support a restructuring of the questionnaire. The altered questions were tested and the general structure was considered suitable. Some ambiguous questions were altered and some unobvious instructions were clarified. Some of the questions were experienced differently in the Swedish and the English versions of the surveys and therefore corrected. 3.5.3 Translation bias
There are many biases to consider when conducting research, and one of them is the issue of translation. The surveys used for this study have been both in Swedish and English. This has been used due to the differences in respondent groups. The programs at the engineering school are mostly taught in Swedish therefore the Swedish surveys are used there. The programs at the business school are mainly taught in English and there are many international students present. Therefore English surveys have been used there. The surveys were built up the exact same way, with the same questions. The translation of the surveys has, however, been done by the researchers of the study. To ensure that the questions still sounded the same way, and the meaning of the questions were understood
20
the same way by the respondents, the surveys were piloted on several bilingual persons representative for the population investigated.
3.6
Generalisability
Vogt (1993) has described generalisability as; "the extent to which you can come to conclusions about one thing (often a population) based on the information about another (often a sample)" (cited in Collins and Hussey, 2003). One of the most important things about generalisability is that the findings should be applicable to another research setting, such as other organisations (Saunders et al. 2007). In other words, this is in line with the definition by Vogt (1993). As to ensure a high degree of a possibility to generalise the sampling method is crucial to avoid the many biases connected to it. Many researchers have argued that the only sufficient sampling method to use to be able to generalise for a population is random sampling. For this study, convenience sampling has been used; this type of sampling has a degree of randomness connected to it. However, it is not the same as random sampling, the respondents are picked due to the accessibility and, as discussed in section 3.3, this creates an underrepresentation of people not attending classes where the population was sampled. Further, the sampling turned out to be unevenly distributed in terms of gender and school. To be able to fully generalise for the population, the sample should be representative for the characteristics of the population. This implies that the sample should have the same gender and school distribution as the population as a whole to avoid a bias. For survey research, the issue of generalisability is closely linked to the issue of validity. (Ruane, 2006). Even though a pilot study was conducted, one can not completely rule out that the respondents have interpreted the questionnaire questions differently than the researchers intended them to do. This lowers the degree of validity in the findings, causing a decrease in the level of generalisability. Due to the discussed flaws, connected to the study, the authors find that the degree of generalisability has been affected negatively. Therefore, the results presented should be interpreted as an estimate and not as the truth. As Ruane (2006) argues it is important to point out is that research does not describe the truth but instead gives an estimate on how the reality looks like. There will always be biases present.
3.7
Reliability
In general, a reliable measurement is one that yields the same result if measured more than once. This holds only if there has not occurred any change in the variables measured (Ruane, 2006). Further, reliable measures do not fluctuate due to random events, but yield consequent results. This is quite easily controlled when a particular instrument is used, such as a thermometer that measures a temperature of something. Measuring the same thing more than once to check reliability of measurement is called test-retest (Gomm, 2004). However, when investigating people and their behaviour and opinions it gets more complicated. To ask a person to do a test twice can seem rather awkward and the person is likely to just repeat what was just stated. This is also why the time aspect is of interest. If doing a retest within an interval that is to narrow, the answers attained are most likely influenced more by the memory than the actual feelings (Ruane, 2006). For this study a retest is not possible, since the respondents are not available more than once. Also, there was another aspect to consider. Since this study aimed to investigate perceptions, it can not be guaranteed that the respondents feel the same way today as they do, for example, in a month. As presented in section 2.2.2, there are several ways that
21
reputations are created, and as the respondents get more impressions of the companies their perceptions may change. Therefore it is not accurate to do a retest after a long time period. Ruane (2006), however, states that this does not necessarily have to indicate low reliability. There are other ways to assure that what is measured is reliable. In survey research it is common to instead ask several questions on the same concept and then compare the answers. For example, a question on age can be asked both as how old a person is, as well as the year the person was born to assure that the same age is given in the two questions. One type of question asked like this is called single items (Ruane, 2006). There are also multiple items that are used when you ask several questions regarding the same aspect. This is not as straight forward as age, but still useful to assure reliability (Ruane, 2006). The Fombrun-Harris, Reputation Quotient is constructed in this way, being a reliable scale to measure corporate reputation. The surveys used for this study are constructed with the RQ as a base but altered to better fit the population investigated. When you are investigating one phenomenon, there is a common technique used for measuring reliability, namely Split Half, used for checking a scale for internal consistency (Ruane, 2006). The questions that are investigated are divided into two sections, and then the scores of the two groups are compared by looking at how closely they are correlated. This coefficient is called the Cronbach's alpha and varies between 0.0 and 1.0, with 1.0 being perfectly correlated (Ruane, 2006). Since a measurement of corporate reputation is done in this study, the Split Half technique can be applied and calculated. Pallant (2006) suggests a number of steps when checking for reliability with the help of SPSS. First, all mean scores should be examined to see whether they are what could be expected. Secondly, the Cronbach alpha value should be calculated to check internal consistency of the scales. If the alpha values are too low, you may need to consider removing items with low item-total correlations. Pallant (2006) claims that an alpha value over 0.7 is ideal and shows that the scale is reliable with the sample. For this study all means of all questions, for all companies are within expected ranges. Three out of four Cronbach alpha values where over the required value for reliable scales and thus showing internal consistency. The only faulty alpha value was the one for Adecco (0.665). The Alpha if item deleted option has been used in order to see what impact the removal of each item has on the alpha value. The results show that when removing the Marketing item, the alpha value was 0.775 and thus above 0.7. Reflection has been made if the item of the scales needed to be excluded. Since the remaining three companies show no similar signs to the removal of the Marketing item, it will be kept for analysis, also for Adecco. This result will, however, be accounted for when further analysing the data. All statistics from the reliability testing can be found in Appendix 3.
3.8
Validity
There are many different aspects of validity, but mostly it regards whether using concrete, observable support for claims on the reality or not (Ruane, 2006). The aspect of measurement validity refers to whether the study measures what it is supposed to measure. Ruane (2006) stresses the importance of the issue by exemplifying the debate of lie detector tests. The instrument is argued to have low validity due to the fact that it does not really measure if a person is lying or not, but other factors that can indicate that a person is lying. In this study the measurement validity is of great importance, due to that the questionnaires used for empirical data collection have to be valid and measure corporate reputation. Since the Harris-Fombrun Reputation Quotient is a recognised way to measure corporate reputation, this leads to a higher level of validity. However, not all items in the RQ are
22
applicable as aspects to consider for all stakeholder groups. Therefore, the items investigated have been altered so that the answers retrieved from the respondents are credible perceptions, see table 3-2. This is also of value when considering the aspect of external validity. External validity refers to what extent the results of a study are applicable to other settings and groups (Ruane, 2006). The findings of this study can be applicable to other settings. The findings of this study can also give an estimate of how the reality looks like for this setting. However, since corporate reputation is a widely debated concept that deals with perceptions, it can not be argued that the findings from one population also hold for another. External validity is also linked to the sampling method. When sampling people as respondents to answering a questionnaire, it created a bias, this due to the under representation of the uncooperative respondents Gomm (2004). However, for this study, this has not been a major problem since almost all the respondents that were given a questionnaire also returned it. The under represented group in this case would instead be students not attending classes. This can possibly lower the validity and the possibility to generalise for the population as a whole. Yet, referring back to section 3.6, the findings of this report are an estimate and not an exact picture of the truth. There are several more biases that can reduce validity of a study, but one might be more significant than others; there is no guarantee that the respondents have interpreted the statements in the surveys as the researchers intended them to do. In survey research this can never fully be eliminated, but for this study the pilot study was used to try to reduce this type of bias as much as possible.
3.9
Analysing data
When conducting statistical analysis today most of the methods involve applying a statistical computer program to help deal with the empirical data. There exists a large variety of statistical software for analysing purposes both for professional and personal use. Some of the more frequently used software packages are SAS and Minitab although the most widely used statistical software today is SPSS (Statistical Package for the Social Science) (Ghauri & Grønhaug, 2005). The software SPSS has been used in order to compile and analyse the results from the survey data. According to Byrne (2002), the use of SPSS for a presentation of a large number of variables is best created through an exploration process. An exploration is a procedure of organising the data so that it can be ordered in some way to find out if any meaningful patterns can be extracted from it. The summarising is done through standard descriptive statistics, which is a method to describe the overall character and patterns of the variables. Examples of descriptive statistics are graphs and tables (Byrne, 2002). Quantitative data can be divided into two groups; categorical and quantifiable data. Also Microsoft Excel has been used to be able to present more intuitive graphs and tables. Categorical data, sometimes called nominal data, refers to data which values cannot be measured numerically. It can, however, be classified into categories (Saunders et al., 2007). The initial question (Q1) together with Q16 in the questionnaire yield nominal data since the variables cannot be measured numerically, but set into a specific classification according to the responses given. Nominal data is categorical data where the order of the categories is arbitrary (Saunders et al., 2007). Ordinal data is a more precise form of categorical data, where there is a logical ordering to the categories (Byrne, 2002). The following questions in the survey (Q2-Q15) are of ordinal
23
data character since the format of Likert scale is used to classify the respondent's perception towards the different companies. To put the answers in relation to the population sampled and to describe the characteristics of the sample, a set of control variables has been included in the bottom part of the questionnaire. The remaining questions (Q17, Q18) are not a subject of further analysis, but used solely as screening questions to exclude non representative persons for the population. Pallant (2005) suggests a methodical way of processing data gathered through with the use of SPSS. This way first includes preparing the data by defining variables and coding responses. Then data should be entered in a data file and then screened and cleaned to check and correct possible errors in the data. If errors are found, one should identify them is SPSS and go back to the questionnaires in order to correct them to the accurately given values (Pallant, 2005). These steps have been performed for the respective four companies in separate data files. The response variables have also been manipulated in the sense that the coding for the negatively worded items has been reversed. When designing the study, the choice was made to include a non-response item - have no opinion. The screening of the data showed that a fairly large number of cases (persons) have marked the answer have no opinion in the questionnaire, indicating that they have no opinion about the particular perceptual question. These responses have been treated as missing data in the analyses. Pallant (2006) suggests using the option exclude cases pairwise when analysing the data in SPSS, which excludes the cases with missing data only if they are missing data required for a specific analysis. The cases will still be included in the analyses for which they have the necessary data. Further, the screening showed that the number of actual missing data (i.e. no answer given at all) was very small - only a few cases had a one or a few blank response boxes. These answers have been treated as missing data, i.e. been given no value in order not to effect the further analyses. In a separate analysis about the have no opinions these actual missing values and the have no opinion responses will be merged, in order to separately analyse these missing values through SPSS. This procedure can also be supported by reflecting on the likeliness that a missing answer might in fact be an actual statement saying that the respondent had no opinion or knowledge about the subject in question. Apart from this, the data did not need further corrections, since the data was within all limits of expected values. The characteristics of the samples, drawn from the descriptive statistics used in these steps, can be found in section 3.3. Further, the data have been checked for reliability. These results can be found in section 3.7. A Reputation Quotient has been calculated for each company from the means of each respective question for each respective company. To put the answers in relation to the population sampled, the Reputation Quotient has been analysed in relation to the gender, age, and respective school of the respondents, and whether the respondents could consider working for the respective companies. This creates more depth and reflection upon the mean values and ranks presented as the Reputation Quotient in the league table. An analysis without the purpose of making statistical inferences has been performed in order to highlight patterns, similarities, and differences for each company as well as in relation to the other companies. The analysis concerning the age control variable did not show any noticeable difference in any aspect, and is therefore not presented or dealt with in the analysis.
24
The perceptual answers have also been analysed separately in relation to the corresponding answers for each respective company. As to see where the relative differences in perceptions may lie between the respective companies, the aggregated RQ-measure has been disaggregated by each dimension and item. An analysis has been made, not only to compare the companies, but also to give an idea of the internal relation for each respective dimension of the RQ. The nominal data questions (Q1, Q16) have been analysed in relation to the findings and analysis of the questions regarding the corporate reputation and the actual RQ findings of the companies. This has given another dimension and perspective to deepen the analysis on the companies.
25
4
Results and Analysis
This chapter provides the reader with descriptive statistics that present the results of the empirical findings. Further, the Reputation Quotient will be analysed both from an aggregated as well as disaggregated perspective.
4.1
The Reputation Quotient
Reputation Quotient 51,79 52 51 50 49 48 47 46 45 44 43 42 Adecco Manpower Poolia Academic Work 45,9 RQ 47,97 48,07
Figure 4 - The total Reputation Quotient
In this study, the corporate reputations of four staffing and recruitment companies have been calculated. 125 graduating business and engineering students from Jönköping University have been questioned in total. Through convenience sampling a set of 31 survey answers have been collected regarding Adecco, Poolia, Academic Work, and Manpower - having one more respondent. These findings are summarised and the companies are ranked in a league table (table 4-1). The values presented in the league table are derived from the perceptual responses on questions CR2-CR15 in the questionnaire answered by the graduating students (see Appendix 6). The individual responses to the questions have been assigned the values 1-5. The values that make up the league table are the calculated mean values on each question, per company. The sum of all mean values for each respective company makes up the Reputation Quotient, which is used to rank the companies in relation to the others. The Reputation Quotient is also displayed in figure 4. Since the responses have been assigned the value 1-5, the RQ can theoretically range from 14 to 70. As can be seen in figure 4, the RQ differs from each of the four investigated companies. The findings show that Academic Work is ranked highest out of the four staffing and recruitment companies, by the graduating students, with a total RQ of 51.79. Poolia and Manpower both have very similar RQ values of 48.07 and 47.97 respectively. Adecco is the
26
company that received the least positive RQ value with 45.9, when graduating students had their say about the corporate reputation of these four companies.
Table 4-1 - Company league table Companies Academic Work Poolia Manpower Adecco
CR2 3,5 3,64 3,55 3,38 CR3 2,92 2,62 2,46 2,61 CR4 3,74 3,8 3,52 3,68 CR5 3,73 3,63 3,52 3,3 CR6 3,76 3,64 3,64 3,52 CR7 3,78 3,09 3,44 3,04 CR8 3,4 2,88 2,85 2,88 CR9 3,9 3,38 3,5 3,35 CR10 3,47 3,52 3,4 2,95 CR11 4,05 3,55 3,64 3,73 CR12 3,44 3,1 3,27 3 CR13 3,8 3,7 3,63 3,24 CR14 4,04 3,64 3,55 3,52 CR15 4,26 3,88 4 3,7 RQ 51,79 48,07 47,97 45,9 RANK 1 2 3 4
A disaggregated league table of the RQ can visualise where the differences are, and in what aspects the companies are more or less positively perceived in respect to its competitors. This is presented in table 4-1. Academic Work has received the highest mean value in 11 of the 14 questions in the RQ scale. Further, table 4-1 shows that Adecco is perceived in the least positive way, in relation to the other companies, in 10 of the 14 questions.
Table 4-2 - Company standard deviation Companies Academic Work
CR2 0,812 0,86 0,767 0,703 CR3 0,83 CR4 0,752 0,78 0,995 0,7 CR5 0,631 0,804 CR6 0,437 0,816 1,021 0,869 CR7 0,85 0,565 0,793 0,852 CR8 1,046 0,727 1,056 0,783 CR9 0,436 0,733 1,189 0,9 CR10 0,874 0,881 0,757 0,823 CR11 0,524 0,941 0,768 0,602 CR12 0,814 0,73 1,099 0,767 CR13 0,676 0,858 0,745 0,766 CR14 0,976 0,831 0,866 0,797 0,571 0,638 0,945 CR15
0,619Poolia 0,77 0,667
1,154Manpower 0,893
0,683Adecco
Table 4-2 presents the standard deviation of each question for each of the four companies, which can explain to what extent opinions differ among the respondents on each question. The standard deviation is the most commonly used measure of summarising data dispersion. The standard deviation reflects the degree to which the values in a distribution differ from the mean. Standard deviation is intuitive in the sense that it is expressed in the original values of the data (Bryman, 1999). The deviations of the answers, if regarding all four companies, are ranging from 0,436 to 1,189. In reflection to the questions posted, the two questions with the lowest deviating values are CR6 and CR13. In respect to the companies, the respondents of Manpower seem to have the most diverse answers. However, as can be seen in table 4-2, the dispersion is fairly similar between most questions and no direct patterns can be found in it. Also in regard to the individual mean values presented for each questions in table 4-1, no direct connections are obvious between high or low mean values and high or low dispersion.
27
4.1.1
The Reputation Quotient separated by gender and school
RQ 54 51,97 52 50,13 50 47,89 48 46 44,34 44 42 40 47,07 46,93 49,4 51,2 Adecco Male Adecco Female Manpower Male Manpower Female Poolia Male Poolia Female Academic Work Male Academic Work Female
Figure 5 - Gender separated Reputation Quotient.
In figure 5, is displayed the perceptions about the four companies when separated by gender of the respondents. This is done to see if there are any gender differences in the perceptions towards the companies. The analysis is done to strengthen the findings of the overall RQ and to give an implication on how the scores are built up in relation to gender. These values have been calculated in the same way as the total RQ, but as separated by gender. Three out of four staffing and recruitment companies are more positively perceived by women than by men. Academic Work is the only company perceived as more positive by men than by women. Still, Academic is more positively perceived than all other companies for both men and women. If comparing Manpower and Poolia, it is surprising that Manpower has a higher value both when comparing men and women with respective category for Poolia, since Poolia received a higher total RQ (figure 4). This can however be explained by the fact that the sample of Poolia have an equal gender distribution whereas the Manpower sample have roughly 66 percent men and 34 percent women. This unequal gender distribution is also shared by the sample of Academic Work to about the same extent, whereas the sample of Adecco is equal in respect to gender (see Appendix 5). This uneven allocation is the result of a haphazard sampling, when distributing the surveys an effort was put into getting an equal amount of responses on each of the companies but not regarding gender distribution. The larger proportion of men and their less positive opinions have a negative impact on the total RQ. This therefore indicate that if the study was conducted repeatedly with a more equal gender distribution throughout the companies the result could have been slightly different result, and it is possible that Manpower would be ranked better than Poolia.
A de cc o M A a l e d e cc o F M em a l a n e po M we r M a n a l po e we r F m e a l P e oli o a M P a l o oa A e li ca F e e d m al A m ic a W or e c de k M mi a l cW o r e kF m e a l e
28
In general, women seem to have less difference in how they perceive the companies than the men. The values for women range from 47.89 to 51.2 whilst the vales for the men range from 44.34 to 54.97. These findings imply that the overall RQ is positively affected by the women in three of the four cases, whereas in the case of Academic Work thepositive effect on the overall rating comes from men.
RQ 54 52 50 48 46,45 46 44 42 A de cc o JIB A S de c c o J M TH a n o p we r JI M B S a n o p we r J TH P o oa li J I BS P A o oa ca li de J T mi H W c A r o ca k JI de B S mi cW o r kJ TH 45,78 47,29 49,84 48,9 47,66 48,62
53,02 Adecco JIBS Adecco JTH Manpower JIBS Manpower JTH Poolia JIBS Poolia JTH Academic Work JIBS Academic Work JTH
Figure 6 - School separated Reputation Quotient.
Figure 6 shows the graduating students' perceptions of the four companies in regard to which school the respondents are enrolled in. The school RQ values are calculated the same way as the total RQ, but divided by schools. It is the sum of the mean values, derived from all questions of the RQ-scale, per company, per school. Students at JTH - the engineering school, perceive three out of four companies in a more positive way than students at JIBS - the business school. Also in this regard, Academic Work is the company that differs from the other companies - being more positively perceived by JIBS students than by JTH students. JTH values are less diverging between the companies than the JIBS values, whereas the value of Academic Work is the one that deviates the most from the other companies. As compared to the other values of the RQ findings, JIBS students' opinions on Academic Work differ quite much from both JIBS students' perception of remaining companies and JTH students' perceptions of all companies. Still, if looking at JTH students' perceptions of Academic Work, they are more positive than both schools' students' perceptions on any other company. If looking at the values of the schools separately, the JIBS values correspond to the ranking found in the total RQ (as presented in figure 4). This, however, is not true for the JTH values, in which Manpower have received a higher RQ value than Poolia, thus being ranked as number two. The samples of the four companies have a fairly equal share of students from each school, except for Adecco which sample is composed of 61 percent from JIBS and 39 percent from JTH. Since the JIBS students have been more negative than the JTH
29
students towards three of the four companies, it can indicate that Adecco, would have been perceived more positively if the distribution would have been 50 percent sampled form the two schools. However, since the values from the two schools are relatively similar, the effect of having equal share from both schools is likely to be similar to the total RQ. Since the share of respondents at JTH were allocated as 39 percent and 61 percent at JIBS, it implies that each individual respondent at JTH influence the JTH-score to a larger extent than each individual respondent at JIBS. It might be coincidental that the JTH students are more positive than the JIBS students in three of the four cases, or there might be an undetected flaw connected to having fewer respondents in general, leading to a higher score. This problem is also present in the gender divided RQ, due to the uneven allocation between men and women. The explanation of the problem can be found in the sampling method. As the respondents were sampled in classes, this was the determining factor of how many people were sampled from each school and the gender distribution. Further, when eliminating the responses that were not representative for the population the largest share of the cases turned out to be from women. An observation made during the sampling process was that the classes at JTH tended to be smaller than the classes at JIBS; this was a contributing reason that more people were sampled from JIBS. As in the separation of RQ by gender, the distribution of the perceptions was more unequal among the JIBS students than for JTH students, ranging between 45.78 to 53.02 and 46.45 to 49.84 respectively. This also implies that overall for the study, the respondents that have been most positive are the men enrolled in a program at JIBS answering surveys on Academic Work, whereas, the male respondents at JIBS answering surveys on Adecco was found to be most negative. As Academic Work perceived most positive at both JIBS and JTH and by both women and men, the findings imply for the same as for males, that women at JIBS are most positive towards Academic Work. Women at JTH have also given a positive perception towards Manpower. The women at JIBS and JTH were both most negative towards Adecco, however, the results imply that women over all were more positive towards Adecco than the men. This discussion is, however, only an assumption. In order to get an accurate estimate of the situation, the separation has to be done dividing up the result by company, gender, and school originating from the total RQ. In this study such a separation would be biased, the share of responses per company divided by both gender and school would yield a too small number of respondents for each measure, and no indications could have been concluded.
30
4.1.2
Disaggregated Reputation Quotient
Disaggregated RQ
4,5 4 3,5 3 Adecco 2,5 2 1,5 1 0,5 0 CR2 CR3 CR4 CR5 CR6 CR7 CR8 CR9 CR10 CR11 CR12 CR13 CR14 CR15 Manpower Poolia Academic Work
Emotional
Workplace
Products/Services
Finance Vision/Leadership
Social
Figure 7 - The disaggregated Reputation Quotient.
