Effective information collection on international inbound visitors

Description
Data collection from inbound tourists is a repetitive activity. This paper’s main purpose is to
show that, unless something useful about the nature of change is being established, repetitious
collection of data from, for example, inbound visitors results in ineffective accumulation of data. The
paper also aims to elucidate what it means for data to be ineffective for practical application or theory
development.

International Journal of Culture, Tourism and Hospitality Research
Effective information collection on international inbound visitors
Wu-Chung Wu You-De Dai Hsiou-Hsiang J ack Liu
Article information:
To cite this document:
Wu-Chung Wu You-De Dai Hsiou-Hsiang J ack Liu, (2012),"Effective information collection on international inbound visitors", International
J ournal of Culture, Tourism and Hospitality Research, Vol. 6 Iss 1 pp. 54 - 69
Permanent link to this document:http://dx.doi.org/10.1108/17506181211206252
Downloaded on: 24 January 2016, At: 22:18 (PT)
References: this document contains references to 35 other documents.
To copy this document: [email protected]
The fulltext of this document has been downloaded 420 times since 2012*
Users who downloaded this article also downloaded:
J oseph O'Leary, Tzung-Cheng Huan, (2012),"International tourist behavior: an IJ CTHR special issue", International J ournal of Culture,
Tourism and Hospitality Research, Vol. 6 Iss 1 pp. 3-7http://dx.doi.org/10.1108/17506181211206289
Chia-Huh J oy Liang, Hung-Bin Chen, Ming-Yang Wang, (2012),"Units, populations, and valid analyses", International J ournal of Culture,
Tourism and Hospitality Research, Vol. 6 Iss 1 pp. 70-80http://dx.doi.org/10.1108/17506181211206261
Antónia Correia, Metin Kozak, J oão Ferradeira, (2013),"From tourist motivations to tourist satisfaction", International J ournal of Culture,
Tourism and Hospitality Research, Vol. 7 Iss 4 pp. 411-424http://dx.doi.org/10.1108/IJ CTHR-05-2012-0022
Access to this document was granted through an Emerald subscription provided by emerald-srm:115632 []
For Authors
If you would like to write for this, or any other Emerald publication, then please use our Emerald for Authors service information about
how to choose which publication to write for and submission guidelines are available for all. Please visit www.emeraldinsight.com/
authors for more information.
About Emerald www.emeraldinsight.com
Emerald is a global publisher linking research and practice to the benefit of society. The company manages a portfolio of more than
290 journals and over 2,350 books and book series volumes, as well as providing an extensive range of online products and additional
customer resources and services.
Emerald is both COUNTER 4 and TRANSFER compliant. The organization is a partner of the Committee on Publication Ethics (COPE) and
also works with Portico and the LOCKSS initiative for digital archive preservation.
*Related content and download information correct at time of download.
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
2
:
1
8

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
Effective information collection on
international inbound visitors
Wu-Chung Wu, You-De Dai and Hsiou-Hsiang Jack Liu
Abstract
Purpose – Data collection from inbound tourists is a repetitive activity. This paper’s main purpose is to
show that, unless something useful about the nature of change is being established, repetitious
collection of data from, for example, inbound visitors results in ineffective accumulation of data. The
paper also aims to elucidate what it means for data to be ineffective for practical application or theory
development.
Design/methodology/approach – The approach was to examine three years of data from inbound
visitors to Taiwan to consider what would make data collection more effective.
Findings – Collecting many speci?c variables relating to travel by inbound tourists can result in
recognizing segments and other matters important for applied research or theory development.
Analysis shows detailed information can have limited use and high cost when different details apply to
different segments. After identifying segments to study, effective information collection can require
segment speci?c questioning, special sampling and segment speci?c studies.
Originality/value – While various countries conduct special studies, annual collection of a wide variety
of information from inbound tourists is a common practice. This research provides new perspectives on
why some data collection practices should be modi?ed.
Keywords Survey data, Data collection, Effectiveness, Inbound, Taiwan, Tourism management,
Data handling
Paper type Research paper
Introduction
Data collection from inbound tourists to countries is a repetitive activity of many countries. For
example, one can see information on data collection from international visitors by the US at the
TourismIndustries (TI) website (e.g. see Travel Industries, 2009). For Canada, documentation on
data collection on international travel is posted on the website of Statistics Canada (2009).
Issues have been raised about both the USA’s and Canada’s data collection. In 2009, a meeting
regarding improving TI’s inbound data was convened by Frechling at George Washington
University (Frechtling, 2009). The meeting produced recommendations for improvement.
Meetings regarding problems with Canada’s international travel data have been a regular
occurrence resulting in changes in the survey to improve effectiveness (e.g. see Bradford et al.,
2001). This research uses Taiwan Tourism Bureau’s inbound visitor survey data for analysis of
matters involved in effective information collection. Though analysis is only of data collected
fromTaiwan’s inbound visitors in 2000, 2001 and 2003 (see Taiwan TourismBureau, 2001, 2002,
2004), results apply to other countries and to surveys of outbound and domestic travelers.
The strategy in this research is simple. Introduction of Taiwan’s data collection from
international tourists is followed by a general and then topic speci?c introduction to literature
relevant to data collection from international travels to a country. Presentation of hypotheses
focuses on data being collected using an appropriate amount of resources and actually
being useful in theory development or in planning, marketing and management. Analysis
PAGE 54
j
INTERNATIONAL JOURNAL OF CULTURE, TOURISM AND HOSPITALITY RESEARCH
j
VOL. 6 NO. 1 2012, pp. 54-69, Q Emerald Group Publishing Limited, ISSN 1750-6182 DOI 10.1108/17506181211206252
Wu-Chung Wu is an
Associate Professor in the
Graduate School of
Hospitality Management,
National Kaohsiung
Hospitality College,
Hsia-Kang, Taiwan. You-De
Dai is an Assistant
Professor at the Graduate
Institute of Recreation,
Tourism and Hospitality
Management, National
Chiayi University, Chiayi,
Taiwan. Hsiou-Hsiang Jack
Liu is an Associate
Professor in the Department
of Tourism Management,
National Kaohsiung
University of Applied
Sciences, Kaohsiung,
Taiwan.
Received September 2009
Revised April 2010
Accepted July 2010
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
2
:
1
8

