Description
This paper aims to show the need for tourism researchers to identify clearly units of
observation and measurement.
International Journal of Culture, Tourism and Hospitality Research
Units, populations, and valid analyses
Chia-Huh J oy Liang Hung-Bin Chen Ming-Yang Wang
Article information:
To cite this document:
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 - 80
Permanent link to this document:http://dx.doi.org/10.1108/17506181211206261
Downloaded on: 24 January 2016, At: 22:18 (PT)
References: this document contains references to 29 other documents.
To copy this document: [email protected]
The fulltext of this document has been downloaded 308 times since 2012*
Users who downloaded this article also downloaded:
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. 54-69http://dx.doi.org/10.1108/17506181211206252
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
Daniel Briggs, Tim Turner, (2012),"Understanding British youth behaviors on holiday in Ibiza", International J ournal of Culture, Tourism and
Hospitality Research, Vol. 6 Iss 1 pp. 81-90http://dx.doi.org/10.1108/17506181211206270
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
)
Units, populations, and valid analyses
Chia-Huh Joy Liang, Hung-Bin Chen and Ming-Yang Wang
Abstract
Purpose – This paper aims to show the need for tourism researchers to identify clearly units of
observation and measurement.
Design/methodology/approach – Using examples focusing on international tourism, discussion
shows how terms commonly used in tourism research can be vague, ambiguous or invalid for
formulating theory or analysis or for generating and presenting research results.
Findings – A tourism research best practice is needed regarding identifying populations and
observation and measurement units so ambiguous or invalid use of terms like person, visits, visitor and
travelers does not occur and speci?c terms, for example, person-visit, person-visit-day and party-visit
are used to communicate clearly.
Originality/value – The paper clari?es the necessity of using terms like person-visit, person-visit days
and nuclear-family party visit to give research clear meaning and, in some cases, to avoid propagating
questionable or invalid analysis.
Keywords Unit, Dimensional analysis, Person-visit, Party-visit, International travel, Weighting, Tourism,
Sampling theory, Data analysis
Paper type General review
Introduction
This paper is to prompt discussion by raising issues. The underlying thesis is that ?aws are
occurring in tourism research because of vague use of terminology about who or what is
being considered both in theoretical and practical work. Reading of Beaman and Redekop
(1990) regarding weighting and units of survey estimates prompted thoughts about using
dimensional analysis. Fourier (1822), who developed procedures for ?nding wave structures
associated with mathematical functions, introduced dimensional analysis. Dimensional
analysis is a simple and powerful technique for learning about the dependence of a
phenomenon on the dimensions and properties of a system (see, e.g. Langhaar, 1951;
Rayleigh, 1915). Dimensional analysis is based on the simple principle that the units on each
side of any valid physical equation must be the same.
Discussion with colleagues of the units of tourism research resulted in recognizing the value
of applying dimensional analysis ideas in tourism research. Confusion exists in some
documents and articles about theory, data collection and analysis because units of
observation, measurement or analysis are inadequately de?ned or speci?ed. Unit of analysis
has been de?ned as ‘‘the major entity that is being analyzed in a study. It is the ‘what’ or
‘whom’ that is being studied’’ (e.g. see unit of analysis at www.socialresearchmethods.net/
kb/unitanal.htm; see units of measurement and of sampling and/or observation at http://
davidakenny.net/u_o_a.htm). A problem with unit of analysis being ‘‘the major entity being
studied’’ is that what the major entity is need not be clear. When data on people and their
trips are being studied to learn about decision making, is the major entity the person, trip,
person-trip, party-trip or possibly a decision?
PAGE 70
j
INTERNATIONAL JOURNAL OF CULTURE, TOURISM AND HOSPITALITY RESEARCH
j
VOL. 6 NO. 1 2012, pp. 70-80, Q Emerald Group Publishing Limited, ISSN 1750-6182 DOI 10.1108/17506181211206261
Chia-Huh Joy Liang is an
Assistant Professor in the
Department of Tourism and
Leisure Management,
National Penghu University,
Makung, Taiwan. Hung-Bin
Chen is an Assistant
Professor in the Department
of Hospitality Management,
National Penghu University,
Makung, Taiwan.