Figure 7 describes a disaggregated view of how each of the companies is perceived in regard to each question of the RQ scale. Each value on each question is the calculated mean value of the responses on each company. As can be seen in the graph, the responses in regard to the companies more or less follow a pattern of peaks and pits on each respective question. The most deviating company is Academic Work, which on several questions (CR3, CR7, CR8, CR9, CR11, and CR14) clearly has a more positive mean value than the other three companies. As the original Harris-Fombrun Reputation Quotient consisted of 20 items divided into 6 dimensions, this structure was also used in this study. As can be seen in (table 31) in section 3.2.2 the dimensions are Emotional Appeal (CR2- CR4), Workplace Environment (CR5-CR6), Products and Services (CR7-CR10), Vision and Leadership together with Financial Performance (CR11-CR13) and Social Responsibility (CR14- CR15). For each dimension, most of the items were reformulated, as can be seen in the table 3-2. This was done for enhancing clarity for the respondents. Some of the items were altered or deleted to get a better fit to the population investigated. As this was done, the dimensions still remained and the altered items could be placed into the same original dimensions. However, as Financial Performance and Vision and Leadership were found to have least fit to the population and setting, a few of the items were removed completely, and therefore these dimensions were merged into one (CR11CR13). If looking at the responses on the Emotional Appeal dimension, the first eye-catching difference is that overall the respondents for the four companies have answered positively on the question on whether they respect the company, but none of the companies were admired by the respondents to a larger extent. In the original Harris-Fombrun RQ item, these two aspects were put together; if that would have been done in this setting theanswers would have been misleading. Overall, all companies are positively perceived in this dimension with only small differences, but not on the question on admiration where all the companies were perceived more negatively. The relatively large difference between this
31
question and the others in the dimension raises the question whether it is suitable as to be included in the dimension. On the dimension of Workplace Environment, the results show that the companies are quite equally perceived, and CR6 regarding whether the company has a well functioning organisation, is the aspect where the mean values are most equal in the study overall. One could argue that in the eyes of the investigated stakeholder group none of the companies have a strong competitive edge on this aspect, as they are perceived similarly. The third dimension concerning Products and Services was also ranked as an important factor when looking at a company (Q1). On the three first questions (CR7-CR9), Academic Work distinguishes itself from the rest of the companies. Academic Work is, compared to the other companies, more positively perceived when it comes to providing good job offers and being an attractive employer as well as creating a good merit of employment. If the companies would be interested in investigating this target group, it can be suggested that the Products and Services factors as they are stated in the questionnaire are the most important for getting a picture of how they are perceived by a possible employee. However, on the question of whether a company was using innovative marketing strategies, three of the four companies were perceived similarly with Adecco being perceived more negatively. This question, however, does not necessarily imply that the marketing efforts of the company are bad, as the question refers to the innovativeness of the companies marketing efforts. Further, it can also be argued that this aspect could have been reformulated to make it easier for the respondents to take a stance or to give the respondents a possibility to distinguish between good and bad marketing. The statement could have been formulated in terms of effectiveness instead of innovativeness. The next dimension investigated is Vision and Leadership together with Financial Performance. All firms were found to have a high score on the question on future prospects, which is valuable information to the industry as a whole. Academic Work outperformed its competitors in this aspect as it was ranked relatively higher than the other three firms. Further, Academic Work also got the highest score on the question if the company tends to outperform its competitors. In this question all three companies received similar scores and Manpower, which is ranked the third, got a better score than Poolia, ranked the second in the overall RQ measure. The respondents found all firms to be exploring the market opportunities. However, Adecco was found to have a significantly lower score than its competitors investigated. The last dimension investigated was Social Responsibility, where all the companies scored relatively high values, and all firms were assigned to be significant to the job market and fulfilling an important role in society. CR15, concerning if the company was important to the job market, received the highest scores for three of the investigated companies compared to the other items measured.
32
4.1.3
Analysis of "have no opinions"
Adecco
16,13%
Table 4-3 - Table of non-response
Questions
CR2 CR3 CR4 CR5 CR6 CR7 CR8 CR9 CR10 CR11 CR12 CR13 CR14 CR15
Manpower
9,38%
Poolia
9,68% 16,13% 19,35% 12,90% 29,03% 25,81% 19,35% 16,13% 25,81% 29,03% 32,26% 35,48% 19,35% 22,58%
Academic
1W,13% 6 ork 25,81% 12,50% 19,35% 12,90% 32,26% 25,81% 16,13% 16,13% 35,48% 29,03% 38,71% 32,26% 12,90% 12,90% 9,38% 15,63% 21,88% 21,88% 15,63% 12,50% 21,88% 21,88% 31,25% 25,00% 3,13% 3,13%
22,58% 25,81% 29,03% 45,16% 25,81% 35,48% 32,26% 45,16% 38,71% 48,39% 51,61% 25,81% 25,81%
The cases where the respondents answered have no opinion on question (CR2-CR15) have been analysed separately. This separation was made in order to see to what extent the respondents had no opinion on each question, and how this has effected the measuring of the RQ. As discussed in the section 3.5.1, this choice was made to be able to distinguish between neither disagree or agree and have no opinion, which turned out to give quite extensive effects on the results of the study. While screening data for errors it became obvious that a fairly large number of respondents marked have no opinion and the decision was made to analyse these findings separately. Table 4-3 presents these results as percentages of have no opinion answers out of the number of respondents on each question, per company. The high rate of have no opinions can also be an outcome of pattern answering (as discussed in 3.5.1). However, there is no implication that this would affect any specific company since all respondents are as likely to fall into pattern answering. Pattern answering is, however, a possible bias which cannot be disregarded, since it may affect the overall result of the study.
60,00%
50,00%
40,00% Adecco Manpower 30,00% Poolia Academic Work
20,00%
10,00%
0,00% C R 2 C R 3 C R 4 C R 5 C R 6 C R 7 C R 8 C R 9 C R 1 0 C R 1 1 C R 1 2 C R 1 3 C R 1 4 C R 1 5
Figure 8 - Graphical overview of non-respondents.
33
Figure 8 visualises the same findings. The graph shows that the questions on all four companies follow a pattern of have no opinion on each respective question, except on CR8 where the respondents for Academic Work marked a higher level of have no opinion than on the previous question, whereas the respondents for the other companies marked a lower level of have no opinion than on CR7. The have no opinion answers reaches its peak in CR12 and CR13. This could indicate that these questions have least fit to the population, so if the questions where to be altered again, one could consider to exclude them or re-formulate the statements. As the calculation of the Crombach Alpha exclude the have no opinion answers, these type of flaws are undetected in that calculation. This observation is significant to the overall reliability of the measurement of the RQ. Further, Academic Work has a higher degree of have no opinions in 13 out of 14 questions, which is exceptional compared to the other companies investigated. In reflection to this remark, the question arose - to what extent can this high degree of have no opinion answers, not included in measurement of the RQ, have affected the RQ? If, for example, the have no opinion option would not have been included as an option. Making the assumption that the respondents answering have no opinion would have answered neutral (neither agree nor disagree) if no such alternative would have been included in the questionnaire, this would have negatively affected all companies' RQ values. This is particularly interesting, considering Academic Work clearly having more positive RQ values than the other companies, but also having evidently higher degree of have no opinion answers. In order to see to which extent this would have affected the other findings, another analysis was made - namely to replace all have no opinion responses with the neutral neither agree nor disagree. These findings are presented below in figure 9.
Reputation Quotient when "no opinions" = 3
49,00 48,50 48,00 47,50 47,00 46,50 45,74 46,00 45,50 45,00 44,50 44,00 Adecco Manpower Poolia 47,06 46,68
48,52
Altered RQ
Academic Work
Figure 9 - The altered Reputation Quotient.
As can be seen, overall all findings are rather similar as those of the original total RQ (figure 4). All companies have, however, been affected negatively in relation to the total RQ with some variation in effect. Academic Work still deviates clearly as more positively
34
perceived from the other companies and the negative impact from the additional neutral values was not very large, when comparing the "new" values of Academic Work with the "new" values of the other companies. One difference however, from the total RQ, is that Manpower has changed relative rank position with Poolia, and is now ranked as second. The two companies have, however, in both the total RQ and the altered RQ, similar values. This alternative analysis validates the findings in the total RQ, and thus the choice to include a have no opinion response option in the questionnaire. Further, the possibility to separately be able to analyse the cases that marked the have no opinion response have shed light on to which degree the respondents have opinions at all regarding the companies. 4.1.4
60,0%
The relation between corporate reputation and employment
Comp Manpower Adecco AW Poolia
50,0%
40,0%
Percent
30,0%
20,0%
10,0%
0,0% Yes No
Work for company?
Figure 10 - Willingness to work for the companies.
Figure 10 displays the findings from the question "Would you consider working for company X after graduation?" - Q18, in the questionnaire. 50 percent of the respondents of both Adecco and Academic Work would consider working for the companies after graduation. Roughly a 60 percent share of the respondents of Poolia and Manpower is positive towards working for the companies. Manpower and Poolia have been assigned similar values throughout the study, and therefore the findings regarding these companies are not surprising. Of note is that the Adecco respondents are as positive towards working for Adecco as is the Academic Work respondents towards working for Academic Work. If comparing these findings to the previously stated results, it might look surprising that the respondents are not more positive towards working for Academic Work, it clearly being the most positively perceived company among the four and furthest away from Adecco. The implications that can be drawn from this are that either there is a measurement error in either of the measure of RQ or in Q18. There can be some missing factors in corporate reputation that have not been detected in this study that could give another result than the one just presented. If that is
35
not the case, the results might imply that corporate reputation is not directly linked to whether a person is willing to work for the company or not. If neither is true, then one could argue that corporate reputation is a bad measurement for perceptions for this stakeholder group, and other incentives should be measured instead. That, however, is not part of the purpose of this study to examine. Further, to see what single factor the respondents ranked as most important when determining how they perceived a company, one question (Q1) was included into the questionnaire. This question was not specific for the staffing and recruitment industry. However, the results are vivid; reputation is the most important factor when perceiving a company. The findings can be seen in Appendix 1. This finding can also help identify the relation between the dimensions that build up the corporate reputation. However, one can not ascribe it a weighting since the respondents were not asked to weight the questions just pick the alternative that they found most important. If that would have been possible it would have been in line with the free-descriptive method presented by Bromley (2002). These findings do not have to indicate that reputation is important when looking for a job, but rather to give an overall evaluation of a company. Therefore, companies will have to take this into consideration when targeting the investigated stakeholder group. However, since reputation scored highest in question one it gives more strength for corporate reputation being an important field of investigation for companies that want to find out how they are perceived by their stakeholders. To add an initial aspect to the measurement of corporate reputation, Bernstein's model of reputation creation has been used for investigating the source of the perceptions that the respondents hold towards the companies, (Q16). Bernstein's model is based on the definition by Bevis that the reputation creation is "the net result of the interaction of all the experiences, impressions, beliefs, feelings and knowledge that people have about a company"(cited in Davies et. al., 2003 p. 63). Bernstein elaborates further on how these impressions are created and the factors are presented in (figure 3) on how reputation is created. These impressions have been altered to get a better fit to the respondents. Bernstein also presents three different "filters" that the impressions have to pass through before becoming a perception: the filters of country of origin, industry, and company. These filters most likely have an impact on the perceptions but are not a subject of investigation in this study and will be disregarded. Instead, only the source of the perception is investigated. As can be seen in Appendix 2, the most important factor in this study turned out to be advertising, which shows to be the highest occurring factor in three of the four cases. Slightly below 40 percent of the respondents answering surveys for Academic Work and Poolia picked advertising as the most important factor for perception creation. Manpower and Adecco have about 25 percent and 15 percent respectively on the same question. For Adecco, word of mouth has been the most important for the perception creation. Adecco also has a high level of other answers than the ones that were suggested. However, these answers follow a clear pattern: the respondents had no idea of what type of business Adecco is running. The finding that a high level of respondents picked advertising as the most important factor can indicate that the companies marketing efforts are the most important factor when it comes to how corporate reputation is created. However, word of mouth is an important factor and can also contain a lot of information that is unknown in this study. It can be both negative and positive information that have been transferred from person to person. In word of mouth there is also room for peoples' interpretation of, for example, marketing that has been transferred to the respondents sampled in this study by others.
36
Conclusion
In the following chapter the authors present conclusions that have been drawn from the analysis. These conclusions are related back to the purpose. This study will investigate the corporate reputation of four staffing and recruitment companies, as representatives of the industry, from graduatingstudents'perspective. The purpose of this study has been fulfilled by measuring the corporate reputation using the methodology of the Reputation Quotient. The Reputation Quotient (RQ) is a scientifically grounded method for measuring corporate reputation, which validates the findings. To measure such a vaguely defined concept as corporate reputation brings not only definition ambiguity but also measurement concerns. Both the definition and measurement is something that have been widely debated in academia - a debate that has influenced and supported the altering of the method in order to fit the investigated population and subject. A quantitative survey has been designed to gather the empirical data from 125 graduating students, not in order to generalise to a greater degree but nevertheless to give a good estimation of the perceptions of the population of interest. Four corporate reputations have been measured for four staffing and recruitment companies. SPSS have been used to analyse the data and descriptive statistics of various kinds have presented the findings in this report. Analyses of the RQ in relation to characteristics of the samples have been performed in order to reinforce the findings and to put them in various perspectives. In addition, a disaggregated analysis of the RQ components has confirmed the total RQ findings. Additional investigations on how the students' perceptions have been created and whether they would consider working for the companies have been performed, putting the RQ findings in context. The reliability testing of the scale, which lies behind the RQ results, confirmed that the scale was also reliable after alteration of the original RQ. The findings entail that graduating students ascribe Academic Work the most positive corporate reputation out of the four. Poolia is ranked second with a small marginal to Manpower which is ranked third. Adecco is the company that received the least positive RQ value when measuring how graduating students perceive these four staffing and recruitment companies. No evidence has supported that a more positive reputation implies an increased willingness to work for the company, at least in the context of staffing and recruitment companies. Due to what is measured, namely perceptions, it is rather bold to make statements of the population as a whole. Instead, the findings can be seen as estimates of the companies reputations through the eyes of the graduating students at Jönköping University in the fields of business and engineering. However, it can be argued that one could make some claims on the population since the sample size is large relative to the population. Closely linked to the generalisability issue is validity; to have a high level of external validity, the findings should be applicable to another setting. For this study it has not been the aim to apply the findings to other populations than to the one investigated. This limitation has been imposed due to the many external factors present that can affect the creation of perceptions. These are not all incorporated to the corporate reputation measurement and, therefore, are even harder to detect since they fall outside the frame of the study. Further,
37
validity is also linked to the sampling method. For this study the there is a sampling bias present due to that only the population attending classes is sampled, leading to an underrepresentation of students not attending classes. Also, there is an uneven gender and school allocation that affects the validity negatively. To ensure that the measurement conducted is reliable, four Cronbach Alpha values have been calculated and shown positive results for the measurement. Further, since the nonresponse is disregarded in this calculation, a separate analysis has been conducted to enlighten the implications of this. As it turned out, there were two items that showed a high level of non-response, indicating that these statements need to be revised to get a better fit to the population investigated if the study was to be re-tested.
38
5
Discussion
The concluding chapter presents a discussion on the subject with the authors' personal thoughts and opinions. Further, the authors highlight the problems with the method and thus help the reader to understand the precautionary measures taken to help contribute to thoughtful and justifiable conclusions. There is also a part that reflects upon the thesis's limitations, as well as the authors' recommendations for further studies and readings. There have been many attempts by researchers to find a valid measurement of corporate reputation. Some of the measures have been adapted and used more than others. As discussed in earlier sections, the Fortune MAC is probably the most widely known, whereas the Harris-Fombrun Reputation Quotient might be seen as the most accepted. The RQ is, however, far from perfect and Wartick (2002), among others, have appointed critique to the measure. These issues have been dealt with in the initial stages of the study process. When looking at the findings of this study, they can seem rather intuitive. However, it would be hard to see the value of a measured RQ when not seen in relation to competing firms. This is the problem when quantifying feelings and perceptions as the depth behind why the respondents perceive the firms as they do disappear from the results. The results received can, on the other hand, easily be ranked and compared to other companies and thus one can decide which company has the best perceived corporate reputation. One could say that measures of corporate reputation yield valued outcomes only when put in relation with other firms within the same industry. It can therefore be argued that the only ones that should be interested in measuring corporate reputation are firms that are interested in how they are perceived in relation to its' competitors. This opinion is already included in some definitions but put in relation to the effort spent by researchers trying to find a universal measurement of corporate reputation; one can question where the academic value can be found in that type of research when, in practice, corporate reputation is only valuable to rank a firm in relation to their competitors. Would it not be of more interest for companies and academia, if more effort was spent on what actually can be made to improve corporate reputation? This does, however, require a clearly stated definition which can receive the gathered attention of academia. This brings us to the discussion on the definition of corporate reputation. The researchers are as divided in the question as in the issue of how to measure the concept. It has been discussed by several researchers that the assessment of corporate reputation includes the perceptions of all the firm's constitutes, indicating an aggregated perspective. Others have appointed harsh critique to this view arguing that a good assessment can only be collected if a disaggregated perspective is taken. There are also separate opinions on whether corporate reputation is a concept measured over time or as a snap shot of the reality. This ambiguous stance from academia on the definition of corporate reputation makes it difficult to know from where to take off for measuring the concept. To add credibility to the field overall, a canonical working definition needs to be established. This would enhance the possibility to compare the results from two studies. As it is today, one can not be certain if the researchers are measuring the same concept or two different concepts with the same name. Another concern that could have given the study another level of credibility would have been if weightings were included. As Bromley (2002) suggests, a free-descriptive method would have given the respondents the possibility to weight the dimensions of the RQ measure after their choice of interest. In this way further analysis could have been made on
39
what items and dimensions are considered essential for how to perceive a company. This could have given the study a different outcome. For example, if most of the factors where Poolia was ranked higher than Manpower were assigned low importance, and all the questions where Manpower was ranked higher than Poolia were assigned high importance, it would have given Manpower a boost in the overall perception of the company. 5.1.1 Limitations
There are many flaws to the concept and measurement of corporate reputation. To ensure that this has been regarded, it is important to ensure that a solid psychometric framework is used. As Bromley suggests, an ideal process to attain a good scale is to gather a fair amount of attributes that are relatively independent from each other but still correlated to the overall score or criterion. As this study is based on measuring the RQ, even though it is argued to fit all stakeholder groups, one should closely think of the purpose of the research and what, and whom, is about to be investigated. In this study the scale was altered to fit the investigated population to a larger extent than the original items in the RQ measurement. This way the measurement validity is increased and the scale actually measures what it is supposed to measure, namely the perceptions of graduating students at Jönköping University in the fields of business and engineering. In this way a better psychometric framework is assessed, which suites both subject and population sampled. There is however, a problem connected to the amount of items used for the investigation. Some researchers imply that the amount of respondents have to be up to 20 times the items investigated in order to have a sample big enough to draw conclusions. Thus, this number tends to vary from researcher to researcher, and some even claim that five respondents per item are sufficient. As convenience sampling was used for this sample, the sample was based on how many responses that could be collected, which turned out to be 125. A suggested analysis of dimensions, such as the ones in the RQ measure, could have been properly performed with factor analysis. However, as Tabachnick & Fidell (2001) suggests, for factor analysis 300 responses is needed, and no less than 150 (cited in Pallant, 2005). Therefore, the usage of factor analysis was not made possible, instead a disaggregated assessment was made to analyse the dimensions of the RQ in order to get a multilateral analysis of the RQ. Some of the issues that remain are the underrepresentation of students that are not present at lectures where the population was sampled and that the gender distribution for the sample is not representative for the population as a whole. If addressed, these issues could have given the study another overall result. Further, the allocation of students between JIBS and JTH has also been disregarded. This could be of great importance when wanting to generalise for the population as the distribution should be representative for the population. The issue of gender and school representation has been disregarded in this study due to the constraints of the sampling method, implying that the sample is not fully representative for the population. To ensure that the distribution is equal, one should use another sampling method and approach the students in another way. Some researchers would also argue that random sampling is the only credible way of sampling if generalisation is to be made on the population. This awakens another issue connected both to reliability and validity; if these results are applicable for another setting than the one just examined. What is measured in the end is perceptions which are individual and very dependent on the individual experiences of the respondents. One might be able to see the same patterns if the study was repeated with other respondents or another time. As Balmer & Stephen (2003) suggests reputations tend
40
to be stable but can change significantly due to external or internal events such as industry disasters or ethical lapse by the company. This indicates that this study could yield the same result if conducted within a near future, unless there is some striking change in the staffing and recruitment industry. One can also argue that the findings then would give about the same results and therefore imply that the findings also could be generalised for the graduating students at Jönköping University in the fields of engineering and business. The generalisation can, however, not be made for a setting outside Jönköping since the companies investigated can profile themselves differently in another city. 5.1.2 Implications
The results of this study imply that Academic Work is perceived in the best way in terms of corporate reputation and Adecco as the worst, of the four staffing and recruitment companies. However, Academic Work was also the company with the highest rate of have no opinion answers implying that the company may in fact be the least recognised company and thus has to improve the awareness of the company. However, a high rate of have no opinion was also recurrent for the other companies, which might imply that all the companies should appoint a lot of marketing effort to create high awareness among this important stakeholder group investigated. This statement is also backed up by the findings, showing that the most important factor for creating perceptions was advertising. This can also imply that advertising also is the most effective way to reach out to this stakeholder group. However, word of mouth, scored as the second most important factor, implying that one person's experience have great potential to travel far and wide and reach many ears. Therefore the companies should be aware that all communicated messages from the firm are in line with what is desired. Thus, employees are crucial, since their experiences can also be told to the investigated stakeholder group that has the potential to become employees. However, it is also important to state that in this study there was found no relation between whether one would like to work for a company and a good corporate reputation measured in RQ. Therefore, the companies should also look at other means to investigate how to attract this stakeholder group to become an employee. As for the graduating students, this study might not be able to provide such good information about what company to work for since corporate reputation is mostly a measure for companies. However, the investigated population as well as others who have interest in the staffing and recruitment companies can be informed of how these four companies are perceived in relation to each other by graduating students. 5.1.3 Further Research
During the process of this study a lot of flaws with the concept and measurement of corporate reputation have been detected. The first suggestion for further research is to thoroughly investigate corporate reputation as a concept to come up with a valued definition that could eliminate the root to ambiguity in the research field as it is today. This type of study will also help clarify whether corporate reputation is a concept that is measured over time or if it is a snap-shot of reality. In order to give a multilateral picture to the measurement of RQ, several different aspects have been taken into consideration for this study. However, if the study were to be repeated, it is suggested to apply weightings to the items of the RQ measurement or any scale of measurement used, as to enhance credibility to the measurement of corporate reputation. This thinking has been inspired by Bromleys' (2002) ideas on free-descriptive method.