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
elucidates problems that arise in using Taiwan’s data from exiting international visitors.
Discussion elaborates on problems in using data and on being effective in obtaining data
that meet needs for theory development, planning, marketing and management.
Conclusions cover practical and other implications of the research.
Research, literature, logic and Taiwan’s inbound data
This paper is general. Technicalities associated with data collection (e.g. sample size,
sample selection and observation weighting) are not examined. One can ?nd details on data
collection methods, sampling, etc. for the US, Canada and Taiwan at websites referenced
above or in references given. Yes, getting a representative sample matters in getting an
accurate picture of dollars expended nationally or of numbers of visitors with particular
attitudes and behaviors. However, this research accepts that if adequate resources are
devoted to determining speci?cs of information that is useful then logistics of obtaining the
data should be pursued. In some cases proper application of statistical theory can establish
viable data collection or establish that information acquisition should not proceed because,
for example, adequately accurate information would not be affordable.
Data collection
For Taiwan, the Taiwan Tourism Bureau (TTB) manages collection data from inbound
travelers. Taiwan Inbound Surveys (TIB surveys) are collected by statistical institutes.
Universities have institutes with which TTB can contract for data collection. In different years,
data are collected by different institutes. For 2000, 2001 and 2003, the Graduate Institute of
Applied Statistics of Fu-Jen Catholic University collected the data. Data collection in 2000,
2001 and 2003 resulted in information fromabout 15,000 respondents. Eliminating business,
conference and visiting friends and relation (VFR) visitors and only keeping data from origin
areas (countries or groups of countries) for which enough responses exist to support some
analysis, 4,080 records exist for leisure/sightseeing visitors. By year, numbers of
respondents are: 1,208 (2000), 1,624 (2001) and 1,248 (2003).
This research does not use 2002 data because differences in questionnaire structure and
other matters arose when data collection for 2002 was contracted to Shih Hsin University.
What appear to be subtle differences in survey methodology can have quite large impacts
(e.g. see Beaman et al., 2001). Data having almost the same variables is a different criterion
of similarity than data being collected using identical forms and procedures. Based on a
majority of variables being collected repeatedly, the Taiwan inbound data from 1995 to 2008
are similar from year to year, regardless of the university collecting them.
Table I gives the general categories of data found and speci?cs on some variables available
in 2000, 2001 or 2003. Over all years since the mid-1990s expenditure data has been
collected to allow estimation of tourist expenditures in Taiwan. In this research, examination
of expenditure data is not pursued since analysis would require considering meeting criteria
speci?ed by economists (e.g. see references in Sun, 2005). Furthermore, getting good and
ef?cient estimates with expenditure data is currently a topic in the literature (Pol et al., 2006;
Huan et al., 2008). Dealing with expenditure data is appropriate for a separate paper.
Table I does not make some matters clear regarding responding to TIB surveys. Some
questions are only asked of respondents in a certain segment. Questionnaire designers
recognize that some questions are only appropriate for a segment or some segments
(e.g. for business travelers or for conference attendees). This research deals with data
collected from leisure visitors. Without counting the answering of about 15 expenditure
questions, a sightseeing visitor implicitly gives responses for about 50 variables. Implicitly
refers to respondents supplying some responses while reading to note that many items in
lists (e.g. of activities or expectations) do not need to be checked.
Overview of literature relating to data to collect
This research does not provide extensive references on research relevant to matters raised.
Elaborate referencing is not necessary because the book Woodside and Martin (2008) and
VOL. 6 NO. 1 2012
j
INTERNATIONAL JOURNAL OF CULTURE, TOURISM AND HOSPITALITY RESEARCH
j
PAGE 55
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
2
:
1
8

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
volumes of Consumer Psychology of Tourism, Hospitality and Leisure (e.g. Vol. 2 edited by
Mazanec et al., 2001; and Vol. 3 edited by Crouch et al., 2004) provide overview articles with
extensive lists of references. However, research that merits mention based on inferences
made in this article is in March and Woodside (2005). The book is a good resource for seeing
how in depth interviews can be used in tourism research. One may be reluctant to base
theory or large investments in development on information from a relatively few interviews.
However, the book shows how in depth interviews can be used to establish perspectives on
behavior that facilitate appropriate data collection by establishing an understanding of
variables needed, of segments and of cases/respondents needed to have good information
for theory building and practical applications. Martin (2008) and DeCrop (2001) also cover
the matter of using in depth interviews in tourism research.
Research, logic and data collected
Those familiar with the TIB questionnaires (Taiwan Tourism Bureau, 2001, 2002, 2004) will
notice the organization of Table I does not entirely parallel question grouping in the
questionnaires used in setting up the table. The table’s organization facilitates its use in this
paper. Commentary in the paper proceeds in the order of the headings of the table except
for addressing demographics. The last item in the table is demographics. Demographic
Table I General categories of information in Taiwan’s inbound surveys collected for
sightseers with some speci?c information on variables for 2000, 2001 and 2003
General categories
Arrival/departure
Duration (nights); with whom (seven variables to allow checking off wife, child, relation, friends,
associates, etc.); travel arrangements (?ve categories from tour starting abroad to self organized
travel); conveyance (air or boat by location of interview)
Information and trip planning
Planning (e.g. started when, speci?c in?uences
a
)
Pre visit information sources (newspapers, etc. with 9 or more responses to check)
Information in Taiwan (not collected in 2000, 2001 and 2003 but, e.g. in 2002)
Information to be used at home (not collected in 2000, 2001 and 2003 but, e.g. in 2002)
Attraction to Taiwan (checking a category indicated a reason for coming)
General attractions (14þ variables – categories such as scenery, price of goods and cuisine)
Speci?c attractions (10 þ variables – e.g. Lantern Festival and white water rafting)
Purposes (primary and secondary purpose for coming – e.g. sightseeing, VFR, business)
Activities during visiting
Activity by category (15 þ variables – e.g. golf, shopping and hot spring-spa visits)
Scenic spots visited (10 þ variables to identify locations)
Meeting expectations/satisfaction (hotel satisfaction is listed with accommodation)
Taiwan meeting expectations (ten categories with ratings better, met or poorer)
Arrangements meeting expectations (5 þ categories for satisfaction with tour/guides)
Satisfaction affecting visiting (5 þ categories
a
covering visa processing, security, etc.)
Reactions to places visited (favorite, least favorite, reasons for these)
Intentions and history of visiting
Intention to return to Taiwan (yes-no and provisions for reasons)
Visits to Taiwan (visits last 3 years, last purpose
a
, year of last visit
a
)
Visits to Competitors (e.g. where
a
, how attributes compare to Taiwan
a
, preferences
a
)
Accommodation
Types used, satisfaction and desirable attributes (5þ variables for attributes)
Speci?c hotels stayed in (5 þ can be identi?ed)
Preferences for accommodation (price, location, type and 5 þ services to check as wanted)
Expenditures/spending (including number of persons to whom expenditures apply)
Information allowing estimation of spending of different types in Taiwan
Demographics
Age, sex, occupation, income, education, nationality and country of residence
Note:
a
Indicates not a variable for all of 2000, 2001 and 2003
PAGE 56
j
INTERNATIONAL JOURNAL OF CULTURE, TOURISM AND HOSPITALITY RESEARCH
j
VOL. 6 NO. 1 2012
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
2
:
1
8

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
information collected by Taiwan is that usually collected in survey research in which
attributes of the respondents is considered important (e.g. see web references for
demographic information in the US and Canadian surveys). Dolnicar (2008) and Binge´ et al.
(2008) provide numerous references to research using demographic information. Anyway,
the demographic information being that usually collected is, as far as the authors can
determine, Taiwan Tourism Bureau’s justi?cation for its collection.
Data items listed under arrival/departure (Table I) are justi?ed based on the literature.
However, one can note that surveys of inbound visitors include variables like the ones listed
because for various types of trips logic implies the relevance of the variables. When one
thinks about people traveling, duration of a trip, the composition of the travel party and
transport mode(s) are variables that do contribute to explaining travel related decision
making and behavior (see articles in Woodside and Martin, 2008). Nevertheless, as
re?ected by the conditional asking of questions in TIB surveys, some questions for those on
a tour are not appropriate for people organizing their own travel (e.g. see Taiwan Tourism
Bureau, 2001). A thorough enumeration of appropriate questions would be hierarchical.
Questions being speci?c to one, or to more than one, segment is a matter pursued below.
Response patterns, responses to or responding to questions being highly dependent on
segments, receives much attention in developing insights from in depth interview
information (e.g. see chapters in March and Woodside, 2005; Martin, 2008).
Information search is acknowledged as important by inclusion of commonly asked questions
in TIB surveys. For references on this topic, Zins (2007) comments on important information
search articles, introduces literature relevant to 2009 and cites literature on trip planning.
Data listed in Table I under information and trip planning are consistent with suggestions in
the literature Zins cites.
A variety of data listed in Table I identify what tourists respond about a trip such as desired,
expected, purposes, satisfaction, frequency of past visiting of Taiwan and other matters. For
literature one can see Hsu and Huang (2008) for a review of travel motivation concept
development. In both Dolnicar (2008) and Binge´ et al. (2008) a variety of studies are
identi?ed and analysis techniques are discussed for data listed under attraction to Taiwan
and activities during visiting. Looking at research studies can lead one to recognize long lists
of things to ask about. In the 2003 Taiwan inbound survey under reasons for visiting Taiwan,
14 choices exist. For activities, 15 are listed. For meeting expectations, the list has ten items.
For satisfaction, the list is divided into three parts with a total of 14 items. For accommodation
services, eight are listed. The authors did not ?nd articles that dealt explicitly with restricting
lists so that one met limits on analysis related to having too many categories (e.g. see
Dolnicar, 2008). Getting usable multivariate categories is a topic in analysis.
For questions on motivation, expectation, meeting expectation and intention to return, a
range of literature provides guidance for theory development and analysis. Fallon (2008)
covers articles discussing expectancy and the discon?rmation model. The discussion
addresses expectations motivating repurchase and consequences of failing to meet
expectations affecting purchasing. Taiwan’s data identi?ed in the last paragraph seem to
invite developing expectancy-discon?rmation theory or applying this theory to making
inferences. From another perspective, importance-performance analysis (IPA) is endorsed
for drawing practical inferences from data on wants and satisfaction (Fallon, 2008). Beaman
and Huan (2008) address problems in using IPA. Also, TIB data have information on trip
purpose and segmentation by trip purpose is an analysis approach covered in the literature
(e.g. see citations in Binge´ et al., 2008). Segmentation by purpose can involve a priori or
common sense division into groups by selecting visitors based on purpose and other
variables (Laesser et al., 2006; Dolnicar, 2008). For example, recognizing that a person is
visiting friends or relations (VFR) can be important because VFR visitors do not require the
same services as those using commercial services. Asking VFR visitors about services that
are irrelevant creates confusion between, for example, zero for chooses not to use and
irrelevant. Recognizing tour visitors with different purposes (e.g. visiting temples or
wilderness) can clarify such matters as type of accommodation services used and
participation in activities. People in tour busses do not need personal vehicle services and,
VOL. 6 NO. 1 2012
j
INTERNATIONAL JOURNAL OF CULTURE, TOURISM AND HOSPITALITY RESEARCH
j
PAGE 57
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
2
:
1
8