Ming-Yang Wang 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
)
Consider terminology used for units in tourism research. A person who in providing survey
data from home reports international travel in a certain time frame is an international traveler.
Also, someone interviewed on reentering his or her home country is an international traveler.
However, for a period of time, say 2008, the number of people in a country who can report
traveling internationally with return in 2008 is less than the number of international travelers
returning, unless all international travelers in 2008 make only one trip in 2008. Ambiguity exits.
Furthermore, when data are collected from people at their homes, one can give demographic
distributions for unique people who travel internationally (i.e. people returning fromat least one
international trip in 2008). One may ?nd demographic information (e.g. age-gender-income
distributions) published so marketers understand the attributes of international visitors. Now
consider getting travel data for 2008 by collecting data from visitors returning from abroad or
from foreign visitors who are departing, for example, from Taiwan in 2008. Is giving
demographic distributions by tabulations for these international visitors the same as getting
demographics information from data collected when people are at home? The unit of
observation for people arriving back in a home country or leaving a country is person-visit and
tabulations of demographics for a randomsample froma population of person-visits is not the
same as for the observation unit of people of the population that traveled internationally in
2008. Using person-visit data (e.g. by weighting) to give distributions for unique people that
traveled abroad is not trivial unless, for example, one has data from customs and immigration
with people identi?ed (e.g. by passport number) so a person with multiple international trips is
only counted once. One can see Tyrrell and Johnston (2002) and Johnston and Tyrrell (2003)
for considerations in determining unique visitors. Unique is time frame dependent (e.g. for a
month, year or life time up to a point in time). Beaman and Redekop (1990) refer to considering
unique visitors in a year in allowing one vote or expression of opinion rather than letting
frequent users (say in 2008) have a vote per visit.
The ideas of the last paragraph show the need to be clear about units used in tourism
research. Traveler, visitor, tourist or some combination of these is the only terminology one
?nds in many articles. Person-visit is found in lists of de?nitions used by agencies like Tourism
Manitoba (http://ti.travelmanitoba.com/assets/pdf/tourism_de?nitions_and_concepts.pdf)
and Statistics Canada (e.g. see Statistics Canada, 1997). The term is also found in some
practically oriented material (e.g. see South Grow, 2009). Neither the term person-visit nor
person-visit-day is de?ned in the Dictionary of Travel, Tourism and Hospitality (Medlik, 2003).
Is tourism research science? If so, what does that imply? In the physical sciences if units are
not made clear, research is recognized as ?awed. If you are dealing with acceleration and
you report a result in meters per second or give a number and write that an object speeded
up, people know your result is invalid or the research or its reporting lacks rigor. If
researchers use terms like international traveler, party or visitor in ambiguous ways, is the
research scienti?c? Is using terms like person-visit or party-visit-days necessary? Is the use
of such undesirable because using them is not necessary? Do some researchers see being
precise as distracting from the readability of an article or only being unnecessary for survey
designers and statisticians? Beaman and Redekop (1990) show that international mountain
park visitors to Canada can be described as majority foreign or majority domestic by
changing the unit of measurement. Given their result, presumably, how you measure a
majority of visitors being foreign, impacts planning and management decisions. From
another perspective, while one can commend South GrowRegional Initiative for reporting on
person-visits to southern Alberta as a percent of person-visits to Alberta, if impact of the
visitors relates more closely to party-visit-days and person-visit-days than person-visits,
should reporting be different? While person-visits or party-visits to a country re?ect the work
load at entry/exit points, the case is made subsequently that party-visit days and person-visit
days spent circulating about a country relate to locals’ perceptions about what visitors are
like and to services used and thus services demanded.