41
References
Abratt, R. (1989). A new approach to corporate image management process. Journal of Marketing Managemant, 5, 63-76. Almega. (2008-02-14). Bemanningsbranschen uppåt för tredje året i rad. Retrieved 2008-02-14, from http://www.newsdesk.se/pressroom/almega/pressrelease/view/ bemanningsfoeretagen-uppaat-foer-tredje-aaret-i-rad-195869 Almega. (2008-01-29). Arbetssökande föredrar bemanningsföretag och jobbsajter före Ams. Retrieved 2008-02-05, from http://www.newsdesk.se/pressroom/almega/pressrelease/view/arbetssoekandefoeredrar-bemanningsfoeretag-och-jobbsajter-foere-ams-192609 Armstrong, M. (2006). Handbook of Human Resource Management Practice (10th ed.). London: Kogan Page Limited. Balmer, M. T. J., Stephen, A. (2003). Revealing the corporation: perspectives on identity, image, reputation, corporate branding, and corporate-level marketing. London: Routlegde Bemanningsföretagen. (2005-10-14). Bemanningsbranschen. Retrieved 2008-02-05, from http://www.almega.se/Files/ALMEGA/Caradoc_Members/PM/Kartläggninge_Be manningsbransc_A.DOC Bolander, L., & Nord, A. (1999). Arbetssituationen för anställda på bemanningsföretag eller hur mår "Nisse" egentligen? Master thesis in Business Administration. Linköping: Linköping University. Brace, I. (2004). Questionnaire Design. London: Kogan Page Ltd. Bromander, L. (2007). Uthyrning av arbetskraft - Ett rättssocialogiskt perspektiv på en flexibilitetsstrategi. Retrieved 2008-02-04, from https://gupea.ub.gu.se/dspace/bitstream/2077/7334/1/Lena%20Bromander.pdf Bromley, D. (2002). Comparing Corporate Reputations: League Tables, Quotients, Benchmarks or Case Studies? Corporate Reputation Review, 5(1), 35-50. Buglear, J. (2005). Quantitative Methods for Business - The A-Z of QM. Oxford: Elsevier Butterworth-Heinemann. Caruana, A. (1997). Corporate Reputation: concept and measurement. Journal of product & brand management, 6(2), 109-118. Collins, J., Hussey, R. (2n003). Business Research - A practical guide for undergraduate and postgraduate students (2d ed.). Hampshire: Palgrave Macmillan. Davies, G., Chun, R., Silva, D. R. V., Roper, S. (2003). Corporate Reputation and Competitiveness. Trowbridge, Wiltshire: The Cromwell Press. Dutton, J. E., Dukerich, J. M., Harquail, C. V. (1994). Organisation images and member identification. Administrative Science Quarterly, 39, 239-263. Englund, S., Nordström, L., Ohlsson, A. (2000). "Nisse" på besök i verkligheten. Bachelor thesis in Business Administration. Jönköping: Jönköping International Business School.
42
Fink, A. (1995). The Survey Handbook. London: SAGE Publications. Fombrun, C. J. (1996). Reputation: Realizing Value from the Corporate Image. Boston: Harvard Business School Press. Fombrun, C. J., Gardberg, N. A., Sever, J. M. (2000). The Reputation Quotient: A multistakeholder measure of corporate reputation. The Journal of Brand Management, 7(4), 241256. Fowler, J. F. (2002). Survey Research Methods (3rd ed.). Thousand Oaks, CA: SAGE Publications. Ghauri,rdP., Grønhaug, K. (2005). Research Methods in Business Studies. A Practical Guide. (3 ed.). Harlow: Pearson Educated Limited. Gomm, R. (2004). Social Research Methodology a critical introduction (1st ed.). Hampshire: Palgrave MacMillan Gotsi, M., & Wilson, A. M. (2001). Corporate reputation: Seeking a definition. Corporate Communications: An International Journal, 6(1), 24-30. Groenland, E. A. G. (2002). Qualitative Research to Validate the RQ-Dimensions. Corporate Reputation Review, 4(4), 308-315. Hillenbrand, C., & Money, K. (2006). Using Reputation measurement to create value: An analysis and integration of existing measures. Journal of General Management, 32(1), 1-12. Ind, N. (1997). The Corporate Brand. London: The MacMillan Press. Jacobsen, D. I. (2002). Vad, hur och varför? Om metodval i företagsekonomi och andra samhällsvetenskapliga ämnen. Lund: Studentlitteratur Johansson, M. (2006). Bemanningsbranchen - en kvantitativ studie av dess utveckling I Sverige. Master Thesis in Economics and Statistics. Växsjö: Växsjö Universitet. Litwin, S. M. (1995). How to Measure Survey Reliability and Validity. Thousand Oaks, CA: SAGE Publications. Moingeon, B., Soenen, G. (2002). Corporate and Organizational Identity: Integrating Strategy, Marketing, Communication and Organizational Perspectives. London: Routledge. Mahone, J. F., Wartick, S. L. (2003). Dealing with Stakeholders: How Reputation, Credibilty and Framing Influence the Game. Corporate Reputation Review, 6(1), 19-35. Olofsdotter, G. (2004). Bemanningsföretag - en möjlighet för vem? In Wolvén, L-E., Ekstedt, E. (Eds.), Företagande och gemenskap - små företags kultur, samverkan och konkurrens. (p. 149-168). Stockholm: Elanders Gotab. Pallant, J. (2005). SPSS Survival manual. (2nd ed.). New York: McGraw-Hill Education. Ruane, J. M (2006). A och O i samhällsvetenskaplig forskning. Lund: Studentlitteratur. Saunders, M. N. K., Thornhill, A., Lewis, P., McMillan, K., Weyers, J. (2007). Research Methods for Business Students. London: Pearson Higher Education.
43
Strbac, S. (2006). The staffing process for product development projects: A case study at 3P, Volvo AB. Dissertation in International Project Management. Göteborg: Chalmers University of Technology. Wartick, L. S. (2002). Measuring Corporate Reputation, Definition and Data. Business & Society, 41(4), 371-392. Zikmund, W. G. (2000). Business Research Methods (6th ed.). Orlando: Hartcourt Inc.
44
Appendices
Appendix 1 - Most important factor when looking at a company.
40,0%
30,0%
Percent
20,0%
10,0%
0,0%
Gut feel ng i Products/service Vision/leadership Workplace env s Fin results Social resp Carismatic repr Reputation Other
Most imortant factor Most imortant factor Frequency Valid Gut feeling Products/services Vision/leadership Workplace env Fin results Social resp Carismatic repr Reputation Other Total Missing Total System 14 27 11 6 6 2 3 43 4 116 9 125 Percent 11,2 21,6 8 ,8 4 ,8 4 ,8 1 ,6 2 ,4 34,4 3 ,2 92,8 7 ,2 100,0 Valid Percent 12,1 23,3 9 ,5 5 ,2 5 ,2 1 ,7 2 ,6 37,1 3 ,4 100,0 Cumulative Percent 12,1 35,3 44,8 50,0 55,2 56,9 59,5 96,6 100,0
45
Appendix 2 - Major Source in creating perceptions? The following graphs are the summarised answers from question 16. Not all of the response options have been picked and therefore do not show in the graphs. Academic Work
40,0%
30,0%
Percen t
20,0%
10,0%
0,0% Personal contact Media/internet Word of mouth Advertisment Other
Source
Manpower
25,0%
20,0%
Percent
15,0%
10,0%
5,0%
0,0% Personal contact Media/internet Word of mouth Advertisment Experience Other
Source
46
Adecco
40,0%
30,0%
Percen t
20,0%
10,0%
0,0% Personal contact Media/internet Word of mouth Advertisment Mail/email Other
Source
Poolia
40, 0%
30, 0%
Percen t
20, 0%
10, 0%
0,0% Personal contact Media/internet Word of mouth Advertisment Experience Other
Source
47
Appendix 3 - Reliability testing statistics Academic Work
Reliability Statistics Cronbach's Alpha ,841
N of Items 14 Item-Total Statistics Scale Variance if Item Deleted 22,750 24,444 28,278 27,028 29,750 27,528 29,028 30,028 28,250 26,194 27,000 30,694 29,861 29,500 Corrected Item-Total Correlation ,844 ,788 ,740 ,898 ,333 ,568 ,188 ,356 ,522 ,725 ,455 ,180 ,141 ,378 Cronbach's Alpha if Item Deleted ,799 ,806 ,823 ,813 ,839 ,825 ,857 ,838 ,829 ,815 ,834 ,845 ,857 ,836
Scale Mean if Item Deleted Good feeling Admire Trust Good company Organisation Offerings Att employer Good merit Marketing Future Competition Opportunities Society Job market 49,67 50,22 49,44 49,56 49,67 49,44 49,44 49,44 49,33 49,22 49,67 49,22 48,89 48,67
48
Manpower
Reliability Statistics Cronbach's Alpha ,893
N of Items 14 Item-Total Statistics Scale Variance if Item Deleted 47,829 48,981 45,400 46,114 49,067 51,667 46,695 52,114 49,524 50,543 52,410 51,410 48,952 49,571 Corrected Item-Total Correlation ,640 ,538 ,717 ,662 ,626 ,373 ,614 ,460 ,504 ,642 ,420 ,533 ,715 ,728 Cronbach's Alpha if Item Deleted ,882 ,887 ,878 ,882 ,883 ,894 ,884 ,890 ,889 ,884 ,891 ,887 ,880 ,881
Scale Mean if Item Deleted Good feeling Admire Trust Good company Organisation Offerings Att employer Good merit Marketing Future Competition Opportunities Society Job market 44,40 45,47 44,60 44,40 44,27 44,67 45,13 44,40 44,67 44,60 44,53 44,53 44,33 44,00
49
Adecco
Reliability Statistics Cronbach's Alpha ,665
N of Items 14 Item Statistics Item-Total Statistics Scale Variance if Item Deleted 15,302 16,681 17,764 16,841 15,610 16,264 17,187 14,066 22,555 15,104 16,335 17,940 16,335 15,786 Corrected Item-Total Correlation ,641 ,282 ,089 ,315 ,554 ,393 ,228 ,549 -,528 ,556 ,472 ,107 ,330 ,469 Cronbach's Alpha if Item Deleted ,601 ,649 ,677 ,645 ,612 ,633 ,656 ,597 ,775 ,605 ,627 ,670 ,641 ,621
Scale Mean if Item Deleted Good feeling Admire Trust Good company Organisation Offerings Att employer Good merit Marketing Future Competition Opportunities Society Job market 42,07 42,71 42,07 42,07 41,93 42,43 42,57 42,29 42,64 41,79 42,21 42,36 41,79 41,64
50
Poolia
Reliability Statistics
Cronbach's Alpha ,955
N of Items 14 Item-Total Statistics Scale Variance if Item Deleted 87,344 90,500 89,956 89,344 87,511 90,889 83,733 85,733 90,011 82,844 84,667 88,278 90,100 76,989 Corrected Item-Total Correlation ,853 ,718 ,718 ,903 ,749 ,648 ,702 ,792 ,757 ,973 ,814 ,896 ,550 ,880 Cronbach's Alpha if Item Deleted ,950 ,953 ,953 ,950 ,952 ,954 ,955 ,951 ,952 ,946 ,951 ,950 ,957 ,951
Scale Mean if Item Deleted Good feeling Admire Trust Good company Organisation Offerings Att employer Good merit Marketing Future Competition Opportunities Society Job market 43,30 44,50 43,20 43,30 43,20 44,00 44,20 43,80 43,70 43,80 44,00 43,50 43,90 43,90
51
Appendix 4 - Mean values and std. deviation
Academic Work
Descriptive Statistics N Good feeling Admire Trust Good company Organisation Offerings Att employer Good merit Marketing Future Competition Opportunities Society Job market Valid N (listwise) 26 24 23 22 17 23 20 21 17 19 16 15 23 23 9 Mean 3,50 2,92 3,74 3,73 3,76 3,78 3,40 3,90 3,47 4,05 3,44 3,80 4,04 4,26 Std. Deviation ,812 ,830 ,752 ,631 ,437 ,850 1,046 ,436 ,874 ,524 ,814 ,676 ,976 ,619
Manpower
Descriptive Statistics N Good feeling Admire Trust Good company Organisation Offerings Att employer Good merit Marketing Future Competition Opportunities Society Job market Valid N (listwise) 29 28 29 27 25 25 27 28 25 25 22 24 31 31 15 Mean 3,55 2,46 3,52 3,52 3,64 3,44 2,85 3,50 3,40 3,64 3,27 3,63 3,55 4,00 Std. Deviation 1,021 ,793 1,056 1,189 ,757 ,768 1,099 ,745 ,866 ,638 ,767 ,770 ,995 ,683
52
Adecco
Descriptive Statistics N Good feeling Admire Trust Good company Organisation Offerings Att employer Good merit Marketing Future Competition Opportunities Society Job market Valid N (listwise) 26 23 25 27 21 23 26 26 20 22 19 21 27 27 14 Mean 3,38 2,61 3,68 3,30 3,52 3,04 2,88 3,35 2,95 3,73 3,00 3,24 3,52 3,70 Std. Deviation ,852 ,783 ,900 ,823 ,602 ,767 ,766 ,797 ,945 ,703 ,667 ,700 ,893 ,869
Poolia
Descriptive Statistics N Good feeling Admire Trust Good company Organisation Offerings Att employer Good merit Marketing Future Competition Opportunities Society Job market Valid N (listwise) 28 26 25 27 22 23 25 26 23 22 21 20 25 24 10 Mean 3,64 2,62 3,80 3,63 3,64 3,09 2,88 3,38 3,52 3,55 3,10 3,70 3,64 3,88 Std. Deviation ,780 ,804 ,816 ,565 ,727 ,733 ,881 ,941 ,730 ,858 ,831 ,571 ,860 1,154
53
Appendix 5 - Sample characteristics
Academic Work
School Cumulative Percent 45,2 100,0
Frequency Valid JTH JIBS Total 14 17 31
Percent 45,2 54,8 100,0 Sex
Valid Percent 45,2 54,8 100,0
Frequency Valid Female Male Total 10 21 31
Percent 32,3 67,7 100,0
Valid Percent 32,3 67,7 100,0
Cumulative Percent 32,3 100,0
Descriptive Statistics N Age Valid N (listwise) 30 30 Minimum 21 Maximum 31 Mean 23,77
Manpower
School Cumulative Percent 43,8 100,0
Valid
JTH JIBS Total
Frequency 14 18 32
Percent 43,8 56,3 100,0 Sex
Valid Percent 43,8 56,3 100,0
Frequency Valid Female Male Total 11 21 32
Percent 34,4 65,6 100,0
Valid Percent 34,4 65,6 100,0
Cumulative Percent 34,4 100,0
Descriptive Statistics N Age Valid N (listwise) 31 31 Minimum 20 Maximum 42 Mean 25,06
54
Adecco
School Cumulative Percent 38,7 100,0
Frequency Valid JTH JIBS Total 12 19 31
Percent 38,7 61,3 100,0
Valid Percent 38,7 61,3 100,0
Sex Cumulative Percent 48,4 100,0
Frequency Valid Female Male Total 15 16 31
Percent 48,4 51,6 100,0
Valid Percent 48,4 51,6 100,0
Descriptive Statistics N Age Valid N (listwise) 31 31 Minimum 20 Maximum 28 Mean 23,61
Poolia
School Cumulative Percent 51,6 100,0
Valid
JTH JIBS Total
Frequency 16 15 31
Percent 51,6 48,4 100,0 Sex
Valid Percent 51,6 48,4 100,0
Frequency Valid Female Male Total 15 16 31
Percent 48,4 51,6 100,0
Valid Percent 48,4 51,6 100,0
Cumulative Percent 48,4 100,0
Descriptive Statistics N Age Valid N (listwise) 30 30 Minimum 20 Maximum 28 Mean 23,37
55
Appendix 6 - Questionnaires Below follows the questionnaires used for collecting empirical data. The data gathering was done by handing out 31 surveys to three of the staffing and recruitment companies and 32 surveys to the fourth (see section 3.3.2). Since the surveys were identical apart from the different company names; Adecco, Poolia, Manpower, Academic Work the authors have decided to simply include templates, and use the company name "X" for the English survey and the company name "Y" for the Swedish survey.
56
A study on the corporate reputation of Company "X"
This study will be used as empirical data for a bachelor thesis concerning corporate reputation. All respondents will remain anonymous. Thank you very much for participating!
For all following questions, please mark one alternative that suites your perception best.
Ex:
1. Which factor do you consider the most important in determining how you perceive a company? (Choose one factor, not specific for Company "X") "Gut" feeling Financial performance National origin Products and services Social responsibility Reputation Vision and leadership Charismatic representatives Other: Workplace environment
To what extent do you agree with the following statements? Strongly Disagree Disagree Neither agree nor disagree Agree Strongly Have no opinion
2. I have a good feeling about Company "X" 3. I admire Company "X" 4. I do not trust Company "X"
5. Company "X" looks like a good company to work for 6. Company "X" seems to have a well functioning organisation
7. Company "X" offers high quality work opportunities 8. Company "X" is an attractive employer 9. To work for Company "X" will yield a valuable experience 10. Company "X" does not use innovative marketing strategies
11. Company "X" has strong prospects of becoming more successful in the future 12. Company "X" tends to outperform its competitors 13. Company "X" does not take advantage of market opportunities
Please turn!
57
Stongly Disagree
Disagree
Neither agree nor disagree
Agree
Strongly Agree
Have no opinion
14. Company "X" has an important role in the society 15. Company "X" is important for the job market
16. What has been the major source in creating your above stated perceptions of Company "X"? (Choose one factor) Personal contact Direct mail, e-mail Media comments, internet Prior experience Word of mouth Other: Advertising
Are you currently enrolled in a program at:
J TH
JIBS
Not a program student
After graduation, where do you intend to look for a job? (feel free to choose more than one)
Sweden
Nordic countries (other then Sweden)
Europe (other than the Nordic countries)
Rest of the world
When do you expect to apply for a job?
0 - 6 months
6 - 12 months
12 - 24 months
Already have a job (fulltime)
Would you consider working for Company "X" after graduation?
Yes
No
Female
Male
Age:
Thank you very much for your participation and help!
Please state your e-mail address if you want to be informed of the results of this study:
58
En studie om företaget "Ys" anseende.
Denna studie kommer att användas som empirisk data i en kandidatuppsats om företags anseende. Alla svarande kommer förbli anonyma. Stort tack för er medverkan!
För följande frågor, välj ett alternativ som överensstämmer bäst med din uppfattning. Ex:
1. Vilken faktor anser du är viktigast när det gäller att avgöra hur du uppfattar ett företag? (välj en faktor, ej specifikt för företag "Y") Emotionell Utstrålning Finansiella Resultat Nationellt ursprung Produkter och Tjänster Socialt Ansvar Rykte Vision och Ledarskap Karismatiska representanter Annat Arbetsplats Miljö
Till vilken utsträckning håller du med dessa påståenden? Håller inte alls med Håller inte med Håller varken Håller med eller delvis inte med Håller helt med Har ingen uppfattning
2. Jag har en god känsla kring företag "Y" 3. Jag beundrar företag "Y" 4. Jag litar inte på företag "Y" 5. Företag "Y" verkar vara ett bra företag att arbeta för 6. Företag "Y" verkar ha en välfungerande organisation
7. Företag "Y" erbjuder högkvalitativa jobberbjudanden 8. Företag "Y" är en attraktiv arbetsgivare 9. Att arbeta på företag "Y" ger en god merit 10. Företag "Y" använder inte innovativa marknadsföringsstrategier
11. Företag "Y" har starka utsikter för att bli mer framgångsrika i framtiden 12. Företag "Y" tenderar till att överträffa dess konkurrenter 13. Företag "Y" utnyttjar inte marknadsmöjligheterna
Var god vänd!
59
Håller inte alls med
Håller inte med
Håller varken Håller med eller delvis inte med
Håller helt med
Har ingen uppfattning
14. Företag "Y" fyller en viktig samhällsfunktion 15. Företag "Y" är betydande för arbetsmarknaden
16. Vad har varit den största källan till din ovan angivna uppfattning av företag "Y"? (Välj en)
Personlig kontakt
Mediauppgifter, Internet
Ryktesspridning (word of mouth)
Annonsering, reklam
Direkt post, e-post
Tidigare egna erfarenheter
Annat
Jag är för närvarande inskriven på ett program på:
J TH Ej programstudent
JIBS
Var kommer du att söka jobb efter avslutad utbildning? (välj ett eller flera alternativ)
Sverige Europa (ej Norden) Resterande världen
Norden (ej Sverige)
När räknar du med att ansöka om ett jobb?
0 - 6 mån 12 - 24 mån
6 - 12 mån Har redan jobb (heltid)
Skulle du kunna tänka dig att jobba för företag "Y" efter din examen?
Ja
Nej
Kvinna
Man
Ålder:
Om du önskar att ta del av resultatet av denna undersökning, skriv din e-mail adress:__________________________________
Stort Tack för er medverkan och hjälp!
60
doc_671167722.docx
Many businesses have public relations departments dedicated to managing their reputation. In addition, many public relations firms describe their expertise in terms of reputation management. The public relations industry is growing due to the demand for companies to build corporate credibility and hence reputation.[citation needed] Incidents which damage a company's reputation for honesty or safety may cause serious damage to finances.
A Case study on corporate reputation
Abstract:The purpose of this study is to measure and compare the corporate reputation of four staffing and recruitment companies, from the perspective of graduating students. Corporate reputation is a concept that has interested researchers for the last four decades, and still the concept incorporates a number of varied definitions, contexts and measurement methods (Gotsi &Wilson, 2001). Along with definition ambiguity comes also measurement concerns such as whether to aggregate or disaggregate several perspectives into one general index of a company's reputation. Most researchers agree that corporate reputation is a suitable concept when describing the perceptions that stakeholders have of a company, and that it can only be ascribed to one company. This study takes the perspective to investigate one external stakeholder group of staffing and recruitment companies - graduating students. The group is of interest because it can possibly become an internal stakeholder group - employees. For this study a quantitative method has been applied and an investigation has been made through a survey. Methodology about measuring corporate reputation is discussed, and lay ground for the empirical data collection. An altered version of the Harris-Fombrun Reputation Quotient (RQ) is used, in order to achieve better fit of the instrument and to attain higher reliability and validity. The purpose of this study has been fulfilled by measuring corporate reputation from the perspective of graduating students. Empirical data from 125 students has been collected, and four corporate reputations of staffing and recruitment companies have been measured. The findings are not valid for generalisability but enoughto give a good estimation of the population of interest. In addition, a disaggregated analysis of the RQ components has confirmed the findings. Using the methodology of the theoretically based Harris- Fombrun Reputation Quotient further validates the findings. Graduating students perceive Academic Work in the most positive way, in respect to the other companies.
ii
Kandidatuppsats i företagsekonomi Titel: Författare: Handledare: Datum: Ämnesord: Hur avgångsstudenter uppfattar bemannings- och rekryteringsindustrin - En studie om företagets anseende Christopher Broman, Robin Cabander, Emilia Karlsson Börje Boers 200805-28 Företagets anseende, Corporate Reputation, Reputation Quotient, League table, Bemanningsföretag, Rekrytering
Sammanfattning
Syfte: Syftet med denna studie är att mäta och jämföra anseendet av fyra bemannings- och rekryteringsföretag, utifrån avgångsstudenters perspektiv. Företagets anseende är ett teoretiskt koncept som har intresserat forskare de senaste årtiondena och fortfarande råder delade meningar om definitioner, sammanhang och mätmetoder rörande konceptet (Gotsi & Wilson, 2001). Med tvetydiga definitioner kommer även oklarhet gällande mätningsmetod, som till exempel om man ska summera flera intressenters åsikter till ett övergripande index eller inte. Forskare är eniga om att företagets anseende är ett passande koncept när man ska beskriva de åsikter intressenter har om ett företag. Denna studie undersöker en specifik grupp av bemannings- och rekryteringsföretags externa intressenter, nämligen avgångsstudenter. Denna grupp är intressant då de kan komma att bli interna intressenter, via anställning av ett företag. Denna studie har en kvantitativ inriktning och undersökningen har utförts med hjälp av en enkät. Metodiken i att mäta företags anseende genom att använda sig av league tabeller och pilotstudier diskuteras och ligger till grund för empiriinsamling. Användandet av en anpassad version av Harris-Fombrun Reputation Quotient (RQ) ger studien ökad validitet och reliabilitet. Syftet med denna studie har blivit uppnått genom att mäta företagens anseende ur avgångsstudenters perspektiv. Empirisk data från 125 studenter har samlats ihop och fyra bemannings- och rekryteringsföretags företagsanseenden har blivit uppmätta. Resultatet kan inte generaliseras för en större population, men den kan användas för att ge en bra indikation hur åsikterna för den specifika populationen är. Vidare har en uppdelad analys av komponenterna i RQ gjorts vilket har kunnat säkerställa resultaten i den totala RQ. Användandet av den teoretiskt grundade RQ ger ytterligare validitet åt resultaten. Avgångsstudenter har mest positiva åsikter om Academic Work i jämförelse med de andra undersökta företagen.