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
because of being shepherded from entry to Taiwan to exit may not need services asked
about in TIB surveys. The problemis that respondents may ?nd TIB questions about services
used confusing and analysts may have no way to recognize answers do not mean what they
think they do. This research does not pursue this matter.
Questions about intentions (e.g. intention to return) complement other data mentioned in this
section because ?nding what is wanted and being satis?ed, in theory, are causes of return
and loyalty while disappointment promotes not returning (e.g. see Niininen and Riley, 2004).
While Kozak et al. (2002) acknowledge the importance of asking about intention to return,
they question the value of vague information (e.g. responses like ‘‘I’m likely’’ or ‘‘not likely’’ to
return). They suggest that without a time frame in which return is likely, stated intention to
return yields virtually no information about return ?ow per time interval (e.g. year) of visitors.
Furthermore, Wang et al. (2012) produce evidence that for less than 50 percent return (ever)
of leisure inbound to person-visits to Taiwan for which a respondent states she/he intends to
return.
As for expenditure related data, for Taiwan, and other countries, an important purpose of
collecting data from inbound tourists is learning about and tracking their economic impact.
For Taiwan and expenditures, see, for example, Sun (2005). Expenditure data can be used
for other purposes than estimating economic impact (Mok and Iverson, 2000). However, as
implied above, discussion of the use of expenditure data should not be pursued because of
the need to diverge into requirements of economists for making estimates.
Hypotheses
The general hypothesis behind this research is H1. What is stated has various implications.
One is that year after year collection of data similar to those described in Table I is needed.
An aspect of evaluation is examination of TIB surveys being used to make good decisions or
to reach valid conclusions in developing theory. Another matter to consider regarding
continuing collection of data on wants, information search, activities, etc. is variables or
analyses involving newly collected data being valuable given the data already available and
research done (or that could have been done). Does one need certain results every year or
few years is a question to ask. From another perspective, are all variables being collected of
some use in developing theory or meeting practical needs? A variable can be important in
theory but not in practice because of inaccuracy (e.g. high bias or variability meaning the
true value of x has a high probability of being, e.g. 25 percent different from x):
H1. Collecting the kind of inbound data identi?ed in Table I is an effective way to spend
resources if the goal is to develop an understanding of inbound visitor decision
making and behavior to support planning, managing and marketing.
The following hypotheses re?ect the ideas implicit in formulating data collection such as
identi?ed in Table I. H2 expresses the reasonable expectation that data collected based on
theory can be used to con?rm the validity of the theory. H3 suggests that con?rming theory
and practical results will result from obtaining useful information from long lists of
expectations, activities, events, etc. H4 highlights repeated collection (e.g. annually) as
important in planning, marketing and management. Note that, for example, accepting H2
does not mean accepting H4:
H2. Analysis of TIB data shows that Taiwan’s inbound visitors decision making and
behavior are as expected by theory.
H3. TIB questions with multiple responses on who travels, expectations, expectations
being met, satisfaction, activities, etc. give one the detailed insights into visitors’
decision making and behavior needed for planning, managing and marketing of
inbound tourism.
H4. TIB data collected repeatedly is necessary and effective for con?rming/developing
theory and/or establishing trends that are useful for planning, managing and
marketing of inbound tourism.
PAGE 58
j
INTERNATIONAL JOURNAL OF CULTURE, TOURISM AND HOSPITALITY RESEARCH
j
VOL. 6 NO. 1 2012
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
2
:
1
8

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
Analysis and results
Accepting H1 depends on accepting H2, H3 and H4. For example, if H2 is rejected,
accepting H1 is not appropriate because data collection indicated in Table I is based on
behavior conforming to theory. As for H3, one could say that you accept the hypothesis if you
get any useful insights. An alternative is rejecting the hypothesis because, for example,
having too many categories is an impediment to getting useful insights. This research
examines detailed in data collected affecting gaining useful insights. As with H3, H4 being
con?rmed is approached by considering producing useful information rather than, for
example, accepting or rejecting a null hypothesis. For example consider an equation for
estimating numbers of golfers is found to be signi?cant and have signi?cant coef?cients.
Estimating the equation for a given year may contribute to theory and understanding.
Estimating it for several years may bolster con?dence in the result. Continuing to estimate it
every year, year after year has limited value. Estimating a time series equation for ten years
and getting a signi?cant growth rate greater than zero that has a 90 percent probability of
being between 1 and 5 percent is not producing an equation with practical value. Similarly, if
applying a time series or other equation to estimate, for example, the number of golfers
coming to a facility in a year helps to develop theory but prediction for the facility is expected
to be out by ,25 percent or more 50 percent of the time, the model has little practical value.
Being good enough for theory development does not mean good enough for practical use.
In another context, the coef?cient of variation of a national estimate expressed as a percent
(i.e. 100* standard deviation/estimate) can indicate usefulness for planning. Based on
having large samples, some national statistics (e.g. for the US or Canada) are quite reliable
(e.g. have a CV of 5 percent). However, while a national estimate is acceptably reliable, a
comparable statistic for a state or province can be of little use (e.g. have CVof 25 percent or
more). Producing statistics that are inaccurate or likely to be inaccurate is an indicator of
being ineffective.
Modeling and ambiguous conclusions
Consider showing that stated intention to return depends on various in?uences. Reason
suggests that for consumer products positive responses on meeting or exceeding
expectations results in intent of repurchase. For example, shopping or visiting historic
locations exceeds expectation, this contributes to deciding to return while negatives
contribute to deciding not to return. Table II gives information on the ten variables for which
responses on meeting expectations are obtained. Hereafter, factor refers to a variable such
as the rating of meeting expectations on shopping. A na? ¨ve model formulation is intention to
return for a respondent (r
r
) is a linear function of the expectation variables (say, 1
i,r
).
However, for TIB data only some of the 1
i,r
apply to each individual. This is because
nonresponse can indicate that a variable is not relevant. Let m
r
be the mean of r’s 1
i,r
ratings.
If a respondent gives ratings for three of the ten factors m
r
would be the average of the three
responses (e.g. gives 3, 2 and 1 for factors 2, 4 and 7 with m
r
¼2). Equation (1) implies that
intention to return depends on m
r
and hotel satisfaction (h). Since over 95 percent of
respondents stay in hotels (see ‘‘Stay at’’ in Table III), h is included in the model because
being satis?ed with accommodation presumably impacts returning. Failing to consider h
could result in not ?nding an effect when one should be found:
r
r
¼ c
0
þ c
1
m
r
þ c
2
h þ e ð1Þ
where r
r
, m
r
and h are as de?ned above and c
0
and c
1
are constants estimated by regression
and e is random error.
In Table IV, one sees regression results for estimation of equation (1) (see results for all).
When the model is run for speci?c years (2000, 2001 and 2003), the variance explained
remains highly signi?cant (F,0.0001). Also, the coef?cients c
0
and c
1
have consistent
values from year to year. The coef?cient c
1
being positive shows doing better at meeting
expectations results in a greater likelihood of return. Given that one is testing the hypothesis
that c
1
.0, c
1
being over three times its standard deviation means the probability of c
1
being
zero is ,0.005. In other words, the results support theory.
VOL. 6 NO. 1 2012
j
INTERNATIONAL JOURNAL OF CULTURE, TOURISM AND HOSPITALITY RESEARCH
j
PAGE 59
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
2
:
1
8