Research context and literature
Unit of measurement mattering has been illustrated. However, to clarify the scope of this
article, consider information on problems with a service or reactions to an offering (e.g. see
VOL. 6 NO. 1 2012
j
INTERNATIONAL JOURNAL OF CULTURE, TOURISM AND HOSPITALITY RESEARCH
j
PAGE 71
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
)
Cooper and Schindler, 2003, p. 200). A unit of observation, member of a population, need
not be a person. All complaints ?led in 2006 by visitors to a country is a population (i.e.
population of complaints ?led). If complaints are on a form, a complaint form is the unit of
observation. Tabulations of such complaint information should not be treated as if one has
complaint information obtained by a survey of exiting visitors. Even relating the number of
complaint forms ?led in a month to, for example, total person-visits in the month can be
misleading unless all complaints from a person for a visit are counted as appearing on one
form. When a person can ?le multiple complaint forms, dividing numbers of complaints
forms by the number of person visits can give a false impression about complaining. Say 1
percent of person-visits result in any complaint, those who ?led multiple complaint forms
could result in complaint forms per person-visit being 2 percent. In a survey, one can have a
similar situation. While only 2 percent of visitors register one or more complaints, one can
have 70 complaints from 20 people. Most, say 60, complaints could come from ten people
who complain about a lot of things. Simply stating that complaints per person-visit was 7
percent does not convey the fact that only 1 percent of visitors had more than one complaint.
Raising issues about good reporting is not to suggest that complaints ?led about problems
with visiting a country or with services at any destination should not be investigated.
However, investigation is facilitated if complaints are collected with information that
facilitates recognizing how complaints relate to complainants and potential complainants.
The ideas and issues introduced do not just apply to tourism or leisure researchers. In the
business oriented text of Cooper and Schindler (2003, p. 179) one reads: ‘‘a population
element is the subject on which measurement is being taken.’’ Subsequently, they state, ‘‘the
de?nition of the population may be apparent from the management problem or research
question(s) but often it is not (p. 186).’’ The text gives practical examples and acknowledges
‘‘confusion about whether the population should be persons, households, or families, or a
combination of these’’ should be considered in analysis. Cooper and Schindler are
acknowledging that measurement and analysis can involve multiple units. Does a single
appropriate unit of analysis exist for survey data collected from, for example, international
leisure travels returning home or exiting a given foreign country (e.g. Taiwan)? From material
earlier, the answer is that in general no single unit exists. One may need multiple units to
achieve research objectives.
To amplify on units in observation and analysis, consider real data. In the Canadian Travel
Survey (CTS) households are selected and a person 16 þ in the household is selected so
that the unit of observation for weighted data is a member of the non-institutional population
of residents of Canada 16 þ (Statistics Canada, 1997). Now, this article is not endorsing
continual repetition of information like ‘‘non institutional population of residents of Canada
16 þ ’’ or providing further details on why the sample is not perfect. Rather the concern is
that at some point a clear statement is made. For Taiwan’s survey similar to the CTS, phone
numbers are selected and a person 18 þ associated with the phone number is selected to
report on travel (Taiwan Tourism Bureau, 2008). The population for the survey is reported as
Taiwanese residents 18 þ . With weighting the unit of observation is a Taiwanese resident.
For both Canada and Taiwan, person weights must be calculated to relate a person
appropriately to the population sampled. For Canada, data on all trips taken in a month are,
in principal, recorded (re accurate reporting see Beaman et al., 2001). Therefore, Statistics
Canada (e.g. Statistics Canada, 1997) views weighted trip data as approximating a random
sample of person trips by the CTS population. Given the person-trip sample, the person
weight and a zero-one variable (i.e. one is traveled) can be used to estimate the number of
people in the population or segment who are international travelers in a given month (i.e. time
period for monthly data are one month). Also for Canada, data on the number of international
trips taken by a person can be weighted by a person’s weight and added over a group of
people (e.g. people in a segment) to calculate the number of international person trips taken
by segment members. One might think that in Taiwan one could make estimates like those
made using the CTS. If one could and the estimates agreed with the return records kept by
of?cials who record every return by residents, one could check on survey accuracy or bias.