Bakgrund:
Metod:
Slutsats:
iii
Table of Contents
1
1.1 1.2 1.3
Introduction........................................................................... 1
Background ............................................................................................1 Problem discussion ................................................................................2 Purpose ..................................................................................................3
2
2.1 2.2 2.2.1 2.2.2
Frame of Reference .............................................................. 4
Corporate reputation in a broader context ..............................................4 Corporate reputation as a concept .........................................................5 The importance of corporate reputation..................................................7 Reputation creation ................................................................................7
3
3.1 3.2 3.2.1 3.2.2 3.3 3.3.1 3.3.2 3.4 3.5 3.5.1 3.5.2 3.5.3 3.6 3.7 3.8 3.9
Method................................................................................... 9
Research approach ................................................................................9 How to measure corporate reputation ..................................................10 League table rating...............................................................................10 The Reputation Quotient ......................................................................11 Sample selection ..................................................................................13 Company sampling...............................................................................15 Sample characteristics .........................................................................15 Psychometrics ......................................................................................15 Choice of data collection ......................................................................16 Questionnaire design............................................................................17 Pilot Study ............................................................................................19 Translation bias ....................................................................................20 Generalisability .....................................................................................21 Reliability ..............................................................................................21 Validity..................................................................................................22 Analysing data......................................................................................23
4
4.1 4.1.1 4.1.2 4.1.3 4.1.4
Results and Analysis.......................................................... 26
The Reputation Quotient ......................................................................26 The Reputation Quotient separated by gender and school ..................28 Disaggregated Reputation Quotient .....................................................31 Analysis of "have no opinions"..............................................................33 The relation between corporate reputation and employment................35
5 6
6.1.1 6.1.2 6.1.3
Conclusion .......................................................................... 37 Discussion........................................................................... 39
Limitations ............................................................................................40 Implications ..........................................................................................41 Further Research..................................................................................41
References ................................................................................... 42 Appendices .................................................................................. 45
Appendix 1 - Most important factor when looking at a company. ......................45 Appendix 2 - Major Source in creating perceptions?.........................................46 Appendix 3 - Reliability testing statistics...........................................................48 Appendix 4 - Mean values and std. deviation ...................................................52
iv
Appendix 5 - Sample characteristics ................................................................54 Appendix 6 - Questionnaires ............................................................................56
List of Figures
Figure 1 - From identity to reputation ................................................................. 5 Figure 2 - Defining corporate reputation............................................................. 6 Figure 3 - How reputation is created .................................................................. 8 Figure 4 - The total Reputation Quotient .......................................................... 26 Figure 5 - Gender separated Reputation Quotient. .......................................... 28 Figure 6 - School separated Reputation Quotient. ........................................... 29 Figure 7 - The disaggregated Reputation Quotient. ......................................... 31 Figure 8 - Graphical overview of non-respondents........................................... 33 Figure 9 - The altered Reputation Quotient. ..................................................... 34 Figure 10 - Willingness to work for the companies........................................... 35
List of Tables
Table 3-1 - The Reputation Quotient................................................................ 12 Table 3-2 - The altered Reputation Quotient (RQ) ........................................... 13 Table 4-1 - Company league table ................................................................... 27 Table 4-2 - Company standard deviation ......................................................... 27 Table 4-3 - Table of non-response ................................................................... 33
v
1
Introduction
This section introduces the reader to the background and the problem discussion. Furthermore, it introduces the concept of corporate reputation, and presents the purpose and the delimitations for this thesis.
Students are graduating from Universities all over Sweden every semester. There are many different and important decisions that have to be made concerning job opportunities. The staffing and recruitment industry are closely monitoring the students' attempts for getting the necessary working experience they so desperately need in order to compete in the job market. This bachelor thesis focuses upon one external stakeholder group to the staffing and recruitment companies - graduating students. Corporate reputation is a concept that has interested researchers for the last four decades (Gotsi & Wilson, 2001), still the concept incorporates a number of varied definitions, contexts and measurement methods. A corporate reputation can be utilised by a firm to improve its competitive situation and to preserve a market position when under pressure (Mahon & Wartick, 2003). Having a positive reputation has been linked to several beneficial outcomes such as; greater intention to buy a product, higher perceived product quality, attracting investors to a larger extent, and enhanced organisation identification among employees (Caruana, 1997). Premium prices for products, lower costs of capital, and labour are some of the more concrete positive effects that can evolve from having a good reputation (Fombrun, 1996). Theory regarding corporate reputation is something that mostly has been established to suit the perspective of a company, in order to find out in which way its stakeholders view the organisation. Fombrun (1996, p. 78) states the following definition, which has been more widely used than most (Wartick, 2002). "A corporate reputation is a perceptual representation of a company's past actions andfuture prospects that describes thefirm's overall appeal to all of its key constituents when compared with the leading rivals." Corporate reputation research has had trouble reaching an unambiguous definition of the concept, despite the increasing interest in writing about it. Along with definition ambiguity comes also measurement concerns. Two major issues where opinions differ is whether to aggregate several perspectives taken into one grand reputation, and whether to make a general index of a company's reputation based on all attributes investigated to be able to rate it against others. However, most researchers agree that corporate reputation is a suitable concept describing the perceptions that stakeholders have of a company and that a corporate reputation is something that can be ascribed to one company.
1.1
Background
The staffing and recruitment industry in Sweden can be traced back to the 1950s. The development has, however, been moderate up until the last 10-15 years. Since then, the industry has expanded enormously. Before 1992, the development had been constrained by legal regulations forbidding any kind of private recruitment mediating. The first reason for the deregulating was the increase in demand for temporary labour in the early 1990s. Deregulations have also been made in several stages since then, with the intention to improve the function of the labour market, which have lead to the recognition of the importance of the staffing and recruitment industry (Bemanningsföretagen, 2005).
1
The structure of the labour market has changed because of deregulation, but also due to other factors. The demand of hiring skilled personnel and consultants has increased and companies today are in greater need of getting hold of and removing employees quickly and easily. The change can partly be explained by vast technological developments in combination with the globalisation process and the more service-focused society in developed countries. Increased competition from abroad, along with a more short-term oriented consumption, has shortened production life cycles. Therefore companies have to be flexible in order to survive the changing circumstances. This flexibility can partly be attained by having a flexible labour force (Bromander, 2007) and a solution to this is to hire services of a staffing firm. More and more companies have also decided to outsource human resource functions to recruitment firms, to be able to take advantage of their more efficient recruitment skills and wide networks (Armstrong, 2006). The staffing and recruitment industry has an important function in the Swedish economy today. Over 55 000 people are employed in the industry, which is about 0.9 percent of the labour force. The industry showed an increase of 18 percent in turnover in 2007 as compared to 2006 and it has nearly doubled in three years time, showing a clear expansionary trend (Almega, 2008). The rest of the EU has an average of 2.5-3 percent of the total employment in the industry, as compared to Sweden's 0.9 percent. In addition, the interest organisation of the recruitment and staffing companies (Bemanningsföretagen) has an expansionary viewpoint, with the goal of having 3 percent of the total employment in the industry (Bemanningsföretagen, 2005). The above stated arguments are all indications that the industry will continue to grow. There are four main service areas within the industry:
• • • •
Staffing or consulting services; renting out temporary personnel to a client firm Recruitment services include complete or partial outsourcing of the recruitment process. Contracting include complete or partial outsourcing of a functional unit within the client firm. Outplacement is a strategic service that includes consultation regarding changes in labour structure in the client firm.
The staffing services accounts for 90 percent and the recruitment services for 6 percent of the turnover in the industry. The remaining two services are relatively unexploited (Bemanningsföretagen, 2005). More and more companies are outsourcing the recruitment function to the staffing companies, since it is a costly process. The staffing companies have also become very competent in the matching process of employees and clients, due to the specialisation in this area and the close cooperation with their client firms. Therefore, the degree of recruitment services is increasing. Staffing and recruitment services are also often combined, providing clients with a type of trial employment for possible future employees (Johansson, 2006).
1.2
Problem discussion
It is not hard to make the assumption that a positive corporate reputation also has a positive effect on the ability to hire or recruit new skilled employees, which is one of the main challenges for recruitment and staffing firms. Following the rapid development of the staffing and recruitment industry, studies have been made to explain its characteristics.
2
Marketing research has been performed by, for example, Bemanningsföretagen, to study the public's opinion on staffing and recruitment companies. These, however, are not fully available to the public and lack a clear conceptual framework. Much of the research made on the Swedish market has focused on the staffing services with various areas of interest and perspectives taken. Some have taken the perspective of the consultants (Bolander &Nord, 1999; Olofsdotter, 2004), working for the staffing companies and explored how they experience their situation and the positive and negative factors belonging to it. Other research has taken the perspective of the customer companies and explored how they experience the consultants (Englund, Nordström & Ohlsson, 2000) or how they handle the staffing process (Strbac, 2006). This study takes a different perspective, as the aim is to investigate one external stakeholder group - graduating students. This group is of interest because it can possibly become an internal stakeholder group - employees. By measuring corporate reputation of firms, one can investigate the subjective perceptions of individual stakeholders towards companies. The perceptions of this group are of great interest to the staffing and recruitment companies. Since the staffing and recruitment industry is developing, more qualified services can be offered. The interests of the staffing and recruitment companies have therefore turned towards employees with higher education, making this an interesting group to investigate. The usage of the concept corporate reputation to assess how possible employees perceive companies that may have interest in hiring them is also appealing since this focus has not been used previously in corporate reputation studies. This study will not separately study the different services within the industry. Since major actors in the industry perform both staffing and recruitment, which also comprise a major part of the industry turnover, this is where focus will lie. It is also not possible to only investigate one service area when examining how the companies are perceived, since the perceptions of the companies are created through the information, communication and experiences exposing stakeholders concerning all services. This study aims to investigate how the stakeholder group, graduating students - potential employees - perceive four staffing and recruitment companies, as representatives of the industry. The corporate reputation measured on the companies sampled will be used to see whether the stakeholder group investigated perceives recruitment and staffing companies in similar or different ways. These implications are to be used to try to conclude how staffing and recruitment companies are perceived by this external stakeholder group. For this study, the population sampled will also be the definition of graduating students. Graduating students are, in this study, defined as currently enrolled students pursuing an academic degree at Jönköping University within business or engineering, who will apply for a job within twelve months.
1.3
Purpose
The purpose of this study is to measure and compare the corporate reputation of four staffing and recruitment companies, from the perspective of graduating students.
3
2
Frame of Reference
In this chapter the authors present an overview to the concept of corporate reputation, its context, and why thesubjectdeserves interest.
2.1
Corporate reputation in a broader context
Much of the research executed on the concept of corporate reputation is connected to that of corporate identity and corporate image. Therefore, the following discussion about corporate identity will be used to introduce and give explanation to the concept of corporate reputation. Many studies have been made on corporate identity but still thereexists ambiguous opinions regarding a proper definition of the concept (Moingeon &Soenen, 2002). Fombrun (1996, p. 36) states that corporate identity "captures the commonly understood features that employees themselves use to categorise how a company approaches the work it does, the products it makes, and the customers and investors it serves." Further, Fombrun (1996) discusses that the corporate identity derives from a company's success and failures during its existence and is recognised by everyone through the name and the many presentations it makes to describe its actions, plans and intentions. These presentations are interpreted by everyone, more or less favourably, who forms images of the company in question. This (corporate) image may or may not be in line with the corporate identity or the many presentations from the company. Stakeholders and the public evaluate, rate, and discuss the images of the company. Fombrun (1996, p. 37) concludes, based on the American Heritage Dictionary's definition of reputation, that corporate reputation is then the "overall estimation in which a company is held by its constituents." It is the affective or emotional interpretation of customers, investors, employees, and the general public to a company's name. Figure 1 describes the connection between corporate identity, image and reputation. In this study only the corporate reputation will be investigated.
4
Figure 1 - From identity to reputation, Fombrun (1996).
2.2
Corporate reputation as a concept
The area of corporate reputation has interested marketing researchers for the last four decades. Many researchers have very similar concepts and definitions, but still there is no unambiguous, generally accepted definition of the concept (Gotsi & Wilson, 2001). Abratt (1989) suggests that part of the confusion arises from the fact that some academics have used the term corporate image and corporate identity interchangeably. In addition, the term corporate reputation and corporate image have been used as identical (Gotsi & Wilson, 2001). The largest confusion in the research has been the issue of whether to separate corporate image and corporate reputation or to use the terms interchangeably. Another issue that differs among researchers is the perspective of whom the subject of the matter regards, meaning; whom it is that has the perception in question. Some researchers state that it is an external audience (Ind, 1997; Moingeon & Soenen, 2002) or outside members (Dutton, Dukerich, & Harquail, 1994) that are in focus. Additionally, some authors use more broad terms, such as stakeholders (Gotsi & Wilson, 2001; Davies, Chun, Silva,& Roper, 2003), implying that it may also include the perceptions of members inside the organisation. Dutton et al. (1994) argue that corporate reputation represents outside members' perception of corporate image, based on his own conceptual framework. Ind (1997) does not differ between corporate image and corporate reputation and suggests that it is "the picture that an audience has of an organisation through the accumulation of all received messages". In this study one external stakeholder group will be in focus, namely graduating students. As stated earlier, one of the major issues that differ in the research is the view of corporate reputation and corporate image (Gotsi & Wilson, 2001). These concepts have been described as identical, totally separated, or as interrelated phenomenon. Gotsi & Wilson (2001) merge marketing researchers broadly into two dominant schools of thought. The first views corporate reputation as synonymous with corporate image. The second school of thought, which is also more recent, considers the two concepts as different. A majority of the researchers, however, regard the concepts as interrelated.
5
Balmer & Stephen (2003) argues that image and reputation are the perception of an organisation by an individual, group or groups. Image, however, reflects perception at one point in time and reputation over time. Balmer & Stephen (2003) elaborates further, saying that reputation is judgements made of the organisation based on the organisation's behaviours, performance, and the collective experiences of the organisation. Reputations tend to be stable but can change significantly due to external or internal events such as industry disasters or ethical lapse by the company. Gotsi and Wilson, (2001) have done a thorough research on the theory available and concluded a definition from all previous definitions used, trying to find an unambiguous definition for future research. They have concluded that "A corporate reputation is a stakeholder's overall evaluation of a company over time. This evaluation is based on the stakeholder's direct experiences with the company, any other form of communication and symbolism that provides information about the firm's actions and/or a comparison with the actions of other leading rivals" (Gotsi & Wilson, 2001, p. 29). Gotsi and Wilson's article (2001) emphasises on the different opinions regarding corporate reputation versus corporate image and argues that the two concepts are not the same but they are highly interrelated. Corporate reputations are largely dependent on the everyday image that people form of an organisation while at the same time corporate reputation can influence stakeholder's everyday image of the firm. Figure 2 describe the interrelation visually.
Figure 2 - Defining corporate reputation, Gotsi & Wilson (2001).
As discussed in section 2.1, Fombrun (1996, p. 36) states that corporate identity "captures the commonly understood features that employees themselves use to categorise how a company approaches the work it does, the products it makes, and the customers and investors it serves." Fombrun's (1996) opinion is that the corporate identity is dependent on all possible presentations that are linked to a company. These presentations are interpreted by everyone who takes in images of the company in question. These images are interpreted by the public and stakeholders to shape the perceptions (i.e. reputation) towards the company. Further, Fombrun (1996, p. 78) also introduces what the author calls a working definition of corporate reputation; "A corporate reputation is a perceptual representation of a
6
company's past actions and future prospects that describes the firm's overall appeal to all of its key constituents when compared with the leading rivals." According to Wartick (2002) the working definition by Fombrun has been used more widely than most. However, in his reflection about the research made on corporate reputation, one distinct point of reflection is given on Fombrun's definition. Wartick (2002) finds the idea that all stakeholders' perceptions have to be aggregated into one assessment of reputation as exaggerated, foremost of measurement concerns. Also the fact that all constituents are included in the definition means problem because of the fact that all stakeholders may be of different importance and interest to the company. Wartick (2002) believes that the definition is designed in this way to suit the most common way of using measurement of corporate reputation, namely to rate and compare companies towards each other. Wartick (2002) argues that a disaggregated perspective would give a more truthful definition rather than saying that reputation is the overall appeal to all of its constituents. This study will take the critique appointed by Wartick (2002) to the definition into consideration. Therefore a disaggregated perspective will be used in this study to illuminate the perceptions about the companies. The corporate reputation will not be investigated for more than one stakeholder group, namely the graduating students. 2.2.1 The importance of corporate reputation
"To economists, reputations are traits that signal a company's likely behaviour. To strategists, reputation is a barrier to rivals, a source of competitive advantage. To marketers, reputations are perceptual assets with the power to attract loyal costumers." (Fombrun et al., 2000, p. 241) Trends are making stakeholders' perceptions more critical to companies and they invite managers to take a more active, centralised, focused, and scientific approach to communicating with those stakeholders. As the interest to value and manage reputation has grown, so has the academic focus towards the subject (Fombrun et al., 2000). Research and theory regarding corporate reputation is something that mostly has been established to suit the perspective of a company, to find out how its stakeholders view the organisation - in order to manage the reputation in ways that benefit of the company. A positive reputation can have several beneficial consequences. Some of the outcomes that reputation has been linked to are: the intention of buying a product or service, the attitude of buyers towards salespersons and products, perceived product quality, attracting investors, and organisation identification among employees (Caruana, 1997). Mahon &Wartick (2003) declare that most research has been conducted on the basis of two major ways to utilise reputation. The first is in assisting the organisation in improving its competitive position - an offensive use of reputation. The second is to use reputation in a defensive matter, to preserve a market position when being attacked. Fombrun (1996) develops these aspects when meaning that a number of tangible assets will evolve from having a good reputation; possibility to charge premium prices for products, lower costs of capital and labour, improved loyalty from costumers, and a cushion of goodwill when a crisis hits. These are consequences that should all contribute to performance difference between firms. Fombrun (1996) continues stating that in order to acquire a positive, enduring and resilient reputation; managers have to invest heavily in building and maintaining good relationships with the company's constituents. A firm with a good overall reputation owns a valuable asset (Caruana, 1997). 2.2.2 Reputation creation
Bevis (1967, cited in Davies et al., 2003, p. 63) describes reputation and also the creation of it in these words; "It is the net result of the interaction of all the experiences, impressions,
7
beliefs, feelings, and knowledge that people have about a company". Corporate reputation is likely created by more than one contact with the company. The perception of how to see a firm is built up of numerous communications (Davies et al., 2003). Bernstein (1967, cited in Davies et al., 2003) describes the complexity of these factors in his model (figure 3), and argues that all images of an organisation are seen through a number of filters, which implies certain preconceptions that influence the image received. The industry and country "filter" in particular, will influence all images received but will at the same time give the same effect to every company sharing the same filters. What differs in the creation of reputation between specific "filtersharing" companies are other factors, or communications, which are shown in the outer ring of Bernstein's model. This model (figure 3) can describe how different stakeholders get their perceptions that ground the corporate reputation for a specific company.
Figure 3 - How reputation is created, Bernstein (1967, cited in Davies et al., 2003 p. 63).
8
3
Method
The next chapter provides a presentation on howtomeasure corporate reputation. Moreover, population and the sample will be discussed as well as the composition of the survey used for collecting the empirical data. In addition, the later sections will touch upon generalisability, reliability, and validity.
3.1
Research approach
Most research starts with either of two different approaches. The deductive approach refers to the logical process of reaching a conclusion from something known to be true, and the inductive approach which involves the process of establishing a general proposition based on observation and particular facts (Zikmund, 2000). In research both approaches can be present. Experiences allow us to draw conclusions that also can be empirically verified by using scientific methods (Zikmund, 2000). This thesis aims to find out how graduating students perceive the staffing and recruitment industry. There have not been many studies conducted to assess one external stakeholder's view of a staffing and recruitment company, something which gives this study an inductive character because of the possibility of reaching conclusions about this specific subject. However, since the respondents' perceptions will be investigated and evaluated based on scientifically grounded facts concerning corporate reputation, the study also has a deductive character. Another aspect to consider, which should fit the purpose of the research, is what type of study the research conducted should have; whether it should be descriptive, explanatory or exploratory. According to Robson (2002), exploratory studies are viewed as finding out what is occurring in a situation; to find insights and to ask questions that can highlight a phenomenon in a new way (cited in Saunders et al., 2007). This study uses different theories concerning corporate reputation to assess the perceptions of the population sampled. The approach taken has not been applied in previous research, and can therefore shed new light on the situation of the graduating students and on the companies investigated. Robson (2002) also explains the descriptive research as a way of assessing a truthful profile to persons, events, or situations (cited in Saunders et al. 2007). For this study, the aim is to give an accurate picture of the perceptions of graduating students, and therefore the study is of descriptive character. Further, this study aims to compare the perceptions towards the four staffing and recruitment companies. This information is attained by using an altered version of the Harris-Fombrun Reputation Quotient (RQ), (see table 3-1 and table 32). This type of study also indicates explanatory characteristics. Saunders et al. (2007) presents explanatory studies as studies on circumstances or problems made in order to describe relationship variables. Therefore, all three of the different study types are present in different aspects of this study. Additionally, there are generally two different research strategies to collect data for empirical research. The quantitative methods are usually measured in numbers or words coded into numbers. Secondly, the qualitative approach put emphasis on words to give a rich picture of the situation and environment (Jacobsen, 2002). The qualitative data collection can come from observations, individual interviews and group interviews. The quantitative strategy can involve data collection through surveys that can be sent out in various ways or conveyed through an interviewer reading the questions for the respondent (Jacobsen, 2002). There are many pros and cons with both strategies; the approach and purpose of the research have to decide which strategy to choose. When conducting
9
exploratory research, qualitative data can give you deeper knowledge of the concept or the investigated problem rather than giving a greater amount of data. However, the quantitative research approach has a clear advantage considering the enhanced possibility for the researcher to make generalisations about the population investigated (Zikmund, 2000). This study is constructed with support by the theoretically grounded measurement, the Harris- Fombrun Reputation Quotient, which so far has been used foremost in quantitative studies. Thus, this study also has a quantitative approach when using questionnaires as method of data collection. Since the aim of the study has been to make some claims about the population investigated, a quantitative method was best suited because of the increased degree of generalisation. Perceptions are feelings and attitudes, which preferably are rated and analysed through a survey rather than through an interview (Fombrun et al., 2000).