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
The equation (1) model, does not take into account the possibility that performing well on
factors that motivated coming to Taiwan should be more important than performing well on
other factors. By calculating a mean for factors motivating visiting (m
m,r
) and one for other
factors (m
O,r
), equation (2) de?nes a model allowing for doing well on motivating factors while
still considering performance on other factors. A reasonable hypothesis is that c
1
.c
2
.0.
Results of estimation of the model appear in Table IV (see results for Motiv). While F values
show the model explains signi?cant variance, the values of c
0
, c
1
, c
2
and c
3
are not
consistent with c
1
.c
2
.0; c
1
and c
2
are not consistently signi?cant over time (for 2000, 2001
and 2003). A model that could be expected to ?t data better than equation (1), in fact, does
not yield consistent results. Furthermore, though statistically signi?cant relations exists
explanation is poor enough that predictions made using the All and Motiv models are not
reliable enough that they should be used in decision making. The results support rejecting
H2. In other word, one is paying for some questionable data in TIB surveys so changing the
survey could reduce cost or produce more useful data:
r
r
¼ c
0
þ c
1
m
m;r
þ c
2
m
O;r
þ c
3
h þ e ð2Þ
where r
r
, m
m,r
, m
O,r
and h are as de?ned above and c
0
, c
1
and c
3
are constants estimated by
regression and e is random error.
The results support rejecting H2. Is the problem that theory is wrong? Is the problem that one
does not have enough data? Is the problemthat with more detail only segment speci?c models
will support theory? Different segments, s, may have different c
0
, c
1
and c
3
(e.g. have c
0,s
, c
1,s
and c
3,s
). Dolnicar (2008) and Binge´ et al. (2008) make clear the importance of considering
segments. In this research, K-means (Noursis and SPSS Inc., 1993, ch. 4) was used to place
respondents in different numbers of clusters (3, 4, 5, 6, 7, 8 and 9 segments) but nothing useful
was learned. Many respondents being on tours affects activities done and could be in?uencing
expectations being met. Having a high certainty that a model shows that theory conforms to
reality does not mean predictions made using the model have practical utility.
Finally, discussion of equation (1) and (2) models has not pursued R
2
being about 0.02.
Intention to return has an average value of about 96 percent. Having 96 percent of
respondents indicate intent to return means a small percent of respondents is being heavily
Table II Speci?cs on variables included in 2000, 2001 and 2003 Taiwan inbound survey (TIB) data
Motivations for traveling to Taiwan
2000 2001 2003 Participation 2000 2001 Participation 2003
Meeting expectations 2000
2001 2003
(1) Scenery
(2) Speci?c tourism activities
(3) Prices of goods
(4) Cuisine
(5) Fruits
(6) Weather
(7) Night life
(8) Leisure facilities (e.g.
amusement park golf course)
(9) Historic relics
(10) Arrangement by of?ce
(11) Near the place you live
(12) Taiwan custom and culture
(13) Public securities
(14) Democracy
(15) Other
(1) Sightseeing
(2) Outdoor activities
(3) Playing golf
(4) Hot spring (spa)
(5) Shopping
(6) Visiting historic scenic spots
(7) Wedding photography
personal albums of pictures
(8) Cosmetic preparation
(9) Karaoke or K.T.V.
(10) Pubs or night clubs
(11) Attending an exhibition
(12) Massage
(13) Photography
(14) Participating culture activities
(15) Mighty earth tour (a visit to
the earthquake-damaged
area)
(16) Other
(1) Outdoor recreation (e.g.
climbing. Diving or bird
watching)
(2) Gol?ng
(3) Hot spring
(4) Shopping
(5) Historic heritages
(6) Wedding photography salon
shots
(7) Facial spa beauty treatment
or nails art
(8) Massage or acupressure
(9) Karaoke or K.T.V.
(10) Pubs or night clubs
(11) Exhibition
(12) Local festival
(13) Culture events
(14) Spa sauna
(15) Night marketing sightseeing
(16) Other
(1) Historic relics
(2) Scenery
(3) Prices of goods
(4) Cuisine
(5) Weather
(6) Public securities
(7) Recreational facilities
(8) Traf?c
(9) Friendliness of people
(10) Surrounding
sanitation
Notes: Response coding for meeting expectations 2000 and 2001: 1 – better, 2 – same, 3 – worse; Response coding for meeting
expectations 2003 was reversed: 1 – worse, 2 – same, 3 – better
PAGE 60
j
INTERNATIONAL JOURNAL OF CULTURE, TOURISM AND HOSPITALITY RESEARCH
j
VOL. 6 NO. 1 2012
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
2
:
1
8