However, from survey documentation one learns that to prevent data collection from taking
too long, data are only collected on up to three trips. If trips to be reported were selected at
PAGE 72
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
)
random from some preliminary list of all trips, after weighting maybe Taiwan would have a
sample representative of person-trips of the survey population. As the survey is executed,
one has a sample of an ill-de?ned population. One should not make estimates as if trip data
represented person-trips of the population.
Now, consider that research often reports on travels with other people, on travel in a party.
should one report person-trips or party-trips? For parties that travel in one vehicle and
occupy one hotel room, service providers concerned with vehicle transportation and
accommodation presumably want party-trip information. Recognizing that different clients
for statistics ?nd one or the other unit useful, Statistics Canada provides weights for
computing person-trips and party-trips. However, for the CTS parties are only made up of
household members. Friends from different households traveling together for CTS are not
traveling in a party. Knowing what party means in for data or for a theory matters. Units matter
for other than technical reasons. They matter because different units are appropriate in
different contexts.
The picture emerging is that more attention should be paid to units for which theory is
formulated, models are built, analyses are run, etc. Thinking about issues and concepts
provided shows the need for and value of units. A person can be at an international
destination and return home several times in a year. However, just as length of stay can affect
the probability of selecting a person or party (Lucas, 1963; Scheaffer et al., 1996; Sheaffer
and O’Leary, 2005), length of stay can affect the probability of selection of people returning
from abroad. A person who comes back from international trips several times in a year has
more chance of reporting her or his travel than a person who return once. When people who
make a lot of trips are treated as unique people, one gets a distorted picture of travel by
unique people in a country. For trips ending, for example in a month or year, total expenditure
calculated from person-visit data gives a valid estimate of expenditures abroad. However,
average expenditure per person-trip is not the same as average expenditure per person that
travels. One is back to using the right unit for the right purpose.
Problem statement and research strategy
While Beaman and Redekop (1990) focus on response weighting in their discussion of units,
this paper takes a general and non-technical approach to the need for speci?cation of units.
Given the theme of the special issue in which this article appears, examples and discussion
tend to address international tourism. However, hypotheses are general. The generality is
appropriate since material presented is easily altered to apply to other aspects of tourism,
for example, to domestic tourism. The general thesis is that the terminology used in tourism
research should appropriately identify units. As already introduced, evidence exists that
terms for units of observation, measurement and analysis are being used vaguely and
inappropriately. H1 and H2 formalize ideas implicit in the thesis:
H1. For tourism research to be credible science, researchers need to use precise
terminology for units of observation, measurement and analysis in theory
development, in planning of research, in analysis of data and in presentation of
results.
H2. Components of formulations of theory and of data based research are meaningless
when units that statements are about or apply to are not clearly de?ned.
Sampling caveat
A matter that has not been raised is that survey results can be biased because of sampling,
recall or other bias (Burton and Blair, 1991; Krosnick, 1999). A way to ?nd out if a survey
estimate is biased is to compare the estimate with an accurate benchmark (Vaske et al.,
2005). Such measurement of bias has been done for international travel. Statistics Canada
has caused a bias in a person-visit ?ow. The bias, caused by changes in data collection
methods Statistics Canada considered innocuous is estimated at 10 percent (e.g. see
Beaman et al. 2001). To avoid making qualifying about bias in research results a ‘‘no bias
assumption’’ is accepted:
VOL. 6 NO. 1 2012
j
INTERNATIONAL JOURNAL OF CULTURE, TOURISM AND HOSPITALITY RESEARCH
j
PAGE 73
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
)
No bias assumption: all potential respondents to a sample survey respond (no non response)
and, respond accurately; and the sample is, given weighting, representative of its population
within the limits of statistical variability.