3.2
How to measure corporate reputation
In the recent years much more academic research has been devoted to corporate reputation, which has contributed to an increased value of the concept to both practitioners and scholars. As a result, more models and measures for corporate reputation have emerged (Hillenbrand & Money, 2006). Many of the early methods have been developed for ranking companies in, for example, annual beauty contests such as "most admired companies" (Fombrun, 1996), wherein the most visible is probably Fortune's Most Admired Companies (MAC) (Fombrun et al. 2000). 3.2.1 League table rating
One of the approaches to measure corporate reputation is by using league tables (Bromley, 2002), which involves rating companies against each other according to a "reputation score" in a table. Fortune most admired companies (MAC) is one example of league table usage. For that study, the league tables are based on large samples of various kinds of stakeholders, however foremost executives, directors, and security analysts. The sample group then rates a selection of companies based on various attributes that are relevant to corporate success. Further, these companies can be divided into sub-samples representing different business sectors (Bromley, 2002). It can be argued that the most recurrently used and discussed set of data in the field of corporate reputation measurement is the Fortune MAC. It has been annually reported since the 1980s, and is based on the assessment of 300 companies in 40 different industries with a respondent group of thousands of senior executives, outside directors, and financial analysts (Wartick, 2002). The Fortune MAC has, however, been criticised by several researchers (Black, Carnes and Richardson, 2000, Brown and Perry 1994, Wartick, 2002, Fombrun, 1996) for several reasons. Brown and Perry (1994) found that the Fortune MAC is very dependent on the financial performance of the company creating a so called halo effect implying that the financial performance alone becomes too determinate in the overall impression. Brown and Perry (1994) argue that this effect has to be eliminated to make the study accurate. Other researchers have, however, argued that the solution does not come that easily. Harsh critique against the Fortune MAC is presented by Baucus (1995), who argues that data is so methodologically suspect that it should not be used at all. Wood (1995) also compares the findings from Fortune MAC as what the foxes think of the hen house (both cited in Wartick, 2002), implying that the outcome of the studies is rather obvious. Szwajkowski and Figelwicz (1997, cited in Wartick, 2002) are less harsh in their critique and think that the problems can be managed. This is also backed up by Wartick (2002), except for a few objections. Fombrun et al. (2000) present a table with the following methodological limitations of the Fortune ratings;
10
"Biased sampling frames, target firms selected by size of revenue, restriction to publicly traded companies, collusion because of sector membership of respondents, overrepresentation of senior managers, directors and financial analysts in samples, respondents may lack direct experience relevant to some attributes (items in rating scales), mainly pen and paper mail surveys." (cited in Bromley, 2002, p. 35). Bromley (2002) continues by adding to the above mentioned arguments that reports describing the surveys do not always state what instructions the respondents were given and how the questionnaires or rating scale were formed. Despite the critique, a number of industry followers within publications have adapted and based their own reputation rating measurements on the Fortune MAC, among them Financial Times and Asian Business (Fombrun et al, 2000). Moreover, Bromley (2002) presents an alternative method for constructing a league table of reputation using a standard rating scale. This method is called the free descriptive method. The method enables respondents to list various attributions (impressions of attributes) based on their experiences and interests. The respondents would be drawn from selected stakeholder groups depending on the purpose of the research. The more frequently mentioned attributions could then be used to construct a league table of merit using a methodology similar to that associated with the Fortune surveys. Attributions, regardless of how they are selected, could be assigned weights according to their frequency of occurrence or estimated importance. The free descriptive method tends to generate a wide range of attributions, most of which are shared by relatively small proportions of respondents. The differences in occurrence of attributions could indicate their relative importance, familiarity or salience at the time (Bromley, 2002). 3.2.2 The Reputation Quotient
In 1999, a number of researchers developed the Reputation Quotient (RQ), also sometimes referred to as the Harris-Fombrun Reputation Quotient (Davies et al., 2003). When developing the RQ, the starting point was 27 attributes (i.e. the items investigated) drawn from the most commonly used instruments to measure corporate reputation (among them Fortune MAC) along with additional factors suggested by academic literature on reputation and items provided by an commercial marketing firm, Harris Interactive. 32 attributes made up a first prototype model which was tested and refined through an initial pilot testing in the US air force and a following series of focus group interviews. A second prototype was tested in the PC hardware industry and, after a third study to validate the model, involving 8545 respondents, the final version of the RQ was presented. The findings demonstrated that people justified their feelings about companies on one of 20 attributes, which in turn was grouped into 6 dimensions (Fombrun et al., 2000), which are presented in table 3-1. Fombrun et al., (2000) suggest that reputation is a construct that combines two factors - emotional appeal and rational appeal, the latter representing a net assessment of the five dimensions, apart from emotional appeal. The 20 attributes are items used for measuring corporate reputation with the Harris-Fombrun Reputation Quotient. These items are easily applicable to use as perceptual questions about firms.
11
Table 3-1 - The Harris-Fombrun Reputation Quotient, Fombrun et al., (2000)
Emotional Appeal -good feeling about the company -admire and respect the company -trust the company Products and Services -stands behind products/services -offers high quality products/services -develops innovative products/services -offers products/services that are good value Vision and Leadership -has excellent leadership -has a clear vision for the future -recognizes/takes advantage of market opportunities Workplace Environment -is well managed -looks like a good company to work for -looks like it has good employees Financial Performance -record of profitability -looks like a low risk investment -strong prospects for future growth tends to outperform its competitors Social Responsibility -supports good causes -environmentally responsible -treats people well
The Harris-Fombrun Reputation Quotient is an index that sums up people's perceptions on these attributes, which can be used to rank companies in regard to their reputations. This aggregated usage of the RQ is, as when using league tables, a source of critique. Wartick (2002) has the opinion that to use one aggregated score to rank companies is not a credible and justified method. The reason for this frequent measurement approach is that the most common reason for measuring reputation is to rank companies against each other. Despite the criticism about lack of a rigorous conceptual definition in the study (Wartick, 2002; Groenland, 2002), Wartick claims that RQ is good in the sense that it is broad and can be applied on anyone, whether they are customers, employees, or the general public. This is, however, somewhat contradicting Fombrun's (1996) definition of corporate reputation which includes "all constituents". Wartick (2002) implies that the different stakeholders may have different importance in determining a grand aggregated reputation and therefore it is of higher interest to examine stakeholders separately. Groenland (2002) claims that the RQ shows satisfactory psychometric properties and that practical experience with the model, in many different commercial settings, increases confidence in usefulness of the method. According to Wartick (2002), the RQ meet most of the characteristics for a solid measure of reputation even though the entire data set is not available in literature. Groenland (2002) has performed a qualitative study to validate the dimensions of the RQ in which all six dimensions are supported as relevant for measuring corporate reputation. Groenland (2002), however, stresses the importance of being open to altering the dimensions depending on what is meant to be studied. For example, the cultural differences are raised as a factor that might affect the appropriateness of attributes being investigated. Groenland (2002) suggests that empirical analysis should be used to address the issue, in order to design an appropriate model with attributes suitable for the matter of interest. In his study, focus groups were used in an initial stage to validate the suitability of the RQ attributes when measuring corporate reputations in the Netherlands. The outcome of the study was that two additional dimensions (national origin and charismatic representatives) should be added in the Dutch case, and possibly in the original RQ-scale. For this study, inclusion of these dimensions has been considered. However, national origin and charismatic representatives were not found adequate for the setting of investigating the graduating students' perceptions of staffing and recruitment companies. This study has used the RQ to investigate how graduating students perceive the recruitment and staffing companies. The study has measured the attributes on all
12
companies investigated, and an aggregated assessment of the outcome of the RQ attributes are presented in a league table. Further, various analyses have been made to strengthen and validate the findings. The corporate reputations have only been determined through the investigation of one stakeholder group, namely graduating students. This disaggregated assessment of stakeholders' perceptions of corporate reputation is encouraged by Wartick (2002). Groenland (2002) stresses the importance of being open to altering the dimensions depending on what is meant to be studied. Therefore, the suitability of all the attributes from the RQ model have been reflected upon, whether to be included in the main survey, before the pilot study. Alteration has been made on the attributes and dimensions of the Harris-Fombrun Reputation Quotient, in order to suite the area and population of interest. Attributes that were not clearly applicable and understandable in the context of the study were removed to fit the sampled population and the investigated topic. This also led to a change in the dimensions of the Harris-Fombrun Reputation Quotient. The attributes and dimensions concluded appropriate for measuring the corporate reputation of staffing and recruitment companies, through the eyes of graduating students, have made the foundation on which to build the study. This is presented in table 3-2. When mentioning the Reputation Quotient or RQ from hereon in the report, this version is the one referred unless other is stated.
Table 3-2 - The altered Reputation Quotient (RQ)
Emotional Appeal -good feeling about the company -admire the company -trust the company Products and Services -the company offers high quality work opportunities -the company is an attractive employer -the company uses innovative marketing strategies -to work for the company will yield a valuable experience Financial Performance & Vision and Leadership -the company takes advantage of market opportunities -the company tends to outperform its competitors -the company has strong prospects of becoming more successful in the future Workplace Environment -the company seems to have a well functioning organisation -looks like a good company to work for Social Responsibility -the company has an important role in the society -the company is important for the job market
The quotient calculated, using RQ, is aggregated into one assessment, an index based on all attributes measured, to compare to other companies. Wartick (2002) does not support this aggregated measurement, claiming the extent to which the attributes used affect the overall impression of the reputation is not apparent. The concern could be solved partly by assigning different weightings to the attributes but Wartick (2002) advocates that this is hard to accurately determine. Wartick's critique has been addressed in the sense that RQ used in this study have also separated and presented by dimension and question to be able to se what relative differences are apparent, between the respective companies.
3.3
Sample selection
In this study the perceptions of one stakeholder group, namely the graduating students are in focus. This group was selected in order to have a more specified population that is clearly defined. It can, however, be argued that this population is still relatively unspecified to examine. In this study graduating students are defined as; currently enrolled students pursuing an academic degree at Jönköping University within business or engineering who will apply for a job within twelve months. Graduating students are, as a population, rather
13
un-investigated as a stakeholder group to the staffing and recruitment industry. It is, however, an interesting and important group to investigate. As the choice of education and the companies marketing efforts towards graduating students can be significantly different between different geographical areas, it might not be accurate to place all graduating students in the same population. Therefore the population is narrowed down to solely graduating students from Jönköping University in the fields of engineering and business. This delimitation has added validity to the study and enhanced the degree of generalisability. Saunders et al. (2007) discusses the generalisability issue when considering the applicability of findings on more than one organisation, and says that in such cases it should not be the aim to produce theory for all populations but to explain the particular setting of interest. This reasoning is also applicable on this study. Considering all graduating students as a part of one population would reduce the possibilities to accurately describe the perceptions of the students towards the companies investigated. The authors have found another aspect to highlight. Most companies in the staffing and recruitment industry have a common denominator, namely to have more job offers to people with an engineering or a business administration education. Therefore, only students enrolled in such programs were sampled. In this study the sample were picked from graduating students at Jönköping University within the fields of business and engineering. This is because all of the companies investigated are represented in the area of Jönköping, and all the respondents are as likely to have been exposed to same marketing efforts of the companies by various fairs and marketing campaigns on the campus area. Lists of graduating students are easily obtained and access to the research objects through classes is possible. This gives the opportunity of convenience sampling usage. This choice of sampling method has also provided the possibility to investigate a large proportion of the population. Saunders et al. (2007) describes the method as haphazardly picking people that are easily obtained. The method is also called Haphazard sampling. The sampling process is thereafter continued until a quota is reached (Saunders et al. 2007). As it was hard to estimate how many respondents that would be willing to participate in the study, a quota were not set at a specific number. The quota was instead decided upon the amount of responses that could be collected and were valid in regard to the required population characteristics. It is also important to note that it is not the study areas of the students, being business and engineering, that are in focus, but the fact that the companies sampled in the study are hiring people in these fields. 125 respondents were investigated in total, divided as; 31 respondents responded to questions on Poolia, Academic Work, and Adecco, while 32.responded to questions regarding Manpower questions. All methods of sampling are imperfect, and there will be some bias present. In convenience sampling, the respondents are picked due to the accessibility. Convenience sampling does, however, jeopardise the representativeness of the sample (Ruane, 2006). This is a source of bias making it possible for people that are part of the population but not being available, being underrepresented. Saunders et al. (2007) exemplify this as instead of investigating all managers you are investigating managers taking an MBA course due to the convenience of obtaining responses. Further, Saunders et al. (2007) argue that this bias is of minor relevance when there are small variations in the population. It can be argued that the differences between the persons not present for sampling and the persons that are is quite small. This, because all of the sampled students are enrolled in an education program, and are likely to attend at least some lectures. All respondents sampled therefore have been present in the classes and everyone has been asked to complete the survey. However, since
14
only the graduating students in engineering and business should be sampled a set of control questions helped limit the occurrence of the wrong persons being sampled. The respondents not part of the population were excluded from the study. However, it is hard to eliminate the fact that only the students participating in the lectures are sampled. This is further elaborated in section 3.8. 3.3.1 Company sampling
The companies selected for this research have had four prerequisites. They should all be targeting the investigated population through both marketing and possible employment opportunities, they should all be represented in Jönköping, and they should all be engagedin both staffing and recruitment services. To get this information, randomly picked companies have been investigated through various internet sources to see if they match the prerequisites. Again convenience sampling as presented by Saunders et al. (2007) has been used; the first four companies that matched our criteria were sampled for investigation. The quota of four companies was determined because it was the largest number considered manageable in this study. It was thought to be manageable in the sense that each company would be assigned an adequate number of respondents. Fewer respondents on each company would decrease the degree of generalisability. The reason why we are investigating firms active in both staffing and recruitment services is because the staffing service is very dominant in the industry. Recruitment services are, however, still influential on how the companies are perceived. The contracting and outplacement services are still unexploited and therefore not regarded in this study. 3.3.2 Sample characteristics
The four companies that are subjects of investigation for this study are Adecco, Academic Work, Manpower, and Poolia. The number of respondents was 31 across three out of four companies and Manpower had one more respondent. When distributing the surveys, an effort was made to attain equal number of respondents between the four sampled companies. Even though the number of respondents has been equally allocated the share of respondents in regard to gender and school diverge between the companies sampled. Adecco and Poolia have close to a 50 percent share between male and female respondents. Manpower together with Academic Work have a sample that is more male dominated as close to two thirds of the participants were males. Further, no particular difference can be found between the schools, where the sample was evenly allocated between the engineering (JTH) and business (JIBS) students. The mean age of the respondents were fairly similar between the companies, with mean age varying from 23.37 to 25.06 (See appendix 5 for statistics).
3.4
Psychometrics
The word psychometrics refers to the measurement of subjective estimates, such as judgements of a brand's attractiveness or a company's resolve (Bromley, 2002). The traditional rating procedure in social science is based on the following assumptions; "That the various attributes share a common factor, namely a company's ability to provide valued outcomes, and that the attributions (impressions of attributes) are quantifiable, and ratings constitute valid and reliable measures or indicators of these quantities." (Bromley, 2002, p. 40).
15
According to Bromley (2002), Fombrun demonstrates such a common factor in the RQ scale. However, Michell (1990, 2000) and Kline (1998) have reservations against the way psychometrics is used in the social and behavioural sciences (cited in Bromley, 2002). Bromley (2002) implies that when constructing a psychometric scale for assessing corporate reputation, the ideal procedure is to identify a representative set of attributes or a set of attributes within categories that are relatively independent of each other but still correlated with the overall score or criterion. Again, the Fortune surveys are taken as example. It is likely that the usual attributes for corporate reputation used in the Fortune surveys are interrelated, especially when investigating firms within a particular business sector. Further, it is stated that with reference to corporate reputation, a true score consists of the aggregated score a company would achieve on all the quantitatively scaled attributes related to corporate reputation. Bromley (2002) continues by saying that a test of corporate reputation contains a random set of categories. As the measurement is an estimate of a company's true score, it is the sum of the sub-scores from the items in each of the descriptive categories. The number of categories required and the number of items in each category are issues to be settled by empirical inquiry and practical convenience, within the legitimate psychometric framework. From a reliability standpoint, assuming the test is within the respondents' performance capacity, the larger number of categories and the larger the number of items within the categories the better. Bromley (2002) also states the importance of remembering that subjective attributions represent real and supposed attributes. Real attributes have an independent existence; however attributes may or may not be truly presented by a psychometric assessment. The stakeholder group investigated, that are asked to express attributions about a company, should contain either the entire group or a representative sample of it. The procedure for eliciting the attributions, such as a rating scale, checklist or freedescription should be standardised as should the instructions to respondents and the method of recording and analysing the data. The attributions should be reliable and valid, and a company's reputation score only makes sense when in relation to scores of other companies, and sensibly compared (Bromley, 2002). Kline (1998) and Michell (1990, 2000) attribute a problem to the psychometric exercise because it seems that it is not quantitative in the strict sense of the word. This implies that the numbers assigned to corporate attributes are not quantitative enough. The argument continues by exemplifying that there are no real measurement units for subjective attributions. The meaning of the numbers depends on understanding how corporate attributions have been defined and assessed (cited in Bromley 2002). Bromley (2002) also states that some stakeholder groups find certain corporate attributes appealing while others find other attributes appealing. However, the company success is dependent on theproduction of valued outcomes for all stakeholder groups. A solution to the problem would be to devise separate measures of corporate reputation for each stakeholder groups. Further, the companies could then be compared across the profile of stakeholder assessment (Bromley, 2002).
3.5
Choice of data collection
There are several different ways to use quantitative questionnaires, but overall they can be divided into two categories; interview-administrated and self-completion questionnaires. Interview administrated methods implies that the researcher is present during the data collection, whereas the self-completion refers to methods answered without researcher interference (Brace, 2004). Further, there are also mixed methods of the two mentioned,
16
such as interviewer supervised self-completion and interview administrated methods which can contain elements of self completion for the respondents. Brace (2004) has identified three key benefits of interviewer-administrated interviews; queries about the meaning of the question can be dealt with, misunderstood questions can be corrected, and the respondents can be encouraged to give deeper responses to open- ended questions. Saunders et al. (2007), emphasise that quantitative interviews usually are structured, where the interviewer has a set of questions and asks the respondent face-to- face. Brace (2004) points out that there is an interviewer/interviewee bias present when the interviewer is active in the interview, which means that there is a risk that the interviewer cause bias if, for example, asking the questions in different ways. Also respondents can try to impress the interviewer in his/her answers or try to avoid loss of face. Further Brace (2004) implies that these problems are slightly reduced in telephone-interviews, which is one method of interviewer-administrated interviews. For this study it is important to eliminate such bias in order to give higher validity to the perceptions of the respondents, which this research aims to analyse. Interviewer/interviewee bias is eliminated in self- completion surveys since the interviewer is not active (Brace 2004). Another problem with interviewer administrated interviews is the cost of obtaining a sufficient representative sample (Brace, 2004). In this regard, the self administrated surveys allow that a larger sample being collected for a lower cost since the interviewer does not have to be present for each respondent. This increases possibility to generalise for the targeted population. Saunders et al. (2007), however, point out that the response rate tends to be higher in cases where the interviewer is active. Saunders et al. (2007) also present a version of self completion surveys that tends to have a rather high response rate, and significantly higher than internet based and postal self completion surveys - the delivery and collection surveys. As the name implies this type of survey is in paper form and is handed out and later collected by the researchers. This is a method which brings an important advantage for this study. The sampled population in this study, graduating students, is still active in the university environment and attends classes. Therefore, this method can be used and the surveys can be given the students as they attend classes in auditoriums. This will reduce the constraints of administrating costs and time consumption significantly. Self-completion questionnaires have been chosen for data collection in this study. This is due to the possibility to investigate a large sample combined with low costs of obtaining it, as well as because of the reduction of the interviewer bias. There will also be room for a short presentation of the study if the surveys are handed out in an auditorium, which can give a deeper understanding of the research conducted. 3.5.1 Questionnaire design
As this study aims to investigate the corporate reputation for four staffing and recruitment companies from the perspective of graduating students, the investigation will be based on the already grounded theories on corporate reputation measurement. To examine the concept of corporate reputation, this survey has used the twenty items of the Harris- Fombrun Reputation Quotient (table 3-1) as a starting point. This study uses screening questions (Q17, Q19) in order not to include non-representative respondents in the sample. The screening questions are used to be able to only include the people that are representative for the population. When researching attitudes and behaviours, researchers tend to want to study people with certain characteristics (Brace,
17
2004). For this study it is of value to use screening questions to validate that the students attending classes are graduating students, and not students only taking a course or students that will not graduate and apply for a job within a year. The questions regarding the actual topic of interest are called the main questionnaire (Brace, 2004). When setting up the main questionnaire, it should normally start with the most general topics, before getting into detail (Brace, 2004). In this survey there will be an initial question added, where the respondents are asked to pick the most important factor when looking at a company. This introduced the respondents to the following questions and has been subject for analysis. In the response options, there has also be left a space where the respondents will be able to fill in an alternative answer if they would feel that the RQ factors are not sufficient enough for answering this question. This question has also provided information on what dimension in the RQ that is the most important when looking at a company. However, as it is corporate reputation that is investigated, this alternative has also been included to conclude whether or not overall reputation is most important in determining how to perceive a company. After the initial question, the factors of RQ will be investigated. There are many different response scales to choose between. However, since the RQ table (table 3-1) is built up mainly of different claims, a Likert scale seems to fit this situation best. The Likert scale is one type of attitudinal rating scale, which is sometimes called an agree-disagree scale (Brace, 2004). Most of the times the scale ranges between 1-5, where 1 represents strongly agree and 5 strongly disagree (Brace, 2004; Buglear, 2005). When setting up Likert scale type of surveys it is important that the statements are clear, so that the respondents are reluctant to pick the middle answer, neither agree nor disagree, due to the reason that they do not fully understand the statement (Buglear, 2005). Saunders et al. (2007) suggests that one could use an even scale (1-4) to force the respondents to take a stance. However, this can also make the respondents pick alternative that they do not fully agree with, which gives the answers less accuracy. Brace (2004) reflects upon the issue of what the author calls don't knows. People are reluctant o leave a scale blank even if they do not really have an opinion or do not know what to answer. Therefore, it can be good to include the don't know as one option in the scale. Further, Brace (2004) suggests that researchers should consider whether they want to be able to distinguish between mid-point responses, such as neither agree nor disagree, and responses made from respondents stating that they genuinely did not know what to answer or could not answer. This type of non-response can raise issues when using certain data analysis techniques. Taken Bromley's (2004) suggestions into consideration, this study the above mentioned has been taken into consideration, and an uneven scale (1-5) will be used. Besides that there will also be a sixth alternative included, in form of a have no opinion option. The have no opinion is included because the desire to distinguish the ones who may not consider themselves to have sufficient knowledge about the companies to state their perception from the ones knowing, but having a neutral opinion. These non-responses can also be further compared between the companies. Buglear (2005) points out that it is of significance not to include too many choices on the scale and beyond six can make the respondents indifferent between the options. Another important issue to regard is that if you use a series of statements you should keep the same order of response categories to avoid confusing the respondents (Dillman, 2000 cited in Saunders et al., 2007). As the RQ factors are grouped into six dimensions; emotional appeal, products and services, vision and leadership, workplace environment, financial performance and social responsibility.