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
relied on in assessing factors affecting returning; too little variance is shared between the
dependent and explanatory variables. Actually, Wang et al. (2012) provide evidence that
over 50 percent of respondents indicating they would return will never return. If learning
about return is a goal, one is ineffective if models work poorly and the dependent variable
used in modeling does not re?ect reality about half the time. Therefore, results support
rejecting both H2 and H4.
Trends – some issues
Table V provides information that helps one understand why TIB information on participation
in activities has limited value. Shopping and visiting historic respectively have 87 and 72
percent participation. Other activities only reach about 20 percent with photo sessions and
golf at 2 and 3 percent. Now, one may argue that year-to-year ?gures are important for
tracking change. Given 10 percent is based on participate from about 140 respondents, a
standard deviation of about 12 is expected (Beaman and Tompson, 2004). Over three years,
observing patterns like 150, 140 and 130 or 130, 140 and 150 is not unlikely though no
change is taking place. Fro most activities, TIB not yielding good information for establishing
national/overall trends in participation. Even with ?ve or more years of data and if Taiwan
weighted data to re?ect annual visitor volumes, estimates of percent change would not be so
unreliable that use in planning or managing has a high probability of leading to poor
decisions. Getting information from the supply side could however yield valid trend
information (e.g. getting data on gol?ng fromsuppliers). The point is that H4 is not supported
in as much as most information on participation in activities is not useful for establishing
trends for use in planning or management.
Lists as responses and recognizing segments
In discussion rates of participation in activities (Table V), segment speci?c rates are not
examined. When most overall rates are not reliable enough to use in measuring trends, rates
for segments will not be good because they are based on fewer observations. Whether
considering planning for segments based on participation rates or based on numbers
estimated using models, identifying groups of respondents that are homogeneous with
respects to, for example, certain purchase behaviors of interest, is an essential step. In other
words, determining segments is important.
Finding segments can be challenging (e.g. see Dolnicar, 2008). With n variables with levels
l
k
respondents can be in C¼Pl
k
cells (Pl
k
designates the product of the l
k
values). Table II
shows categories for motivations for traveling to taiwan, participation and meeting
expectations. For example for participation in activities (Tables II and III), 15 zero-one
responses can indicate activities of participation (other is not included). Symbolically a
combination is (001011001011001011). All possible cells number 32,768 (¼2
15
). Table III
shows term such as hot spring/spaþshop þ visit historic spots þ massage rather than a
sequence (e.g. 10000011100). With thousands of possible response combinations,
respondents may be spread sparsely over the possible combinations of responses.
Therefore, using computer programs that ?nd structure in data to get information to use in
theory building or to meet practical needs may seem reasonable.
Cluster analysis of data by computer takes different forms (Dolnicar, 2008). When responses
are categories, a reasonable goal is ?nding cells that are similar, for example, based on a
mathematical criterion (e.g. closeness in Euclidean distance) and treating all people in
similar cells as in a cluster. People in a cluster might all do 4 particular activities and some of
them do some other activities. With more clusters, one can have more detail de?ning
clusters. Dolnicar (2008) writes that for determining clusters with h binary variables the
number of respondents should be at least 2
n
and ideally 5*2
n
. Computation shows
2
12
¼ 4; 096 cells, which is about the number of respondents available in this research.
However, as noted above, participation in activities has 15 dichotomous variables
suggesting analysis should be for a minimum of over 30,000 respondents. Not using
computer based clustering is suggested by the 2
n
rule. Given 15 variables for reasons for
visiting Taiwan and ten for meeting expectations (Table V), even with 3 years of data the
number of respondents only meets the 2
n
rule for meeting expectation. Regardless, using
VOL. 6 NO. 1 2012
j
INTERNATIONAL JOURNAL OF CULTURE, TOURISM AND HOSPITALITY RESEARCH
j
PAGE 61
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
2
:
1
8

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
Table III Cell based response patterns for party, participation, want/motivate and stay at –
cells based on possible response combinations for a collection/group of variables
n
Party (for 32 cells, cells with $ 30 responses)
Respondent 492
Respondent þ spouse 531
Respondent þ children 71
Respondent þ spouse þ children 136
Respondent þ relation 413
Respondent þ spouse þ relation 35
Respondent þ colleagues 398
Respondent þ friends 1,763
Respondent þ spouse þ friends 45
Respondent þ relation þ friends 52
Respondent þ colleagues þ friends 51
Participation (for 32,768 ¼ 2**15 cells)
Responses for cells with ,20 responses 799
No participation identi?ed 82
Shop 338
Outdoor activities þ shop 41
Hot spring/spa þ shop 75
Visit historic spots 67
Shop þ visit historic spots 1,046
Outdoor activities þ shop þ visit historic spots 85
Hot spring/spa þ shop þ visit historic spots 330
Outdoor activities þ hot spring/spa þ shop þ historic 64
Shop þ cosmetic preparation 42
Shop þ visit historic spots þ cosmetic preparation 84
Shop þ karaoke or K.T.V. 20
Shop þ visit historic spots þ karaoke or K.T.V. 38
Hot spring/spa þ shop þ visit historic spots þ karaoke 20
Shop þ pub or night clubs 24
Shop þ visit historic spots þ pub or night clubs 47
Hot spring/spa þ shop þ visit historic. . . þ pub-club 27
Shop þ attending exhibit 21
Shop þ visit historic spots þ attending exhibit 57
Hot spring/spa þ shop þ visit historic spots þ exhibit 39
Outdoor þ hotspring/spa þ shop þ historic þ exhibit 43
Shop þ massage 137
Visit historic spots þ massage 50
Shop þ visit historic spots þ massage 381
Hot spring/spa þ shop þ visit historic spots þ massage 53
Shop þ visit historic spots þ cosmetics þ massage 22
Shop þ visit historic spots þ karaoke þmassage 27
Shop þ visit historic spots þ culture activities 21
Want/motivate (16,384 cells)
Responses for cells n , 20 1,474
Company trip 68
Cuisine 318
Cuisine þ fruits 22
Cuisine þ near 102
Cuisine þ near þ safe 33
Cuisine þ nightlife 21
Cuisine þ relics 80
Cuisine þ relics þ near 49
Cuisine þ twculture 59
Cuisine þ weather þ near 23
Near 96
No want/motivate response 466
Prices 24
Prices þ cuisine 115
Prices þ cuisine þ near 69
Prices þ near 41
(Continued)
PAGE 62
j
INTERNATIONAL JOURNAL OF CULTURE, TOURISM AND HOSPITALITY RESEARCH
j
VOL. 6 NO. 1 2012
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
2
:
1
8

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
SPSS’s k-means (Noursis and SPSS, 1993) to determine clusters based on participation in
activities, reasons for visiting and meeting expectations did not group respondents into
closely matching clusters. Is the result ?awed because of too much detail? Or, is overlap not
large because of too much detail and other reasons? Research is needed. TIB data can be
use. Still, results found provides evidence for rejecting H2.
Table III is important in developing a perspective getting useful information from TIB lists of
responses (e.g. participation in activities). Computer use for data-driven segmentation
being problematic for data available is also mentioned. A way of segmenting that does work
parallels approaches used in analysis of in depth interviews (e.g. see chapters in March and
Woodside, 2005; Martin, 2008; DeCrop, 2001). Introducing the idea of respondents being in
cells lays looking at information in Table III and using common sense in determining
segments. One lets the mind determine patterns rather than using arbitrary rules built into
computer programs. Consider making sense of who travels together, party (upper left of the
table). The number of possible categories based on 5 dichotomous variables is 32.
However, 1,763 of 3,987 respondents report just traveling with friends (44 percent) while
traveling with friends, colleagues and family covers 58 percent of respondents. In fact,
analysis might be as good if traveling alone, with spouse, with relation other than just the
spouse, traveling with colleagues and any travel group including friends. For participation in
activities, a big segment would be the 799 listing no activity. What is the meaning of about 20
percent of respondents reporting no activity? About 20 percent not doing any of the 15
activities seems unrealistic and thus a TIB survey problem to be addressed. Anyway, one
?nds relatively large numbers associating a second activity with shopping (e.g. see 330 for
hot spring/spaþshopping þ visiting historic scenic spot). Examination of motivation and
participation information shows that a few of thousands of cells have over 20 occupants.
Anyway, presenting clusters to use in TIB data collection or analysis is not a goal of this
research. However, by showing that most respondents are relatively few cells and fewer
clusters, one has evidence that data could be collected with far fewer categories/variables.
Unfortunately, ?nding out if some people are in cells because tour activities or desires of
friends cause their membership while others people choose particular activities cannot
occur with current TIB data. Therefore, models that implicitly treat respondents as choosing
Table III
n
Rec facilities 21
Relics 49
Relics þ near 22
Scenery 147
Scenery þ cuisine 226
Scenery þ cuisine þ fruits 22
Scenery þ cuisine þ near 63
Scenery þ cuisine þ nightlife 28
Scenery þ cuisine þ relics 62
Scenery þ near 57
Scenery þ prices 23
Scenery þ prices þ cuisine 80
Scenery þ prices þ cuisine þ near 35
Scenery þ prices þ near 25
Scenery þ relics 65
Scenery þ relics þ twculture 21
Scenery þ twculture 34
Twculture 40
Stay at (128 cells)
Responses for cells , 20 43
Community center 26
Hotel 3,887
Hotel þ private home 36
Private home 88
VOL. 6 NO. 1 2012
j
INTERNATIONAL JOURNAL OF CULTURE, TOURISM AND HOSPITALITY RESEARCH
j
PAGE 63
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
2
:
1
8