18
The dimensional structure in the Harris-Fombrun Reputation Quotient has also been implemented in the questions used in this study, to ensure a logical flow of the questions. Another issue that can be of value to investigate in this study is how reputation is created. Bernstein presents a model (figure 3) for how corporate reputation is created, and the factors listed in this model will constitute the base of response options to a question that will be added into the questionnaire. There will, however, be slight modifications on some of the factors. Price points will be excluded from the response options due to the fact that it is not of interest for the stakeholder group studied in this research. Price points are more relevant for firms that are producing a product or service for its customers and customers of the staffing and recruitment companies are the firms that hire them to get staffing labour or recruitment services. Design will also be excluded as response option, since this is related to products or services as is Price points. Direct mail will also include e-mail, and media comment will also include internet sources. There will also be given room for the participants to express another response than the ones provided from the Bernstein model. The inclusion of reputation creation can give the study more credibility as a whole since it gives information on where the perceptions come from. For introduction purposes, some minor instructions prior to the questions are going to be presented in order to help minimise misinterpretations of how to fill in the questionnaire. The survey questions must have a constant and logical flow to help ensure that the questionnaire works as a conversation with the respondent. Most of the time this helps to get the candidate more involved and eager to devote attention and to answer as truthfully as possible. As with all methods there are always some flaws to be aware of when using Likert scales. Brace (2004) has identified four main problems; order effect, acquiescence, central tendency and pattern answering. The order effect arises from the tendency of respondents to pick alternatives to the left (Arlingstal, 1978, cited in Brace, 2004). Acquiescence means that respondents are more reluctant to pick disagree answers than agree (Karlton & Schuman 1982 cited in Brace 2004). Brace (2004) suggests that a possible solution is to place the disagree options to the left of the scale. Central tendency refers to the respondents as more willing to choose a middle option than the extreme ends of the scale. Albaum (1997) suggests the usage of a two-stage-question, which implies that you should divide your 1-5 scale into first asking the respondent if they agree or disagree and then in a second stage ask how strongly you would agree to your just given answer (cited in Brace, 2004). However, since the scale used in this survey will be made a bit differently with the choice of no opinion, it makes the extreme ends look differently than the traditional scale (1-5) and this segmentation was not needed. Brace (2004) further discusses that this type of two- stage-questions can be very time consuming and are more appropriate for telephone interviews. Pattern answering occurs when a respondent starts ticking boxes as a routine without considering the statements. This is often due to boredom or fatigue. Brace (2004) suggests a solution to mix both positive and negative statements, so that the respondents really have to read the questions, which have been considered for this survey. There has also been left some "white space" between the groups of questions representing the six dimensions of the RQ scale, in order to possibly prevent patterned answering. 3.5.2 Pilot Study
Quantitative research methods such as questionnaires almost always have identifiable errors. The errors range from typographical mistakes, form errors, presentation flaws to ambiguous survey instructions. Most of the time this occurs since the researchers are
19
working too close to the project and thus may overlook even the most obvious errors (Litwin, 1995). Researchers decide to conduct pilot studies to measure the range of ideas or opinions that people have or to find out the way that variables are related to each other (Fowler, 2002). For a pilot test to be effective, it is important that the respondents are similar to those who will be asked to participate in the main survey (Fink, 1995). In this study the target group is graduating students from Jönköping University within business or engineering. In order to use a sample as similar to the population as possible, the actual population was sampled. Therefore engineering and business students graduating from Jönköping University constitutes the pilot test sample. There are some key factors in the pilot survey that have been set under specific attention. Since the aim of the survey is to examine corporate reputation based on the already grounded theories on the matter, it is important that the respondents can locate and differentiate their response according to the ranking figures. Pilot surveys are an essential part of that trial process (Brace, 2004). The pilot testing allows the author a chance to correct these flaws before the main survey is handed out and used to gather real data (Litwin, 1995). Brace (2004) argues that failing to pilot the questionnaire thoroughly risk leading to failure with the project. Brace (2004) has pointed out some reasons why a questionnaire should be piloted and in this pilot survey the following is taken into consideration. First of all, the questions have to sound right. Most of the time questions can look acceptable on paper, but errors and misinterpretations are revealed when read out loud. This pilot survey has been both read out to its respondents and after the study was filled in, a complementary interview regarding the questionnaire design and content has been performed. Another thing to keep in mind is if the respondents understand the survey questions (Brace, 2004). Simplified language should be used to avoid complicated words and ambiguous questions which might lead to incomprehensible answers. No leading or loaded questions and no right and wrong should be promoted throughout the survey. Further, the researcher should ask the participants if the survey allowed them to say everything they wanted on the subject. This can help locate if some questions are imperfectly formulated or confusingly explained (Brace, 2004). The performed pilot study along with the following discussion of the questionnaire provided many valuable insights to support a restructuring of the questionnaire. The altered questions were tested and the general structure was considered suitable. Some ambiguous questions were altered and some unobvious instructions were clarified. Some of the questions were experienced differently in the Swedish and the English versions of the surveys and therefore corrected. 3.5.3 Translation bias
There are many biases to consider when conducting research, and one of them is the issue of translation. The surveys used for this study have been both in Swedish and English. This has been used due to the differences in respondent groups. The programs at the engineering school are mostly taught in Swedish therefore the Swedish surveys are used there. The programs at the business school are mainly taught in English and there are many international students present. Therefore English surveys have been used there. The surveys were built up the exact same way, with the same questions. The translation of the surveys has, however, been done by the researchers of the study. To ensure that the questions still sounded the same way, and the meaning of the questions were understood
20
the same way by the respondents, the surveys were piloted on several bilingual persons representative for the population investigated.
3.6
Generalisability
Vogt (1993) has described generalisability as; "the extent to which you can come to conclusions about one thing (often a population) based on the information about another (often a sample)" (cited in Collins and Hussey, 2003). One of the most important things about generalisability is that the findings should be applicable to another research setting, such as other organisations (Saunders et al. 2007). In other words, this is in line with the definition by Vogt (1993). As to ensure a high degree of a possibility to generalise the sampling method is crucial to avoid the many biases connected to it. Many researchers have argued that the only sufficient sampling method to use to be able to generalise for a population is random sampling. For this study, convenience sampling has been used; this type of sampling has a degree of randomness connected to it. However, it is not the same as random sampling, the respondents are picked due to the accessibility and, as discussed in section 3.3, this creates an underrepresentation of people not attending classes where the population was sampled. Further, the sampling turned out to be unevenly distributed in terms of gender and school. To be able to fully generalise for the population, the sample should be representative for the characteristics of the population. This implies that the sample should have the same gender and school distribution as the population as a whole to avoid a bias. For survey research, the issue of generalisability is closely linked to the issue of validity. (Ruane, 2006). Even though a pilot study was conducted, one can not completely rule out that the respondents have interpreted the questionnaire questions differently than the researchers intended them to do. This lowers the degree of validity in the findings, causing a decrease in the level of generalisability. Due to the discussed flaws, connected to the study, the authors find that the degree of generalisability has been affected negatively. Therefore, the results presented should be interpreted as an estimate and not as the truth. As Ruane (2006) argues it is important to point out is that research does not describe the truth but instead gives an estimate on how the reality looks like. There will always be biases present.
3.7
Reliability
In general, a reliable measurement is one that yields the same result if measured more than once. This holds only if there has not occurred any change in the variables measured (Ruane, 2006). Further, reliable measures do not fluctuate due to random events, but yield consequent results. This is quite easily controlled when a particular instrument is used, such as a thermometer that measures a temperature of something. Measuring the same thing more than once to check reliability of measurement is called test-retest (Gomm, 2004). However, when investigating people and their behaviour and opinions it gets more complicated. To ask a person to do a test twice can seem rather awkward and the person is likely to just repeat what was just stated. This is also why the time aspect is of interest. If doing a retest within an interval that is to narrow, the answers attained are most likely influenced more by the memory than the actual feelings (Ruane, 2006). For this study a retest is not possible, since the respondents are not available more than once. Also, there was another aspect to consider. Since this study aimed to investigate perceptions, it can not be guaranteed that the respondents feel the same way today as they do, for example, in a month. As presented in section 2.2.2, there are several ways that
21
reputations are created, and as the respondents get more impressions of the companies their perceptions may change. Therefore it is not accurate to do a retest after a long time period. Ruane (2006), however, states that this does not necessarily have to indicate low reliability. There are other ways to assure that what is measured is reliable. In survey research it is common to instead ask several questions on the same concept and then compare the answers. For example, a question on age can be asked both as how old a person is, as well as the year the person was born to assure that the same age is given in the two questions. One type of question asked like this is called single items (Ruane, 2006). There are also multiple items that are used when you ask several questions regarding the same aspect. This is not as straight forward as age, but still useful to assure reliability (Ruane, 2006). The Fombrun-Harris, Reputation Quotient is constructed in this way, being a reliable scale to measure corporate reputation. The surveys used for this study are constructed with the RQ as a base but altered to better fit the population investigated. When you are investigating one phenomenon, there is a common technique used for measuring reliability, namely Split Half, used for checking a scale for internal consistency (Ruane, 2006). The questions that are investigated are divided into two sections, and then the scores of the two groups are compared by looking at how closely they are correlated. This coefficient is called the Cronbach's alpha and varies between 0.0 and 1.0, with 1.0 being perfectly correlated (Ruane, 2006). Since a measurement of corporate reputation is done in this study, the Split Half technique can be applied and calculated. Pallant (2006) suggests a number of steps when checking for reliability with the help of SPSS. First, all mean scores should be examined to see whether they are what could be expected. Secondly, the Cronbach alpha value should be calculated to check internal consistency of the scales. If the alpha values are too low, you may need to consider removing items with low item-total correlations. Pallant (2006) claims that an alpha value over 0.7 is ideal and shows that the scale is reliable with the sample. For this study all means of all questions, for all companies are within expected ranges. Three out of four Cronbach alpha values where over the required value for reliable scales and thus showing internal consistency. The only faulty alpha value was the one for Adecco (0.665). The Alpha if item deleted option has been used in order to see what impact the removal of each item has on the alpha value. The results show that when removing the Marketing item, the alpha value was 0.775 and thus above 0.7. Reflection has been made if the item of the scales needed to be excluded. Since the remaining three companies show no similar signs to the removal of the Marketing item, it will be kept for analysis, also for Adecco. This result will, however, be accounted for when further analysing the data. All statistics from the reliability testing can be found in Appendix 3.
3.8
Validity
There are many different aspects of validity, but mostly it regards whether using concrete, observable support for claims on the reality or not (Ruane, 2006). The aspect of measurement validity refers to whether the study measures what it is supposed to measure. Ruane (2006) stresses the importance of the issue by exemplifying the debate of lie detector tests. The instrument is argued to have low validity due to the fact that it does not really measure if a person is lying or not, but other factors that can indicate that a person is lying. In this study the measurement validity is of great importance, due to that the questionnaires used for empirical data collection have to be valid and measure corporate reputation. Since the Harris-Fombrun Reputation Quotient is a recognised way to measure corporate reputation, this leads to a higher level of validity. However, not all items in the RQ are
22
applicable as aspects to consider for all stakeholder groups. Therefore, the items investigated have been altered so that the answers retrieved from the respondents are credible perceptions, see table 3-2. This is also of value when considering the aspect of external validity. External validity refers to what extent the results of a study are applicable to other settings and groups (Ruane, 2006). The findings of this study can be applicable to other settings. The findings of this study can also give an estimate of how the reality looks like for this setting. However, since corporate reputation is a widely debated concept that deals with perceptions, it can not be argued that the findings from one population also hold for another. External validity is also linked to the sampling method. When sampling people as respondents to answering a questionnaire, it created a bias, this due to the under representation of the uncooperative respondents Gomm (2004). However, for this study, this has not been a major problem since almost all the respondents that were given a questionnaire also returned it. The under represented group in this case would instead be students not attending classes. This can possibly lower the validity and the possibility to generalise for the population as a whole. Yet, referring back to section 3.6, the findings of this report are an estimate and not an exact picture of the truth. There are several more biases that can reduce validity of a study, but one might be more significant than others; there is no guarantee that the respondents have interpreted the statements in the surveys as the researchers intended them to do. In survey research this can never fully be eliminated, but for this study the pilot study was used to try to reduce this type of bias as much as possible.
3.9
Analysing data
When conducting statistical analysis today most of the methods involve applying a statistical computer program to help deal with the empirical data. There exists a large variety of statistical software for analysing purposes both for professional and personal use. Some of the more frequently used software packages are SAS and Minitab although the most widely used statistical software today is SPSS (Statistical Package for the Social Science) (Ghauri & Grønhaug, 2005). The software SPSS has been used in order to compile and analyse the results from the survey data. According to Byrne (2002), the use of SPSS for a presentation of a large number of variables is best created through an exploration process. An exploration is a procedure of organising the data so that it can be ordered in some way to find out if any meaningful patterns can be extracted from it. The summarising is done through standard descriptive statistics, which is a method to describe the overall character and patterns of the variables. Examples of descriptive statistics are graphs and tables (Byrne, 2002). Quantitative data can be divided into two groups; categorical and quantifiable data. Also Microsoft Excel has been used to be able to present more intuitive graphs and tables. Categorical data, sometimes called nominal data, refers to data which values cannot be measured numerically. It can, however, be classified into categories (Saunders et al., 2007). The initial question (Q1) together with Q16 in the questionnaire yield nominal data since the variables cannot be measured numerically, but set into a specific classification according to the responses given. Nominal data is categorical data where the order of the categories is arbitrary (Saunders et al., 2007). Ordinal data is a more precise form of categorical data, where there is a logical ordering to the categories (Byrne, 2002). The following questions in the survey (Q2-Q15) are of ordinal
23
data character since the format of Likert scale is used to classify the respondent's perception towards the different companies. To put the answers in relation to the population sampled and to describe the characteristics of the sample, a set of control variables has been included in the bottom part of the questionnaire. The remaining questions (Q17, Q18) are not a subject of further analysis, but used solely as screening questions to exclude non representative persons for the population. Pallant (2005) suggests a methodical way of processing data gathered through with the use of SPSS. This way first includes preparing the data by defining variables and coding responses. Then data should be entered in a data file and then screened and cleaned to check and correct possible errors in the data. If errors are found, one should identify them is SPSS and go back to the questionnaires in order to correct them to the accurately given values (Pallant, 2005). These steps have been performed for the respective four companies in separate data files. The response variables have also been manipulated in the sense that the coding for the negatively worded items has been reversed. When designing the study, the choice was made to include a non-response item - have no opinion. The screening of the data showed that a fairly large number of cases (persons) have marked the answer have no opinion in the questionnaire, indicating that they have no opinion about the particular perceptual question. These responses have been treated as missing data in the analyses. Pallant (2006) suggests using the option exclude cases pairwise when analysing the data in SPSS, which excludes the cases with missing data only if they are missing data required for a specific analysis. The cases will still be included in the analyses for which they have the necessary data. Further, the screening showed that the number of actual missing data (i.e. no answer given at all) was very small - only a few cases had a one or a few blank response boxes. These answers have been treated as missing data, i.e. been given no value in order not to effect the further analyses. In a separate analysis about the have no opinions these actual missing values and the have no opinion responses will be merged, in order to separately analyse these missing values through SPSS. This procedure can also be supported by reflecting on the likeliness that a missing answer might in fact be an actual statement saying that the respondent had no opinion or knowledge about the subject in question. Apart from this, the data did not need further corrections, since the data was within all limits of expected values. The characteristics of the samples, drawn from the descriptive statistics used in these steps, can be found in section 3.3. Further, the data have been checked for reliability. These results can be found in section 3.7. A Reputation Quotient has been calculated for each company from the means of each respective question for each respective company. To put the answers in relation to the population sampled, the Reputation Quotient has been analysed in relation to the gender, age, and respective school of the respondents, and whether the respondents could consider working for the respective companies. This creates more depth and reflection upon the mean values and ranks presented as the Reputation Quotient in the league table. An analysis without the purpose of making statistical inferences has been performed in order to highlight patterns, similarities, and differences for each company as well as in relation to the other companies. The analysis concerning the age control variable did not show any noticeable difference in any aspect, and is therefore not presented or dealt with in the analysis.
24
The perceptual answers have also been analysed separately in relation to the corresponding answers for each respective company. As to see where the relative differences in perceptions may lie between the respective companies, the aggregated RQ-measure has been disaggregated by each dimension and item. An analysis has been made, not only to compare the companies, but also to give an idea of the internal relation for each respective dimension of the RQ. The nominal data questions (Q1, Q16) have been analysed in relation to the findings and analysis of the questions regarding the corporate reputation and the actual RQ findings of the companies. This has given another dimension and perspective to deepen the analysis on the companies.
25
4
Results and Analysis
This chapter provides the reader with descriptive statistics that present the results of the empirical findings. Further, the Reputation Quotient will be analysed both from an aggregated as well as disaggregated perspective.
4.1
The Reputation Quotient
Reputation Quotient 51,79 52 51 50 49 48 47 46 45 44 43 42 Adecco Manpower Poolia Academic Work 45,9 RQ 47,97 48,07
Figure 4 - The total Reputation Quotient
In this study, the corporate reputations of four staffing and recruitment companies have been calculated. 125 graduating business and engineering students from Jönköping University have been questioned in total. Through convenience sampling a set of 31 survey answers have been collected regarding Adecco, Poolia, Academic Work, and Manpower - having one more respondent. These findings are summarised and the companies are ranked in a league table (table 4-1). The values presented in the league table are derived from the perceptual responses on questions CR2-CR15 in the questionnaire answered by the graduating students (see Appendix 6). The individual responses to the questions have been assigned the values 1-5. The values that make up the league table are the calculated mean values on each question, per company. The sum of all mean values for each respective company makes up the Reputation Quotient, which is used to rank the companies in relation to the others. The Reputation Quotient is also displayed in figure 4. Since the responses have been assigned the value 1-5, the RQ can theoretically range from 14 to 70. As can be seen in figure 4, the RQ differs from each of the four investigated companies. The findings show that Academic Work is ranked highest out of the four staffing and recruitment companies, by the graduating students, with a total RQ of 51.79. Poolia and Manpower both have very similar RQ values of 48.07 and 47.97 respectively. Adecco is the
26
company that received the least positive RQ value with 45.9, when graduating students had their say about the corporate reputation of these four companies.
Table 4-1 - Company league table Companies Academic Work Poolia Manpower Adecco
CR2 3,5 3,64 3,55 3,38 CR3 2,92 2,62 2,46 2,61 CR4 3,74 3,8 3,52 3,68 CR5 3,73 3,63 3,52 3,3 CR6 3,76 3,64 3,64 3,52 CR7 3,78 3,09 3,44 3,04 CR8 3,4 2,88 2,85 2,88 CR9 3,9 3,38 3,5 3,35 CR10 3,47 3,52 3,4 2,95 CR11 4,05 3,55 3,64 3,73 CR12 3,44 3,1 3,27 3 CR13 3,8 3,7 3,63 3,24 CR14 4,04 3,64 3,55 3,52 CR15 4,26 3,88 4 3,7 RQ 51,79 48,07 47,97 45,9 RANK 1 2 3 4
A disaggregated league table of the RQ can visualise where the differences are, and in what aspects the companies are more or less positively perceived in respect to its competitors. This is presented in table 4-1. Academic Work has received the highest mean value in 11 of the 14 questions in the RQ scale. Further, table 4-1 shows that Adecco is perceived in the least positive way, in relation to the other companies, in 10 of the 14 questions.
Table 4-2 - Company standard deviation Companies Academic Work
CR2 0,812 0,86 0,767 0,703 CR3 0,83 CR4 0,752 0,78 0,995 0,7 CR5 0,631 0,804 CR6 0,437 0,816 1,021 0,869 CR7 0,85 0,565 0,793 0,852 CR8 1,046 0,727 1,056 0,783 CR9 0,436 0,733 1,189 0,9 CR10 0,874 0,881 0,757 0,823 CR11 0,524 0,941 0,768 0,602 CR12 0,814 0,73 1,099 0,767 CR13 0,676 0,858 0,745 0,766 CR14 0,976 0,831 0,866 0,797 0,571 0,638 0,945 CR15
0,619Poolia 0,77 0,667
1,154Manpower 0,893
0,683Adecco
Table 4-2 presents the standard deviation of each question for each of the four companies, which can explain to what extent opinions differ among the respondents on each question. The standard deviation is the most commonly used measure of summarising data dispersion. The standard deviation reflects the degree to which the values in a distribution differ from the mean. Standard deviation is intuitive in the sense that it is expressed in the original values of the data (Bryman, 1999). The deviations of the answers, if regarding all four companies, are ranging from 0,436 to 1,189. In reflection to the questions posted, the two questions with the lowest deviating values are CR6 and CR13. In respect to the companies, the respondents of Manpower seem to have the most diverse answers. However, as can be seen in table 4-2, the dispersion is fairly similar between most questions and no direct patterns can be found in it. Also in regard to the individual mean values presented for each questions in table 4-1, no direct connections are obvious between high or low mean values and high or low dispersion.
27
4.1.1
The Reputation Quotient separated by gender and school
RQ 54 51,97 52 50,13 50 47,89 48 46 44,34 44 42 40 47,07 46,93 49,4 51,2 Adecco Male Adecco Female Manpower Male Manpower Female Poolia Male Poolia Female Academic Work Male Academic Work Female
Figure 5 - Gender separated Reputation Quotient.
In figure 5, is displayed the perceptions about the four companies when separated by gender of the respondents. This is done to see if there are any gender differences in the perceptions towards the companies. The analysis is done to strengthen the findings of the overall RQ and to give an implication on how the scores are built up in relation to gender. These values have been calculated in the same way as the total RQ, but as separated by gender. Three out of four staffing and recruitment companies are more positively perceived by women than by men. Academic Work is the only company perceived as more positive by men than by women. Still, Academic is more positively perceived than all other companies for both men and women. If comparing Manpower and Poolia, it is surprising that Manpower has a higher value both when comparing men and women with respective category for Poolia, since Poolia received a higher total RQ (figure 4). This can however be explained by the fact that the sample of Poolia have an equal gender distribution whereas the Manpower sample have roughly 66 percent men and 34 percent women. This unequal gender distribution is also shared by the sample of Academic Work to about the same extent, whereas the sample of Adecco is equal in respect to gender (see Appendix 5). This uneven allocation is the result of a haphazard sampling, when distributing the surveys an effort was put into getting an equal amount of responses on each of the companies but not regarding gender distribution. The larger proportion of men and their less positive opinions have a negative impact on the total RQ. This therefore indicate that if the study was conducted repeatedly with a more equal gender distribution throughout the companies the result could have been slightly different result, and it is possible that Manpower would be ranked better than Poolia.
A de cc o M A a l e d e cc o F M em a l a n e po M we r M a n a l po e we r F m e a l P e oli o a M P a l o oa A e li ca F e e d m al A m ic a W or e c de k M mi a l cW o r e kF m e a l e
28
In general, women seem to have less difference in how they perceive the companies than the men. The values for women range from 47.89 to 51.2 whilst the vales for the men range from 44.34 to 54.97. These findings imply that the overall RQ is positively affected by the women in three of the four cases, whereas in the case of Academic Work thepositive effect on the overall rating comes from men.
RQ 54 52 50 48 46,45 46 44 42 A de cc o JIB A S de c c o J M TH a n o p we r JI M B S a n o p we r J TH P o oa li J I BS P A o oa ca li de J T mi H W c A r o ca k JI de B S mi cW o r kJ TH 45,78 47,29 49,84 48,9 47,66 48,62
53,02 Adecco JIBS Adecco JTH Manpower JIBS Manpower JTH Poolia JIBS Poolia JTH Academic Work JIBS Academic Work JTH
Figure 6 - School separated Reputation Quotient.