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
T
a
b
l
e
I
V
M
o
d
e
l
s
o
f
v
a
r
i
a
b
l
e
s
a
f
f
e
c
t
i
n
g
s
t
a
t
e
d
i
n
t
e
n
t
i
o
n
t
o
r
e
t
u
r
n
P
a
r
a
m
e
t
e
r
s
P
a
r
a
m
e
t
e
r
s
t
a
n
d
a
r
d
d
e
v
i
a
t
i
o
n
Y
e
a
r
M
o
d
e
l
F
P
r
o
b
.
F
R
S
Q
I
n
t
e
r
c
e
p
t
M
o
t
i
v
a
t
e
d
m
e
a
n
O
v
e
r
a
l
l
m
e
a
n
O
t
h
e
r
m
e
a
n
H
o
t
e
l
s
a
t
i
s
f
a
c
t
i
o
n
I
n
t
e
r
c
e
p
t
M
o
t
i
v
a
t
e
d
m
e
a
n
O
v
e
r
a
l
l
m
e
a
n
O
t
h
e
r
m
e
a
n
H
o
t
e
l
s
a
t
i
s
f
a
c
t
i
o
n
2
0
0
0
M
o
t
i
v
5
.
6
(
3
,
9
7
)
0
.
0
0
8
0
.
0
1
7
0
.
8
1
0
.
0
1
9
0
.
0
2
1
0
.
0
1
8
*
0
.
0
4
3
0
.
0
1
1
0
.
0
1
8
0
.
0
0
6
5
2
0
0
1
M
o
t
i
v
9
.
9
(
3
,
1
4
)
,
0
.
0
0
1
0
.
0
2
0
0
.
7
6
0
.
0
1
5
0
.
0
7
6
*
*
*
0
.
0
0
1
4
0
.
0
3
9
5
0
.
0
1
0
9
0
.
0
1
7
0
.
0
0
5
4
2
0
0
3
M
o
t
i
v
7
.
8
(
3
,
1
0
)
,
0
.
0
0
1
0
.
0
2
2
0
.
8
0
0
.
0
4
2
*
*
0
.
0
1
2
0
.
0
0
7
6
0
.
0
3
9
7
0
.
0
1
1
2
0
.
0
1
6
0
.
0
0
5
6
2
0
0
0
A
l
l
1
3
(
2
,
1
1
)
,
0
.
0
0
1
0
.
0
2
4
0
.
7
8
0
.
0
4
8
*
*
0
.
0
1
9
0
.
0
3
7
0
.
0
1
3
0
.
0
0
6
5
2
0
0
1
A
l
l
1
1
(
2
,
1
5
)
,
0
.
0
0
1
0
.
0
1
4
0
.
8
2
0
.
0
4
6
*
*
0
.
0
0
8
5
0
.
0
3
2
0
.
0
1
3
0
.
0
0
5
4
2
0
0
3
A
l
l
1
4
(
2
,
1
4
)
,
0
.
0
0
1
0
.
0
2
5
0
.
7
8
0
.
0
6
4
*
*
*
0
.
0
0
8
6
0
.
0
3
6
0
.
0
1
4
0
.
0
0
5
6
PAGE 64
j
INTERNATIONAL JOURNAL OF CULTURE, TOURISM AND HOSPITALITY RESEARCH
j
VOL. 6 NO. 1 2012
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
2
:
1
8

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
participation in activities are ?awed. Problems exist because TIB data do not allow for
understanding some aspects of behavior. Another reason for not accepting H2 is that one
does not have some variables needed to recognize segments with differing behavior.
The research provides evidence for rejecting H3. Can the detail in a year’s data or in several
years’ data be useful in ?nding segments or in deciding on sample sizes required to get
reliable participation rates? That it can is what the research shows. However, if participating
in activities is segment dependent and for some respondents an activity is not an option,
appropriate information must be collected to sort out what is affecting what and what is
irrelevant to whom.
Based on evidence providing justi?cation for rejecting H2, H3 and H4, accepting H1 is not
reasonable. However, taking this research as showing TIB data collection is not effective, is
also not reasonable. All the research really supports is the conclusion that improvement in
collection can occur. Given that asking a better question about intention to return produces
measures that are much closer to reality than the current question, better data can be
produced with no increase in cost. Given that annually asking about participation in low
participation rate activities yields information of questionable value, less frequent collection
or using alternative sources could reduce the cost or TIB. Effectiveness can be increased.
How much the increase can be is not clear.
Discussion
This research does not prove that TIB survey data collection is ineffective. Claiming the
collection is ineffective because some evidence shows that collection could be more
effective exaggerates the importance of ?ndings. Collection of data for estimating
expenditures some statistics about economic impacts may be effective in the sense of
obtaining the information needed at near minimum cost. As for TIB data, other than
expenditure data, a better understanding of segments and how detailed information applies
to marketing and providing service on the ground in Taiwan should allow for improving data
collection in ways discussed below.
Being more effective
Improving effectiveness involves seeking improvement in what is done. When one recognizes
that simple regression results con?rmtheory but more realistic modeling does not, one should
confront moving fromestablishing propositions that are so general as to getting useful results.
Yes, Wang et al. (2012) show a problem with is the modeling carried out arises because 97
percent of respondents specify intending to return when lest than half will ever return. Better
Table V Mean activity participations of leisure visitors based on yes ¼ 1 and 0 for not
reporting participation
Mean for all respondents for
yes/participated ¼ 1 and zero otherwise
Variable
2000
n ¼ 1,208
2001
n ¼ 1,624
2003
n ¼ 1,248
All years
n ¼ 4,080 Approximate SD
Intention to return 0.97 0.95 0.97 0.96 0.01
Nature activities 0.06 0.07 0.21 0.11 0.01
Golf 0.03 0.02 0.01 0.02 ,0.01
Hot spring/spa 0.21 0.19 0.32 0.23 0.01
Shopping 0.87 0.93 0.86 0.89 0.01
Visit historic spots 0.72 0.74 0.69 0.72 0.01
Wedding/personal photographs 0.02 0.01 0.01 0.01 ,0.01
Cosmetic treatments 0.09 0.08 0.03 0.07 0.01
Karaoke/K.T.V. 0.10 0.08 0.06 0.08 0.01
Pubs/night clubs 0.10 0.08 0.07 0.08 0.01
Attend exhibition 0.06 0.05 0.14 0.08 0.01
Massage 0.20 0.20 0.29 0.23 0.01
Cultural activities 0.02 0.02 0.08 0.04 0.01
VOL. 6 NO. 1 2012
j
INTERNATIONAL JOURNAL OF CULTURE, TOURISM AND HOSPITALITY RESEARCH
j
PAGE 65
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
2
:
1
8