Figure 6 shows the graduating students' perceptions of the four companies in regard to which school the respondents are enrolled in. The school RQ values are calculated the same way as the total RQ, but divided by schools. It is the sum of the mean values, derived from all questions of the RQ-scale, per company, per school. Students at JTH - the engineering school, perceive three out of four companies in a more positive way than students at JIBS - the business school. Also in this regard, Academic Work is the company that differs from the other companies - being more positively perceived by JIBS students than by JTH students. JTH values are less diverging between the companies than the JIBS values, whereas the value of Academic Work is the one that deviates the most from the other companies. As compared to the other values of the RQ findings, JIBS students' opinions on Academic Work differ quite much from both JIBS students' perception of remaining companies and JTH students' perceptions of all companies. Still, if looking at JTH students' perceptions of Academic Work, they are more positive than both schools' students' perceptions on any other company. If looking at the values of the schools separately, the JIBS values correspond to the ranking found in the total RQ (as presented in figure 4). This, however, is not true for the JTH values, in which Manpower have received a higher RQ value than Poolia, thus being ranked as number two. The samples of the four companies have a fairly equal share of students from each school, except for Adecco which sample is composed of 61 percent from JIBS and 39 percent from JTH. Since the JIBS students have been more negative than the JTH
29
students towards three of the four companies, it can indicate that Adecco, would have been perceived more positively if the distribution would have been 50 percent sampled form the two schools. However, since the values from the two schools are relatively similar, the effect of having equal share from both schools is likely to be similar to the total RQ. Since the share of respondents at JTH were allocated as 39 percent and 61 percent at JIBS, it implies that each individual respondent at JTH influence the JTH-score to a larger extent than each individual respondent at JIBS. It might be coincidental that the JTH students are more positive than the JIBS students in three of the four cases, or there might be an undetected flaw connected to having fewer respondents in general, leading to a higher score. This problem is also present in the gender divided RQ, due to the uneven allocation between men and women. The explanation of the problem can be found in the sampling method. As the respondents were sampled in classes, this was the determining factor of how many people were sampled from each school and the gender distribution. Further, when eliminating the responses that were not representative for the population the largest share of the cases turned out to be from women. An observation made during the sampling process was that the classes at JTH tended to be smaller than the classes at JIBS; this was a contributing reason that more people were sampled from JIBS. As in the separation of RQ by gender, the distribution of the perceptions was more unequal among the JIBS students than for JTH students, ranging between 45.78 to 53.02 and 46.45 to 49.84 respectively. This also implies that overall for the study, the respondents that have been most positive are the men enrolled in a program at JIBS answering surveys on Academic Work, whereas, the male respondents at JIBS answering surveys on Adecco was found to be most negative. As Academic Work perceived most positive at both JIBS and JTH and by both women and men, the findings imply for the same as for males, that women at JIBS are most positive towards Academic Work. Women at JTH have also given a positive perception towards Manpower. The women at JIBS and JTH were both most negative towards Adecco, however, the results imply that women over all were more positive towards Adecco than the men. This discussion is, however, only an assumption. In order to get an accurate estimate of the situation, the separation has to be done dividing up the result by company, gender, and school originating from the total RQ. In this study such a separation would be biased, the share of responses per company divided by both gender and school would yield a too small number of respondents for each measure, and no indications could have been concluded.
30
4.1.2
Disaggregated Reputation Quotient
Disaggregated RQ
4,5 4 3,5 3 Adecco 2,5 2 1,5 1 0,5 0 CR2 CR3 CR4 CR5 CR6 CR7 CR8 CR9 CR10 CR11 CR12 CR13 CR14 CR15 Manpower Poolia Academic Work
Emotional
Workplace
Products/Services
Finance Vision/Leadership
Social
Figure 7 - The disaggregated Reputation Quotient.
Figure 7 describes a disaggregated view of how each of the companies is perceived in regard to each question of the RQ scale. Each value on each question is the calculated mean value of the responses on each company. As can be seen in the graph, the responses in regard to the companies more or less follow a pattern of peaks and pits on each respective question. The most deviating company is Academic Work, which on several questions (CR3, CR7, CR8, CR9, CR11, and CR14) clearly has a more positive mean value than the other three companies. As the original Harris-Fombrun Reputation Quotient consisted of 20 items divided into 6 dimensions, this structure was also used in this study. As can be seen in (table 31) in section 3.2.2 the dimensions are Emotional Appeal (CR2- CR4), Workplace Environment (CR5-CR6), Products and Services (CR7-CR10), Vision and Leadership together with Financial Performance (CR11-CR13) and Social Responsibility (CR14- CR15). For each dimension, most of the items were reformulated, as can be seen in the table 3-2. This was done for enhancing clarity for the respondents. Some of the items were altered or deleted to get a better fit to the population investigated. As this was done, the dimensions still remained and the altered items could be placed into the same original dimensions. However, as Financial Performance and Vision and Leadership were found to have least fit to the population and setting, a few of the items were removed completely, and therefore these dimensions were merged into one (CR11CR13). If looking at the responses on the Emotional Appeal dimension, the first eye-catching difference is that overall the respondents for the four companies have answered positively on the question on whether they respect the company, but none of the companies were admired by the respondents to a larger extent. In the original Harris-Fombrun RQ item, these two aspects were put together; if that would have been done in this setting theanswers would have been misleading. Overall, all companies are positively perceived in this dimension with only small differences, but not on the question on admiration where all the companies were perceived more negatively. The relatively large difference between this
31
question and the others in the dimension raises the question whether it is suitable as to be included in the dimension. On the dimension of Workplace Environment, the results show that the companies are quite equally perceived, and CR6 regarding whether the company has a well functioning organisation, is the aspect where the mean values are most equal in the study overall. One could argue that in the eyes of the investigated stakeholder group none of the companies have a strong competitive edge on this aspect, as they are perceived similarly. The third dimension concerning Products and Services was also ranked as an important factor when looking at a company (Q1). On the three first questions (CR7-CR9), Academic Work distinguishes itself from the rest of the companies. Academic Work is, compared to the other companies, more positively perceived when it comes to providing good job offers and being an attractive employer as well as creating a good merit of employment. If the companies would be interested in investigating this target group, it can be suggested that the Products and Services factors as they are stated in the questionnaire are the most important for getting a picture of how they are perceived by a possible employee. However, on the question of whether a company was using innovative marketing strategies, three of the four companies were perceived similarly with Adecco being perceived more negatively. This question, however, does not necessarily imply that the marketing efforts of the company are bad, as the question refers to the innovativeness of the companies marketing efforts. Further, it can also be argued that this aspect could have been reformulated to make it easier for the respondents to take a stance or to give the respondents a possibility to distinguish between good and bad marketing. The statement could have been formulated in terms of effectiveness instead of innovativeness. The next dimension investigated is Vision and Leadership together with Financial Performance. All firms were found to have a high score on the question on future prospects, which is valuable information to the industry as a whole. Academic Work outperformed its competitors in this aspect as it was ranked relatively higher than the other three firms. Further, Academic Work also got the highest score on the question if the company tends to outperform its competitors. In this question all three companies received similar scores and Manpower, which is ranked the third, got a better score than Poolia, ranked the second in the overall RQ measure. The respondents found all firms to be exploring the market opportunities. However, Adecco was found to have a significantly lower score than its competitors investigated. The last dimension investigated was Social Responsibility, where all the companies scored relatively high values, and all firms were assigned to be significant to the job market and fulfilling an important role in society. CR15, concerning if the company was important to the job market, received the highest scores for three of the investigated companies compared to the other items measured.
32
4.1.3
Analysis of "have no opinions"
Adecco
16,13%
Table 4-3 - Table of non-response
Questions
CR2 CR3 CR4 CR5 CR6 CR7 CR8 CR9 CR10 CR11 CR12 CR13 CR14 CR15
Manpower
9,38%
Poolia
9,68% 16,13% 19,35% 12,90% 29,03% 25,81% 19,35% 16,13% 25,81% 29,03% 32,26% 35,48% 19,35% 22,58%
Academic
1W,13% 6 ork 25,81% 12,50% 19,35% 12,90% 32,26% 25,81% 16,13% 16,13% 35,48% 29,03% 38,71% 32,26% 12,90% 12,90% 9,38% 15,63% 21,88% 21,88% 15,63% 12,50% 21,88% 21,88% 31,25% 25,00% 3,13% 3,13%
22,58% 25,81% 29,03% 45,16% 25,81% 35,48% 32,26% 45,16% 38,71% 48,39% 51,61% 25,81% 25,81%
The cases where the respondents answered have no opinion on question (CR2-CR15) have been analysed separately. This separation was made in order to see to what extent the respondents had no opinion on each question, and how this has effected the measuring of the RQ. As discussed in the section 3.5.1, this choice was made to be able to distinguish between neither disagree or agree and have no opinion, which turned out to give quite extensive effects on the results of the study. While screening data for errors it became obvious that a fairly large number of respondents marked have no opinion and the decision was made to analyse these findings separately. Table 4-3 presents these results as percentages of have no opinion answers out of the number of respondents on each question, per company. The high rate of have no opinions can also be an outcome of pattern answering (as discussed in 3.5.1). However, there is no implication that this would affect any specific company since all respondents are as likely to fall into pattern answering. Pattern answering is, however, a possible bias which cannot be disregarded, since it may affect the overall result of the study.
60,00%
50,00%
40,00% Adecco Manpower 30,00% Poolia Academic Work
20,00%
10,00%
0,00% C R 2 C R 3 C R 4 C R 5 C R 6 C R 7 C R 8 C R 9 C R 1 0 C R 1 1 C R 1 2 C R 1 3 C R 1 4 C R 1 5
Figure 8 - Graphical overview of non-respondents.
33
Figure 8 visualises the same findings. The graph shows that the questions on all four companies follow a pattern of have no opinion on each respective question, except on CR8 where the respondents for Academic Work marked a higher level of have no opinion than on the previous question, whereas the respondents for the other companies marked a lower level of have no opinion than on CR7. The have no opinion answers reaches its peak in CR12 and CR13. This could indicate that these questions have least fit to the population, so if the questions where to be altered again, one could consider to exclude them or re-formulate the statements. As the calculation of the Crombach Alpha exclude the have no opinion answers, these type of flaws are undetected in that calculation. This observation is significant to the overall reliability of the measurement of the RQ. Further, Academic Work has a higher degree of have no opinions in 13 out of 14 questions, which is exceptional compared to the other companies investigated. In reflection to this remark, the question arose - to what extent can this high degree of have no opinion answers, not included in measurement of the RQ, have affected the RQ? If, for example, the have no opinion option would not have been included as an option. Making the assumption that the respondents answering have no opinion would have answered neutral (neither agree nor disagree) if no such alternative would have been included in the questionnaire, this would have negatively affected all companies' RQ values. This is particularly interesting, considering Academic Work clearly having more positive RQ values than the other companies, but also having evidently higher degree of have no opinion answers. In order to see to which extent this would have affected the other findings, another analysis was made - namely to replace all have no opinion responses with the neutral neither agree nor disagree. These findings are presented below in figure 9.
Reputation Quotient when "no opinions" = 3
49,00 48,50 48,00 47,50 47,00 46,50 45,74 46,00 45,50 45,00 44,50 44,00 Adecco Manpower Poolia 47,06 46,68
48,52
Altered RQ
Academic Work
Figure 9 - The altered Reputation Quotient.
As can be seen, overall all findings are rather similar as those of the original total RQ (figure 4). All companies have, however, been affected negatively in relation to the total RQ with some variation in effect. Academic Work still deviates clearly as more positively
34
perceived from the other companies and the negative impact from the additional neutral values was not very large, when comparing the "new" values of Academic Work with the "new" values of the other companies. One difference however, from the total RQ, is that Manpower has changed relative rank position with Poolia, and is now ranked as second. The two companies have, however, in both the total RQ and the altered RQ, similar values. This alternative analysis validates the findings in the total RQ, and thus the choice to include a have no opinion response option in the questionnaire. Further, the possibility to separately be able to analyse the cases that marked the have no opinion response have shed light on to which degree the respondents have opinions at all regarding the companies. 4.1.4
60,0%
The relation between corporate reputation and employment
Comp Manpower Adecco AW Poolia
50,0%
40,0%
Percent
30,0%
20,0%
10,0%
0,0% Yes No
Work for company?
Figure 10 - Willingness to work for the companies.
Figure 10 displays the findings from the question "Would you consider working for company X after graduation?" - Q18, in the questionnaire. 50 percent of the respondents of both Adecco and Academic Work would consider working for the companies after graduation. Roughly a 60 percent share of the respondents of Poolia and Manpower is positive towards working for the companies. Manpower and Poolia have been assigned similar values throughout the study, and therefore the findings regarding these companies are not surprising. Of note is that the Adecco respondents are as positive towards working for Adecco as is the Academic Work respondents towards working for Academic Work. If comparing these findings to the previously stated results, it might look surprising that the respondents are not more positive towards working for Academic Work, it clearly being the most positively perceived company among the four and furthest away from Adecco. The implications that can be drawn from this are that either there is a measurement error in either of the measure of RQ or in Q18. There can be some missing factors in corporate reputation that have not been detected in this study that could give another result than the one just presented. If that is
35
not the case, the results might imply that corporate reputation is not directly linked to whether a person is willing to work for the company or not. If neither is true, then one could argue that corporate reputation is a bad measurement for perceptions for this stakeholder group, and other incentives should be measured instead. That, however, is not part of the purpose of this study to examine. Further, to see what single factor the respondents ranked as most important when determining how they perceived a company, one question (Q1) was included into the questionnaire. This question was not specific for the staffing and recruitment industry. However, the results are vivid; reputation is the most important factor when perceiving a company. The findings can be seen in Appendix 1. This finding can also help identify the relation between the dimensions that build up the corporate reputation. However, one can not ascribe it a weighting since the respondents were not asked to weight the questions just pick the alternative that they found most important. If that would have been possible it would have been in line with the free-descriptive method presented by Bromley (2002). These findings do not have to indicate that reputation is important when looking for a job, but rather to give an overall evaluation of a company. Therefore, companies will have to take this into consideration when targeting the investigated stakeholder group. However, since reputation scored highest in question one it gives more strength for corporate reputation being an important field of investigation for companies that want to find out how they are perceived by their stakeholders. To add an initial aspect to the measurement of corporate reputation, Bernstein's model of reputation creation has been used for investigating the source of the perceptions that the respondents hold towards the companies, (Q16). Bernstein's model is based on the definition by Bevis that the reputation creation is "the net result of the interaction of all the experiences, impressions, beliefs, feelings and knowledge that people have about a company"(cited in Davies et. al., 2003 p. 63). Bernstein elaborates further on how these impressions are created and the factors are presented in (figure 3) on how reputation is created. These impressions have been altered to get a better fit to the respondents. Bernstein also presents three different "filters" that the impressions have to pass through before becoming a perception: the filters of country of origin, industry, and company. These filters most likely have an impact on the perceptions but are not a subject of investigation in this study and will be disregarded. Instead, only the source of the perception is investigated. As can be seen in Appendix 2, the most important factor in this study turned out to be advertising, which shows to be the highest occurring factor in three of the four cases. Slightly below 40 percent of the respondents answering surveys for Academic Work and Poolia picked advertising as the most important factor for perception creation. Manpower and Adecco have about 25 percent and 15 percent respectively on the same question. For Adecco, word of mouth has been the most important for the perception creation. Adecco also has a high level of other answers than the ones that were suggested. However, these answers follow a clear pattern: the respondents had no idea of what type of business Adecco is running. The finding that a high level of respondents picked advertising as the most important factor can indicate that the companies marketing efforts are the most important factor when it comes to how corporate reputation is created. However, word of mouth is an important factor and can also contain a lot of information that is unknown in this study. It can be both negative and positive information that have been transferred from person to person. In word of mouth there is also room for peoples' interpretation of, for example, marketing that has been transferred to the respondents sampled in this study by others.
36
Conclusion
In the following chapter the authors present conclusions that have been drawn from the analysis. These conclusions are related back to the purpose. This study will investigate the corporate reputation of four staffing and recruitment companies, as representatives of the industry, from graduatingstudents'perspective. The purpose of this study has been fulfilled by measuring the corporate reputation using the methodology of the Reputation Quotient. The Reputation Quotient (RQ) is a scientifically grounded method for measuring corporate reputation, which validates the findings. To measure such a vaguely defined concept as corporate reputation brings not only definition ambiguity but also measurement concerns. Both the definition and measurement is something that have been widely debated in academia - a debate that has influenced and supported the altering of the method in order to fit the investigated population and subject. A quantitative survey has been designed to gather the empirical data from 125 graduating students, not in order to generalise to a greater degree but nevertheless to give a good estimation of the perceptions of the population of interest. Four corporate reputations have been measured for four staffing and recruitment companies. SPSS have been used to analyse the data and descriptive statistics of various kinds have presented the findings in this report. Analyses of the RQ in relation to characteristics of the samples have been performed in order to reinforce the findings and to put them in various perspectives. In addition, a disaggregated analysis of the RQ components has confirmed the total RQ findings. Additional investigations on how the students' perceptions have been created and whether they would consider working for the companies have been performed, putting the RQ findings in context. The reliability testing of the scale, which lies behind the RQ results, confirmed that the scale was also reliable after alteration of the original RQ. The findings entail that graduating students ascribe Academic Work the most positive corporate reputation out of the four. Poolia is ranked second with a small marginal to Manpower which is ranked third. Adecco is the company that received the least positive RQ value when measuring how graduating students perceive these four staffing and recruitment companies. No evidence has supported that a more positive reputation implies an increased willingness to work for the company, at least in the context of staffing and recruitment companies. Due to what is measured, namely perceptions, it is rather bold to make statements of the population as a whole. Instead, the findings can be seen as estimates of the companies reputations through the eyes of the graduating students at Jönköping University in the fields of business and engineering. However, it can be argued that one could make some claims on the population since the sample size is large relative to the population. Closely linked to the generalisability issue is validity; to have a high level of external validity, the findings should be applicable to another setting. For this study it has not been the aim to apply the findings to other populations than to the one investigated. This limitation has been imposed due to the many external factors present that can affect the creation of perceptions. These are not all incorporated to the corporate reputation measurement and, therefore, are even harder to detect since they fall outside the frame of the study. Further,
37
validity is also linked to the sampling method. For this study the there is a sampling bias present due to that only the population attending classes is sampled, leading to an underrepresentation of students not attending classes. Also, there is an uneven gender and school allocation that affects the validity negatively. To ensure that the measurement conducted is reliable, four Cronbach Alpha values have been calculated and shown positive results for the measurement. Further, since the nonresponse is disregarded in this calculation, a separate analysis has been conducted to enlighten the implications of this. As it turned out, there were two items that showed a high level of non-response, indicating that these statements need to be revised to get a better fit to the population investigated if the study was to be re-tested.
38
5
Discussion
The concluding chapter presents a discussion on the subject with the authors' personal thoughts and opinions. Further, the authors highlight the problems with the method and thus help the reader to understand the precautionary measures taken to help contribute to thoughtful and justifiable conclusions. There is also a part that reflects upon the thesis's limitations, as well as the authors' recommendations for further studies and readings. There have been many attempts by researchers to find a valid measurement of corporate reputation. Some of the measures have been adapted and used more than others. As discussed in earlier sections, the Fortune MAC is probably the most widely known, whereas the Harris-Fombrun Reputation Quotient might be seen as the most accepted. The RQ is, however, far from perfect and Wartick (2002), among others, have appointed critique to the measure. These issues have been dealt with in the initial stages of the study process. When looking at the findings of this study, they can seem rather intuitive. However, it would be hard to see the value of a measured RQ when not seen in relation to competing firms. This is the problem when quantifying feelings and perceptions as the depth behind why the respondents perceive the firms as they do disappear from the results. The results received can, on the other hand, easily be ranked and compared to other companies and thus one can decide which company has the best perceived corporate reputation. One could say that measures of corporate reputation yield valued outcomes only when put in relation with other firms within the same industry. It can therefore be argued that the only ones that should be interested in measuring corporate reputation are firms that are interested in how they are perceived in relation to its' competitors. This opinion is already included in some definitions but put in relation to the effort spent by researchers trying to find a universal measurement of corporate reputation; one can question where the academic value can be found in that type of research when, in practice, corporate reputation is only valuable to rank a firm in relation to their competitors. Would it not be of more interest for companies and academia, if more effort was spent on what actually can be made to improve corporate reputation? This does, however, require a clearly stated definition which can receive the gathered attention of academia. This brings us to the discussion on the definition of corporate reputation. The researchers are as divided in the question as in the issue of how to measure the concept. It has been discussed by several researchers that the assessment of corporate reputation includes the perceptions of all the firm's constitutes, indicating an aggregated perspective. Others have appointed harsh critique to this view arguing that a good assessment can only be collected if a disaggregated perspective is taken. There are also separate opinions on whether corporate reputation is a concept measured over time or as a snap shot of the reality. This ambiguous stance from academia on the definition of corporate reputation makes it difficult to know from where to take off for measuring the concept. To add credibility to the field overall, a canonical working definition needs to be established. This would enhance the possibility to compare the results from two studies. As it is today, one can not be certain if the researchers are measuring the same concept or two different concepts with the same name. Another concern that could have given the study another level of credibility would have been if weightings were included. As Bromley (2002) suggests, a free-descriptive method would have given the respondents the possibility to weight the dimensions of the RQ measure after their choice of interest. In this way further analysis could have been made on
39
what items and dimensions are considered essential for how to perceive a company. This could have given the study a different outcome. For example, if most of the factors where Poolia was ranked higher than Manpower were assigned low importance, and all the questions where Manpower was ranked higher than Poolia were assigned high importance, it would have given Manpower a boost in the overall perception of the company. 5.1.1 Limitations
There are many flaws to the concept and measurement of corporate reputation. To ensure that this has been regarded, it is important to ensure that a solid psychometric framework is used. As Bromley suggests, an ideal process to attain a good scale is to gather a fair amount of attributes that are relatively independent from each other but still correlated to the overall score or criterion. As this study is based on measuring the RQ, even though it is argued to fit all stakeholder groups, one should closely think of the purpose of the research and what, and whom, is about to be investigated. In this study the scale was altered to fit the investigated population to a larger extent than the original items in the RQ measurement. This way the measurement validity is increased and the scale actually measures what it is supposed to measure, namely the perceptions of graduating students at Jönköping University in the fields of business and engineering. In this way a better psychometric framework is assessed, which suites both subject and population sampled. There is however, a problem connected to the amount of items used for the investigation. Some researchers imply that the amount of respondents have to be up to 20 times the items investigated in order to have a sample big enough to draw conclusions. Thus, this number tends to vary from researcher to researcher, and some even claim that five respondents per item are sufficient. As convenience sampling was used for this sample, the sample was based on how many responses that could be collected, which turned out to be 125. A suggested analysis of dimensions, such as the ones in the RQ measure, could have been properly performed with factor analysis. However, as Tabachnick & Fidell (2001) suggests, for factor analysis 300 responses is needed, and no less than 150 (cited in Pallant, 2005). Therefore, the usage of factor analysis was not made possible, instead a disaggregated assessment was made to analyse the dimensions of the RQ in order to get a multilateral analysis of the RQ. Some of the issues that remain are the underrepresentation of students that are not present at lectures where the population was sampled and that the gender distribution for the sample is not representative for the population as a whole. If addressed, these issues could have given the study another overall result. Further, the allocation of students between JIBS and JTH has also been disregarded. This could be of great importance when wanting to generalise for the population as the distribution should be representative for the population. The issue of gender and school representation has been disregarded in this study due to the constraints of the sampling method, implying that the sample is not fully representative for the population. To ensure that the distribution is equal, one should use another sampling method and approach the students in another way. Some researchers would also argue that random sampling is the only credible way of sampling if generalisation is to be made on the population. This awakens another issue connected both to reliability and validity; if these results are applicable for another setting than the one just examined. What is measured in the end is perceptions which are individual and very dependent on the individual experiences of the respondents. One might be able to see the same patterns if the study was repeated with other respondents or another time. As Balmer & Stephen (2003) suggests reputations tend
40
to be stable but can change significantly due to external or internal events such as industry disasters or ethical lapse by the company. This indicates that this study could yield the same result if conducted within a near future, unless there is some striking change in the staffing and recruitment industry. One can also argue that the findings then would give about the same results and therefore imply that the findings also could be generalised for the graduating students at Jönköping University in the fields of engineering and business. The generalisation can, however, not be made for a setting outside Jönköping since the companies investigated can profile themselves differently in another city. 5.1.2 Implications
The results of this study imply that Academic Work is perceived in the best way in terms of corporate reputation and Adecco as the worst, of the four staffing and recruitment companies. However, Academic Work was also the company with the highest rate of have no opinion answers implying that the company may in fact be the least recognised company and thus has to improve the awareness of the company. However, a high rate of have no opinion was also recurrent for the other companies, which might imply that all the companies should appoint a lot of marketing effort to create high awareness among this important stakeholder group investigated. This statement is also backed up by the findings, showing that the most important factor for creating perceptions was advertising. This can also imply that advertising also is the most effective way to reach out to this stakeholder group. However, word of mouth, scored as the second most important factor, implying that one person's experience have great potential to travel far and wide and reach many ears. Therefore the companies should be aware that all communicated messages from the firm are in line with what is desired. Thus, employees are crucial, since their experiences can also be told to the investigated stakeholder group that has the potential to become employees. However, it is also important to state that in this study there was found no relation between whether one would like to work for a company and a good corporate reputation measured in RQ. Therefore, the companies should also look at other means to investigate how to attract this stakeholder group to become an employee. As for the graduating students, this study might not be able to provide such good information about what company to work for since corporate reputation is mostly a measure for companies. However, the investigated population as well as others who have interest in the staffing and recruitment companies can be informed of how these four companies are perceived in relation to each other by graduating students. 5.1.3 Further Research
During the process of this study a lot of flaws with the concept and measurement of corporate reputation have been detected. The first suggestion for further research is to thoroughly investigate corporate reputation as a concept to come up with a valued definition that could eliminate the root to ambiguity in the research field as it is today. This type of study will also help clarify whether corporate reputation is a concept that is measured over time or if it is a snap-shot of reality. In order to give a multilateral picture to the measurement of RQ, several different aspects have been taken into consideration for this study. However, if the study were to be repeated, it is suggested to apply weightings to the items of the RQ measurement or any scale of measurement used, as to enhance credibility to the measurement of corporate reputation. This thinking has been inspired by Bromleys' (2002) ideas on free-descriptive method.