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
questions on intention to return are needed both to knowwhen return is likely and if it is actually
likely. Kozak et al. (2002) have examined research on return. An important matter is that a
person indicating they are likely to return yields almost no useful information on future tourist
?ow. What do you infer if a person indicates returning and came once in the past 3 years?
Questions that yield reliable information on time frame of return are needed. Furthermore, if
collecting wants and meeting expectations is to be worth doing, viable and valid ways of using
the data need to be established. As for activities, the inadequacy of most overall participation
rates for establishing trends has been introduced. The matter of participation being segment
dependent suggests the need to know participation rates by segments for planning and
management. Respondents being in segments with different return intentions (e.g. only
making one visit regardless of expectations being met – Kozak et al., 2002), activity
participation rates and models needs to be addressed.
Being able to identify segments is an important theme of this paper. Collecting detail from
respondents using lists, for example, for identifying activities or motivations re?ects building
data collection on theory in as much as lists used are consistent with theory supported by
research. However, trying to derive data-driven segment with about 4,000 respondents
potential in about 32,000 cells is not recommended (Dolnicar, 2008). Furthermore, segments
based on motivation, on participation and on meeting expectations do not correspond very
well. Given most respondents are on organized tours and many travel with friends, possibly
activity information has little value for understanding travel by many visitors. The premise
raised is that TIB data need to be studied to determine and some in depth interview research
may be needed to identify segments from which to collect data and what information is
appropriate from particular segments. By special questions from some segments (business
and conference), Taiwan is collecting segment speci?c data but more consideration of
segments is needed.
This research suggests that TIB data get less and less valuable with time. With one year of
TIB data, insights are possible into what data are needed to achieve some objectives. With
three years of compatible data, one may be able to develop insights that will not occur with a
third of the data or one may have con?dence in insights that is not possible with a third of the
data. However, the statistical principle of estimates increasing in reliability as the square root
sample size suggests that moving from three years of data to four has limited value. And, as
one aggregates data over time, the effect of change (even disruptions such as SARS) must
be considered (Hwang et al., 2002). Data with the detail of TIB data being collected for one
to three years may be effective, or is a better word necessary, in developing strategies for
collecting data that meet ongoing needs. Nevertheless, continuing collection (e.g. over
decades) of TIB data in the current form seems to be throwing good money after bad.
Intensive analysis of the TIB should be occurring to determine how practical needs can be
met effectively.
Modeling and data requirements
Consideration of the model de?ned by equation (3) suggests how logical thinking could be
contributing to planning for data collection. Discussion of segments, s, implies segment
speci?c weights for factors (e.g. for P
f,s
). The dependent variable, I
r
, could be ordinal but
re?ect likelihood of returning in a given time frame (e.g. 1 is about twice a year, 2 is about
once a year, etc.). At ?rst glance, equation (3) may appear to be a linear equation. However,
the equation is nonlinear so model estimation might be carried out using nonlinear
regression (e.g. see NLIN in SAS Institute, 1988a, b). Estimation was not pursued because
Taiwan’s inbound data do not allow estimation. The point is that if such a model is
appropriate, data should be collected that allow its estimation. Wang et al. (2012) also show
the value of modeling and, by doing so, show an inadequacy of Taiwan’s inbound data.
I
r
¼ a
0;s
þ a
1;s
ðSP
f ;s
x
f ;r
Þ=ðSd
x
1
Þ ð3Þ
where:
B I
r
¼is an intentions variable; r is in segment s.
B a
0,s
, a
1,s
and P
f,s
for all factors are regression coef?cients.
PAGE 66
j
INTERNATIONAL JOURNAL OF CULTURE, TOURISM AND HOSPITALITY RESEARCH
j
VOL. 6 NO. 1 2012
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
2
:
1
8

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
B ¼rating by person r on f and does not add to the sum if f is not rated (e.g. ¼ 0).
B d
1
x
¼1 if x (actually x
f,r
)is a valid rating and 0 otherwise.
Using accurate and alternative information
An interesting consideration regarding international travel is that accurate statistics are
created in some countries. Taiwan, as many other countries, uses computers in collecting
information from every person entering and leaving the country legally. In other words,
Taiwan has census information on inbound visitors that should be used in analysis and in
checking the validity of survey data (Wang et al., 2012). Given international visitors leaving
Taiwan register both entry and exit, data created by entry-exit control of?cials (EECO),
extensive accurate data can be produced. Based on current (2009) work, production of
statistics on origin, purpose and duration by such passport variables as age and gender is
reasonable. Some accurate information about inbound tourism can be produced without a
survey. In fact, using passport information, person-visits as well as unique persons visiting in
a year can be units used in producing statistics (Wang et al., 2012).
EECO being the collectors of inbound data currently obtained by surveys conducted under
contract may not be reasonable. However, having survey data compatible with EECO
census data is only reasonable. Current TIB data may only tends to approximate some
accurate information because quotas are used in sampling. Therefore, what con?dence can
one have in survey statistics? Comparison of survey results with accurate statistics would
allow determination of and correction for bias is survey statistics (e.g. see Beaman et al.,
2001). Taiwan’s survey data collection would be more effective if coordinated with EECO
statistics.
Continued collection of expenditure statistics from a sample of visitors on departure may be
effective. Canada has changed collection from the international travelers to be more
effective (Bradford et al., 2001; Statistics Canada, 2009). Regardless, given information on
attributes of the inbound to Taiwan (Tables III and V), collecting segment speci?c information
by exit surveys presents challenges. An alternative for getting data frompeople on tours that
control their activity in Taiwan is using tour operators in collecting data. Operators know
numbers, timing, duration, origins, organized activities, some demographics etc. for tour
participants. Getting reliable information on, for example, satisfaction with tour services
would need to be addressed since operators’ opinions do not replace such information from
travelers. Anyway, given heterogeneity of tours, whose needs are current TIB information
meeting? Considering alternatives and adopting good ones is the way to become more
effective.
Conclusions
This research may seem to have taken an odd approach to dealing with getting effective
data on of inbound international travelers and may seemmainly to apply to Taiwan. However,
TIB data are just a vehicle to introduce matters applying to other countries. As well, the
article provides a caution to countries planning to undertake data collection similar to that of
Taiwan. Data collection can be based on theory but not necessarily be good for advancing
theory or for meeting practical needs. As amply demonstrated in chapters of March and
Woodside (2005), asking about a lot of things in detail can be a good ?rst step in determining
how to be effective in getting information. However, too much detail can cause problems
when one moves to relatively large surveys (e.g. to get reliable estimates of tourist ?ows or
expenditures). Creative work by Taiwan in determining how important information (estimates
of person-visits in seasons in segments) can be determined is necessary so data collected
have more value. In other words, considering how data can be analyzed to achieve
inadequately accurate results to meet user needs. Planning should include having sample
sizes for segment that are needed to produce adequately accurate results. Getting some
detail, as from in depth interviews, may allow identifying data needed thereby facilitating
effective research ensuing. Taiwan and other countries using census data fromentry and exit
of international travel is critical to knowing that survey data is reliably related to actual
traveler ?ow (e.g. person-visits per unit of time).
VOL. 6 NO. 1 2012
j
INTERNATIONAL JOURNAL OF CULTURE, TOURISM AND HOSPITALITY RESEARCH
j
PAGE 67
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
2
:
1
8

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
This research is not de?nitive. However, by raising matters using real data, needed research
should be encouraged. Therefore, this research should contribute to better data collection
and analysis.
References
Beaman, J.G. and Huan, T.C. (2008), ‘‘Importance performance analysis (IPA): confronting validity
issues’’, in Woodside, A. and Martin, D. (Eds), Tourism Management; Analysis, Behavior, and Strategy,
Elsevier, Barking, pp. 358-77.
Beaman, J.G. and Tompson, E. (2004), ‘‘Market trends: appropriateness of using a rate of change based
on two independent samples and its reliability’’, in Arsenault, N. (Ed.), Proceedings CD of TTRA Canada
Annual Meeting, Ottawa, Canada.
Beaman, J.G., Beaman, J.P., O’Leary, J.T. and Smith, S.L. (2001), ‘‘The impact of seemingly minor
methodological changes on estimates of travel and correcting bias’’, in Mazanec, J.A., Crouch, G.I.,
Brent Richie, J.R. and Woodside, A.G. (Eds), Consumer Psychology of Tourism, Hospitality and Leisure,
Vol. 2, Cabi Publishing, Wallingford, pp. 49-65.
Binge´ , E., Gnoth, J. and Andreu, L. (2008), ‘‘Advanced topics in tourism market segmentation’’,
in Woodside, A. and Martin, D. (Eds), Tourism Management Analysis, Behavior, and Strategy, Vol. 3,
Elsevier, Barking, pp. 129-50.
Bradford, R., Cheung, S. and Lapierre, J. (2001), ‘‘Canada’s international travel survey: keeping up with
demands’’, Proceedings of International Statistical Institute: Seoul 53rd Session 2001, available at:http://isi.cbs.nl/iamamember/CD2/pdf/1077.PDF (accessed August 2009).
Crouch, G.I., Perdue, R.R., Timmermans, H. and Uysal, M. (Eds) (2004), Consumer Psychology of
Tourism, Hospitality and Leisure, Vol. 3, Cabi Publishing, Wallingford.
DeCrop, A. (2001), ‘‘The antecedents and consequences of vacationers’ dis/satisfaction: tales from the
?eld’’, in Mazanec, J.A., Crouch, G.I., Ritchie, J.R.B. and Woodside, A.G. (Eds), Consumer Psychology
of Tourism, Hospitality and Leisure, Vol. 2, Cabi Publishing, Wallingford, pp. 333-47.
Dolnicar, S. (2008), ‘‘Market segmentation in tourism’’, in Woodside, A. and Martin, D. (Eds), Tourism
Management Analysis, Behavior and Strategy, CABI, London, pp. 129-50.
Fallon, P. (2008), ‘‘Monitoring visitor satisfaction with destinations using expectations, importance and
performance constructs’’, in Woodside, A. and Martin, D. (Eds), Tourism Management Analysis,
Behavior and Strategy, CABI, London, pp. 424-59.
Fodness, D. and Murray, B. (1999), ‘‘A model of tourist information search behavior’’, Journal of Travel
Research, Vol. 37 No. 3, pp. 220-30.
Frechtling, D. (2009), ‘‘Subject: invitation to attend discussion of inbound visitor measurement’’,
personal communication, [email protected], 1/13/09.
Hsu, C.H.C. and Huang, S. (2008), ‘‘Travel motivation: a critical review of the concept’s development’’,
in Woodside, A. and Martin, D. (Eds), Tourism Management Analysis, Behavior and Strategy, CABI,
London, pp. 14-27.
Huan, T.C., Beaman, J.G., Chang, L.H. and Hsu, S.Y. (2008), ‘‘Robust and alternative estimators for
‘better’ estimates for expenditures andother ‘long tail’ distributions’’, TourismManagement, Vol. 29 No. 4,
pp. 796-806.
Hwang, Y.H., Wang, R.Y., Beaman, J., Fesenmaier, D.R. and O’Leary, J.T. (2002), ‘‘Considerations in
temporal aggregation: applications to the US in-?ight survey data’’, Tourism Analysis, Vol. 6 Nos 3/4,
pp. 71-84.
Kozak, M., Huan, T.C. and Beaman, J.G. (2002), ‘‘A systematic approach to non-repeat and repeat
tourism with measurement and destination loyalty concept implications’’, Journal of Travel and Tourism
Management, Vol. 12 No. 4, pp. 19-38.
Laesser, C., Crouch, G.I. and Beritelli, P. (2006), ‘‘Market segmentation by reasons and in?uences to visit
a destination: the case of international visitors to Australia’’, Tourism Analysis, Vol. 11 No. 4, pp. 241-9.
March, R. and Woodside, A. (Eds) (2005), Tourism Behavior: Travelers’ Decisions and Actions, Cabi
Publishing, Wallingford.
PAGE 68
j
INTERNATIONAL JOURNAL OF CULTURE, TOURISM AND HOSPITALITY RESEARCH
j
VOL. 6 NO. 1 2012
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
2
:
1
8

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
Martin, D. (2008), ‘‘Grounded theory of international tourism: building systematic propositions from emic
interpretation of Japanese travellers visiting the USA’’, in Woodside, A. and Martin, D. (Eds), Tourism
Management: Analysis, Behavior and Strategy, Cabi Publishing, Wallingford.
Mazanec, J.A., Crouch, G.I., Ritchie, J.R.B. and Woodside, A.G. (Eds) (2001), Consumer Psychology of
Tourism, Hospitality and Leisure, Vol. 2, Cabi Publishing, Wallingford, pp. 333-47.
Mok, C. and Iverson, T.J. (2000), ‘‘Expenditure-based segmentation: Taiwanese tourists to Guam’’,
Tourism Management, Vol. 21 No. 3, pp. 299-305.
Niininen, O. and Riley, M. (2004), ‘‘Towards the conceptualisation of tourism destiation loyalty’’,
in Crouch, G.I., Perdue, R.R., Timmermans, H. and Uysal, M. (Eds), Consumer Psychology of Tourism,
Hospitality and Leisure, Vol. 3, Cabi Publishing, Wallingford, pp. 275-84.
Noursis, M.J. and SPSS (1993), SPSS for Windows Professional Statistics Release 6.0, SPSS, Chicago,
IL.
Pol, A.P., Pascual, M.B. and Va´ zquez, P.C. (2006), ‘‘Robust estimators and bootstrap con?dence
intervals applied to tourism spending’’, Tourism Management, Vol. 27 No. 1, pp. 42-50.
SAS Institute (1988a), SAS/STAT User’s Guide Release 6.03, SAS Institute, Cary, NC.
SAS Institute (1988b), SAS/STAT User’s Guide Release Vol. 2, GLM_VARCOMP Version 6, 4th ed., SAS
Institute, Cary, NC.
Statistics Canada (2009), ‘‘International travel survey: mail-back questionnaires and air exit survey of
overseas visitors (ITS)’’, available at: www.statcan.gc.ca/cgi-bin/imdb/p2SV.pl?Function¼getSurvey&
SDDS ¼ 3152&lang ¼ en&db ¼ imdb&adm ¼ 8&dis ¼ 2 (accessed August 2009).
Sun, Y. (2005), ‘‘Marginal economic impact estimation for international visitors to Taiwan and policy
evaluation’’, Asia Paci?c Journal of Tourism Research, Vol. 10 No. 3, pp. 309-27.
Taiwan Tourism Bureau (2001), 2000 Annual Survey Report on Visitors Expenditure and Trend in Taiwan,
Taiwan Tourism Bureau, Taipei.
Taiwan Tourism Bureau (2002), 2001 Annual Survey Report on Visitors Expenditure and Trend in Taiwan,
Taiwan Tourism Bureau, Taipei.
Taiwan Tourism Bureau (2004), 2003 Annual Survey Report on Visitors Expenditure and Trend in Taiwan,
Taiwan Tourism Bureau, Taipei.
Travel Industries (2009), ‘‘Survey of international air travelers (in-?ight survey) program’’, available at:http://tinet.ita.doc.gov/research/programs/ifs/index.html (accessed August 2009).
Wang, M.Y., Huan, T.C. and Kan, T.C. (2012), ‘‘Inadequate return questions: return when? Sometime?’’,
International Journal of Culture, Tourism and Hospitality Research, Vol. 6 No. 1, pp. 44-53.
Woodside, A.G. and Martin, D. (2008), Tourism Management: Analysis, Behaviour, and Strategy, Cabi
Publishing, Wallingford.
Zalatan, A. (1998), ‘‘Wives’ involvement in tourism decision processes’’, Annals of Tourism Research,
Vol. 25 No. 4, pp. 890-903.
Zins, A.H. (2007), ‘‘Exploring travel information search behavior beyond common frontiers’’, Information
Technology and Tourism, Vol. 9, pp. 149-64.
Corresponding author
Hsiou-Hsiang Jack Liu can be contacted at: [email protected]
VOL. 6 NO. 1 2012
j
INTERNATIONAL JOURNAL OF CULTURE, TOURISM AND HOSPITALITY RESEARCH
j
PAGE 69
To purchase reprints of this article please e-mail: [email protected]
Or visit our web site for further details: www.emeraldinsight.com/reprints
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
2
:
1
8

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
This article has been cited by:
1. Tzung-Cheng T.C. Huan. 2016. Seeing further: Honoring John Urry's contributions to tourism and hospitality research. Journal
of Business Research 69, 1228-1233. [CrossRef]
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
2
:
1
8

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)

doc_634227191.pdf