41
References
Abratt, R. (1989). A new approach to corporate image management process. Journal of Marketing Managemant, 5, 63-76. Almega. (2008-02-14). Bemanningsbranschen uppåt för tredje året i rad. Retrieved 2008-02-14, from http://www.newsdesk.se/pressroom/almega/pressrelease/view/ bemanningsfoeretagen-uppaat-foer-tredje-aaret-i-rad-195869 Almega. (2008-01-29). Arbetssökande föredrar bemanningsföretag och jobbsajter före Ams. Retrieved 2008-02-05, from http://www.newsdesk.se/pressroom/almega/pressrelease/view/arbetssoekandefoeredrar-bemanningsfoeretag-och-jobbsajter-foere-ams-192609 Armstrong, M. (2006). Handbook of Human Resource Management Practice (10th ed.). London: Kogan Page Limited. Balmer, M. T. J., Stephen, A. (2003). Revealing the corporation: perspectives on identity, image, reputation, corporate branding, and corporate-level marketing. London: Routlegde Bemanningsföretagen. (2005-10-14). Bemanningsbranschen. Retrieved 2008-02-05, from http://www.almega.se/Files/ALMEGA/Caradoc_Members/PM/Kartläggninge_Be manningsbransc_A.DOC Bolander, L., & Nord, A. (1999). Arbetssituationen för anställda på bemanningsföretag eller hur mår "Nisse" egentligen? Master thesis in Business Administration. Linköping: Linköping University. Brace, I. (2004). Questionnaire Design. London: Kogan Page Ltd. Bromander, L. (2007). Uthyrning av arbetskraft - Ett rättssocialogiskt perspektiv på en flexibilitetsstrategi. Retrieved 2008-02-04, from https://gupea.ub.gu.se/dspace/bitstream/2077/7334/1/Lena%20Bromander.pdf Bromley, D. (2002). Comparing Corporate Reputations: League Tables, Quotients, Benchmarks or Case Studies? Corporate Reputation Review, 5(1), 35-50. Buglear, J. (2005). Quantitative Methods for Business - The A-Z of QM. Oxford: Elsevier Butterworth-Heinemann. Caruana, A. (1997). Corporate Reputation: concept and measurement. Journal of product & brand management, 6(2), 109-118. Collins, J., Hussey, R. (2n003). Business Research - A practical guide for undergraduate and postgraduate students (2d ed.). Hampshire: Palgrave Macmillan. Davies, G., Chun, R., Silva, D. R. V., Roper, S. (2003). Corporate Reputation and Competitiveness. Trowbridge, Wiltshire: The Cromwell Press. Dutton, J. E., Dukerich, J. M., Harquail, C. V. (1994). Organisation images and member identification. Administrative Science Quarterly, 39, 239-263. Englund, S., Nordström, L., Ohlsson, A. (2000). "Nisse" på besök i verkligheten. Bachelor thesis in Business Administration. Jönköping: Jönköping International Business School.
42
Fink, A. (1995). The Survey Handbook. London: SAGE Publications. Fombrun, C. J. (1996). Reputation: Realizing Value from the Corporate Image. Boston: Harvard Business School Press. Fombrun, C. J., Gardberg, N. A., Sever, J. M. (2000). The Reputation Quotient: A multistakeholder measure of corporate reputation. The Journal of Brand Management, 7(4), 241256. Fowler, J. F. (2002). Survey Research Methods (3rd ed.). Thousand Oaks, CA: SAGE Publications. Ghauri,rdP., Grønhaug, K. (2005). Research Methods in Business Studies. A Practical Guide. (3 ed.). Harlow: Pearson Educated Limited. Gomm, R. (2004). Social Research Methodology a critical introduction (1st ed.). Hampshire: Palgrave MacMillan Gotsi, M., & Wilson, A. M. (2001). Corporate reputation: Seeking a definition. Corporate Communications: An International Journal, 6(1), 24-30. Groenland, E. A. G. (2002). Qualitative Research to Validate the RQ-Dimensions. Corporate Reputation Review, 4(4), 308-315. Hillenbrand, C., & Money, K. (2006). Using Reputation measurement to create value: An analysis and integration of existing measures. Journal of General Management, 32(1), 1-12. Ind, N. (1997). The Corporate Brand. London: The MacMillan Press. Jacobsen, D. I. (2002). Vad, hur och varför? Om metodval i företagsekonomi och andra samhällsvetenskapliga ämnen. Lund: Studentlitteratur Johansson, M. (2006). Bemanningsbranchen - en kvantitativ studie av dess utveckling I Sverige. Master Thesis in Economics and Statistics. Växsjö: Växsjö Universitet. Litwin, S. M. (1995). How to Measure Survey Reliability and Validity. Thousand Oaks, CA: SAGE Publications. Moingeon, B., Soenen, G. (2002). Corporate and Organizational Identity: Integrating Strategy, Marketing, Communication and Organizational Perspectives. London: Routledge. Mahone, J. F., Wartick, S. L. (2003). Dealing with Stakeholders: How Reputation, Credibilty and Framing Influence the Game. Corporate Reputation Review, 6(1), 19-35. Olofsdotter, G. (2004). Bemanningsföretag - en möjlighet för vem? In Wolvén, L-E., Ekstedt, E. (Eds.), Företagande och gemenskap - små företags kultur, samverkan och konkurrens. (p. 149-168). Stockholm: Elanders Gotab. Pallant, J. (2005). SPSS Survival manual. (2nd ed.). New York: McGraw-Hill Education. Ruane, J. M (2006). A och O i samhällsvetenskaplig forskning. Lund: Studentlitteratur. Saunders, M. N. K., Thornhill, A., Lewis, P., McMillan, K., Weyers, J. (2007). Research Methods for Business Students. London: Pearson Higher Education.
43
Strbac, S. (2006). The staffing process for product development projects: A case study at 3P, Volvo AB. Dissertation in International Project Management. Göteborg: Chalmers University of Technology. Wartick, L. S. (2002). Measuring Corporate Reputation, Definition and Data. Business & Society, 41(4), 371-392. Zikmund, W. G. (2000). Business Research Methods (6th ed.). Orlando: Hartcourt Inc.
44
Appendices
Appendix 1 - Most important factor when looking at a company.
40,0%
30,0%
Percent
20,0%
10,0%
0,0%
Gut feel ng i Products/service Vision/leadership Workplace env s Fin results Social resp Carismatic repr Reputation Other
Most imortant factor Most imortant factor Frequency Valid Gut feeling Products/services Vision/leadership Workplace env Fin results Social resp Carismatic repr Reputation Other Total Missing Total System 14 27 11 6 6 2 3 43 4 116 9 125 Percent 11,2 21,6 8 ,8 4 ,8 4 ,8 1 ,6 2 ,4 34,4 3 ,2 92,8 7 ,2 100,0 Valid Percent 12,1 23,3 9 ,5 5 ,2 5 ,2 1 ,7 2 ,6 37,1 3 ,4 100,0 Cumulative Percent 12,1 35,3 44,8 50,0 55,2 56,9 59,5 96,6 100,0
45
Appendix 2 - Major Source in creating perceptions? The following graphs are the summarised answers from question 16. Not all of the response options have been picked and therefore do not show in the graphs. Academic Work
40,0%
30,0%
Percen t
20,0%
10,0%
0,0% Personal contact Media/internet Word of mouth Advertisment Other
Source
Manpower
25,0%
20,0%
Percent
15,0%
10,0%
5,0%
0,0% Personal contact Media/internet Word of mouth Advertisment Experience Other
Source
46
Adecco
40,0%
30,0%
Percen t
20,0%
10,0%
0,0% Personal contact Media/internet Word of mouth Advertisment Mail/email Other
Source
Poolia
40, 0%
30, 0%
Percen t
20, 0%
10, 0%
0,0% Personal contact Media/internet Word of mouth Advertisment Experience Other
Source
47
Appendix 3 - Reliability testing statistics Academic Work
Reliability Statistics Cronbach's Alpha ,841
N of Items 14 Item-Total Statistics Scale Variance if Item Deleted 22,750 24,444 28,278 27,028 29,750 27,528 29,028 30,028 28,250 26,194 27,000 30,694 29,861 29,500 Corrected Item-Total Correlation ,844 ,788 ,740 ,898 ,333 ,568 ,188 ,356 ,522 ,725 ,455 ,180 ,141 ,378 Cronbach's Alpha if Item Deleted ,799 ,806 ,823 ,813 ,839 ,825 ,857 ,838 ,829 ,815 ,834 ,845 ,857 ,836
Scale Mean if Item Deleted Good feeling Admire Trust Good company Organisation Offerings Att employer Good merit Marketing Future Competition Opportunities Society Job market 49,67 50,22 49,44 49,56 49,67 49,44 49,44 49,44 49,33 49,22 49,67 49,22 48,89 48,67
48
Manpower
Reliability Statistics Cronbach's Alpha ,893
N of Items 14 Item-Total Statistics Scale Variance if Item Deleted 47,829 48,981 45,400 46,114 49,067 51,667 46,695 52,114 49,524 50,543 52,410 51,410 48,952 49,571 Corrected Item-Total Correlation ,640 ,538 ,717 ,662 ,626 ,373 ,614 ,460 ,504 ,642 ,420 ,533 ,715 ,728 Cronbach's Alpha if Item Deleted ,882 ,887 ,878 ,882 ,883 ,894 ,884 ,890 ,889 ,884 ,891 ,887 ,880 ,881
Scale Mean if Item Deleted Good feeling Admire Trust Good company Organisation Offerings Att employer Good merit Marketing Future Competition Opportunities Society Job market 44,40 45,47 44,60 44,40 44,27 44,67 45,13 44,40 44,67 44,60 44,53 44,53 44,33 44,00
49
Adecco
Reliability Statistics Cronbach's Alpha ,665
N of Items 14 Item Statistics Item-Total Statistics Scale Variance if Item Deleted 15,302 16,681 17,764 16,841 15,610 16,264 17,187 14,066 22,555 15,104 16,335 17,940 16,335 15,786 Corrected Item-Total Correlation ,641 ,282 ,089 ,315 ,554 ,393 ,228 ,549 -,528 ,556 ,472 ,107 ,330 ,469 Cronbach's Alpha if Item Deleted ,601 ,649 ,677 ,645 ,612 ,633 ,656 ,597 ,775 ,605 ,627 ,670 ,641 ,621
Scale Mean if Item Deleted Good feeling Admire Trust Good company Organisation Offerings Att employer Good merit Marketing Future Competition Opportunities Society Job market 42,07 42,71 42,07 42,07 41,93 42,43 42,57 42,29 42,64 41,79 42,21 42,36 41,79 41,64
50
Poolia
Reliability Statistics
Cronbach's Alpha ,955
N of Items 14 Item-Total Statistics Scale Variance if Item Deleted 87,344 90,500 89,956 89,344 87,511 90,889 83,733 85,733 90,011 82,844 84,667 88,278 90,100 76,989 Corrected Item-Total Correlation ,853 ,718 ,718 ,903 ,749 ,648 ,702 ,792 ,757 ,973 ,814 ,896 ,550 ,880 Cronbach's Alpha if Item Deleted ,950 ,953 ,953 ,950 ,952 ,954 ,955 ,951 ,952 ,946 ,951 ,950 ,957 ,951
Scale Mean if Item Deleted Good feeling Admire Trust Good company Organisation Offerings Att employer Good merit Marketing Future Competition Opportunities Society Job market 43,30 44,50 43,20 43,30 43,20 44,00 44,20 43,80 43,70 43,80 44,00 43,50 43,90 43,90
51
Appendix 4 - Mean values and std. deviation
Academic Work
Descriptive Statistics N Good feeling Admire Trust Good company Organisation Offerings Att employer Good merit Marketing Future Competition Opportunities Society Job market Valid N (listwise) 26 24 23 22 17 23 20 21 17 19 16 15 23 23 9 Mean 3,50 2,92 3,74 3,73 3,76 3,78 3,40 3,90 3,47 4,05 3,44 3,80 4,04 4,26 Std. Deviation ,812 ,830 ,752 ,631 ,437 ,850 1,046 ,436 ,874 ,524 ,814 ,676 ,976 ,619
Manpower
Descriptive Statistics N Good feeling Admire Trust Good company Organisation Offerings Att employer Good merit Marketing Future Competition Opportunities Society Job market Valid N (listwise) 29 28 29 27 25 25 27 28 25 25 22 24 31 31 15 Mean 3,55 2,46 3,52 3,52 3,64 3,44 2,85 3,50 3,40 3,64 3,27 3,63 3,55 4,00 Std. Deviation 1,021 ,793 1,056 1,189 ,757 ,768 1,099 ,745 ,866 ,638 ,767 ,770 ,995 ,683
52
Adecco
Descriptive Statistics N Good feeling Admire Trust Good company Organisation Offerings Att employer Good merit Marketing Future Competition Opportunities Society Job market Valid N (listwise) 26 23 25 27 21 23 26 26 20 22 19 21 27 27 14 Mean 3,38 2,61 3,68 3,30 3,52 3,04 2,88 3,35 2,95 3,73 3,00 3,24 3,52 3,70 Std. Deviation ,852 ,783 ,900 ,823 ,602 ,767 ,766 ,797 ,945 ,703 ,667 ,700 ,893 ,869
Poolia
Descriptive Statistics N Good feeling Admire Trust Good company Organisation Offerings Att employer Good merit Marketing Future Competition Opportunities Society Job market Valid N (listwise) 28 26 25 27 22 23 25 26 23 22 21 20 25 24 10 Mean 3,64 2,62 3,80 3,63 3,64 3,09 2,88 3,38 3,52 3,55 3,10 3,70 3,64 3,88 Std. Deviation ,780 ,804 ,816 ,565 ,727 ,733 ,881 ,941 ,730 ,858 ,831 ,571 ,860 1,154
53
Appendix 5 - Sample characteristics
Academic Work
School Cumulative Percent 45,2 100,0
Frequency Valid JTH JIBS Total 14 17 31
Percent 45,2 54,8 100,0 Sex
Valid Percent 45,2 54,8 100,0
Frequency Valid Female Male Total 10 21 31
Percent 32,3 67,7 100,0
Valid Percent 32,3 67,7 100,0
Cumulative Percent 32,3 100,0
Descriptive Statistics N Age Valid N (listwise) 30 30 Minimum 21 Maximum 31 Mean 23,77
Manpower
School Cumulative Percent 43,8 100,0
Valid
JTH JIBS Total
Frequency 14 18 32
Percent 43,8 56,3 100,0 Sex
Valid Percent 43,8 56,3 100,0
Frequency Valid Female Male Total 11 21 32
Percent 34,4 65,6 100,0
Valid Percent 34,4 65,6 100,0
Cumulative Percent 34,4 100,0
Descriptive Statistics N Age Valid N (listwise) 31 31 Minimum 20 Maximum 42 Mean 25,06
54
Adecco
School Cumulative Percent 38,7 100,0
Frequency Valid JTH JIBS Total 12 19 31
Percent 38,7 61,3 100,0
Valid Percent 38,7 61,3 100,0
Sex Cumulative Percent 48,4 100,0
Frequency Valid Female Male Total 15 16 31
Percent 48,4 51,6 100,0
Valid Percent 48,4 51,6 100,0
Descriptive Statistics N Age Valid N (listwise) 31 31 Minimum 20 Maximum 28 Mean 23,61
Poolia
School Cumulative Percent 51,6 100,0
Valid
JTH JIBS Total
Frequency 16 15 31
Percent 51,6 48,4 100,0 Sex
Valid Percent 51,6 48,4 100,0
Frequency Valid Female Male Total 15 16 31
Percent 48,4 51,6 100,0
Valid Percent 48,4 51,6 100,0
Cumulative Percent 48,4 100,0
Descriptive Statistics N Age Valid N (listwise) 30 30 Minimum 20 Maximum 28 Mean 23,37
55
Appendix 6 - Questionnaires Below follows the questionnaires used for collecting empirical data. The data gathering was done by handing out 31 surveys to three of the staffing and recruitment companies and 32 surveys to the fourth (see section 3.3.2). Since the surveys were identical apart from the different company names; Adecco, Poolia, Manpower, Academic Work the authors have decided to simply include templates, and use the company name "X" for the English survey and the company name "Y" for the Swedish survey.
56
A study on the corporate reputation of Company "X"
This study will be used as empirical data for a bachelor thesis concerning corporate reputation. All respondents will remain anonymous. Thank you very much for participating!
For all following questions, please mark one alternative that suites your perception best.
Ex:
1. Which factor do you consider the most important in determining how you perceive a company? (Choose one factor, not specific for Company "X") "Gut" feeling Financial performance National origin Products and services Social responsibility Reputation Vision and leadership Charismatic representatives Other: Workplace environment
To what extent do you agree with the following statements? Strongly Disagree Disagree Neither agree nor disagree Agree Strongly Have no opinion
2. I have a good feeling about Company "X" 3. I admire Company "X" 4. I do not trust Company "X"
5. Company "X" looks like a good company to work for 6. Company "X" seems to have a well functioning organisation
7. Company "X" offers high quality work opportunities 8. Company "X" is an attractive employer 9. To work for Company "X" will yield a valuable experience 10. Company "X" does not use innovative marketing strategies
11. Company "X" has strong prospects of becoming more successful in the future 12. Company "X" tends to outperform its competitors 13. Company "X" does not take advantage of market opportunities
Please turn!
57
Stongly Disagree
Disagree
Neither agree nor disagree
Agree
Strongly Agree
Have no opinion
14. Company "X" has an important role in the society 15. Company "X" is important for the job market
16. What has been the major source in creating your above stated perceptions of Company "X"? (Choose one factor) Personal contact Direct mail, e-mail Media comments, internet Prior experience Word of mouth Other: Advertising
Are you currently enrolled in a program at:
J TH
JIBS
Not a program student
After graduation, where do you intend to look for a job? (feel free to choose more than one)
Sweden
Nordic countries (other then Sweden)
Europe (other than the Nordic countries)
Rest of the world
When do you expect to apply for a job?
0 - 6 months
6 - 12 months
12 - 24 months
Already have a job (fulltime)
Would you consider working for Company "X" after graduation?
Yes
No
Female
Male
Age:
Thank you very much for your participation and help!
Please state your e-mail address if you want to be informed of the results of this study:
58
En studie om företaget "Ys" anseende.
Denna studie kommer att användas som empirisk data i en kandidatuppsats om företags anseende. Alla svarande kommer förbli anonyma. Stort tack för er medverkan!
För följande frågor, välj ett alternativ som överensstämmer bäst med din uppfattning. Ex:
1. Vilken faktor anser du är viktigast när det gäller att avgöra hur du uppfattar ett företag? (välj en faktor, ej specifikt för företag "Y") Emotionell Utstrålning Finansiella Resultat Nationellt ursprung Produkter och Tjänster Socialt Ansvar Rykte Vision och Ledarskap Karismatiska representanter Annat Arbetsplats Miljö
Till vilken utsträckning håller du med dessa påståenden? Håller inte alls med Håller inte med Håller varken Håller med eller delvis inte med Håller helt med Har ingen uppfattning
2. Jag har en god känsla kring företag "Y" 3. Jag beundrar företag "Y" 4. Jag litar inte på företag "Y" 5. Företag "Y" verkar vara ett bra företag att arbeta för 6. Företag "Y" verkar ha en välfungerande organisation
7. Företag "Y" erbjuder högkvalitativa jobberbjudanden 8. Företag "Y" är en attraktiv arbetsgivare 9. Att arbeta på företag "Y" ger en god merit 10. Företag "Y" använder inte innovativa marknadsföringsstrategier
11. Företag "Y" har starka utsikter för att bli mer framgångsrika i framtiden 12. Företag "Y" tenderar till att överträffa dess konkurrenter 13. Företag "Y" utnyttjar inte marknadsmöjligheterna
Var god vänd!
59
Håller inte alls med
Håller inte med
Håller varken Håller med eller delvis inte med
Håller helt med
Har ingen uppfattning
14. Företag "Y" fyller en viktig samhällsfunktion 15. Företag "Y" är betydande för arbetsmarknaden
16. Vad har varit den största källan till din ovan angivna uppfattning av företag "Y"? (Välj en)
Personlig kontakt
Mediauppgifter, Internet
Ryktesspridning (word of mouth)
Annonsering, reklam
Direkt post, e-post
Tidigare egna erfarenheter
Annat
Jag är för närvarande inskriven på ett program på:
J TH Ej programstudent
JIBS
Var kommer du att söka jobb efter avslutad utbildning? (välj ett eller flera alternativ)
Sverige Europa (ej Norden) Resterande världen
Norden (ej Sverige)
När räknar du med att ansöka om ett jobb?
0 - 6 mån 12 - 24 mån
6 - 12 mån Har redan jobb (heltid)
Skulle du kunna tänka dig att jobba för företag "Y" efter din examen?
Ja
Nej
Kvinna
Man
Ålder:
Om du önskar att ta del av resultatet av denna undersökning, skriv din e-mail adress:__________________________________
Stort Tack för er medverkan och hjälp!
60
doc_671167722.docx