Making the no bias assumption may seem inappropriate or unnecessary. The assumption is
made because the authors feel that without it, some of the statements in following sections
should be quali?ed or clari?ed. In fact, making the assumption does not compromise this
research. Making the assumption simply acknowledges that in much tourism research the
no bias assumption is accepted. As implied above, for the CTS Statistics Canada assumes
bias is negligible, though statistically signi?cant violation of the no bias assumption is
established. In other words, accepting the assumption is not an assertion that the
assumption is correct, or a good approximation, where adopted. Rather, just as making a
?awed measurement of a magnitude in the physical sciences does not affect the unit that
should be used, recognizing correct units that apply to tourism theory or statistics is a
separate matter from, for example, unbiased measurement.
Research strategy
The research strategy of this paper is simple. Introductory material has provided some
ideas. The analysis and results section develops those ideas. The Discussion pursues
issues and implications of matters covered in Analysis and results. In concluding, discussion
turns to the need for a best practice for tourism research of appropriately identifying units of
observation, measurement and analysis.
Analysis and results
Based on what has been presented earlier in this paper, issues exist regarding units to use
and being used in tourism research. From a dimensional analysis perspective, a key matter
is to whom or to what statements refer. Consider a statement like a ‘‘sample of visitors was
obtained as the visitors entered ?ight departure areas’’ or ‘‘the table gives demographic and
expenditure information for tourists’’. Referring to travelers, tourists or visitors instead of
person-trips may make some readers comfortable. However, do the statements given make
clear who is in the sample (i.e. the population) or if demographic information is for persons or
person visits? If linguistic variety and readability yields questionable or invalid interpretation
of results or hides ?awed formulation of research, tourism research is not good science.
General considerations
Thinking about units and people’s tourism related activities is facilitated by considering data
collection about travel. A type of data collection introduced and the one for which a subsection
occurs in this section is collection of person and trip data frompeople at home. Another type of
data collection already commented on is for complete destination/trip experiences.
Statements above suggest serious problems in making inferences about unique people at
an origin traveling from complete destination/trip experiences (e.g. see truncation bias in
Smith and Desvouges, 1985; or Phaneuf and Smith, 2006). The problemis unit related. People
who do not travel are not in the person-visit populations data collected. One can make
inferences about people who did not travel but are like those that did. For example, in
modeling one can infer that international travelers from a country who tend to travel
internationally with a given frequency (e.g. about every two years) are like an equal number of
people at home. Units problem exists when research treats person-visit information as person
information. Collection of data on complete destination/trip experiences is not pursued in this
section because hypotheses can be accepted without further consideration of such sampling.
Intercepting people during an experience and observations such as instant counts are even
introduced in this article because conceptual matters regarding units of analysis and data
collection can be intricate (Schreuder et al., 1975).
Sampling to get data for making valid estimates for speci?c populations involves many
issues. Different general sources on sampling are useful. Though dated, Sudman’s Applied
Sampling (Sudman, 1967) is readable and covers a broad range of material. Scheaffer et al.
(1996) cover such matters as systematic sampling (Ch. 7), cluster sampling (Ch. 8-9) and
PAGE 74
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
)
estimating population size (ch. 10). Many other publications are listed with annotation at
www.bettycjung.net/Surveys.htm useful (accessed August 2009).
Sampling, party and being speci?c
Examples illustrate matters of concern. In the statement above about sampling, consider
interviews were with non institutionalized persons 16 þ with data collected on all trips of
certain types (e.g. over 50km from home and not work related) ending in 2008 and by
weighting data yield unbiased estimates for the population, for person-trips and for household
party trips. Given caveats on the population sampled and about observations representing
their population, the article is not suggesting continual repletion of caveats. If the term
person-trip has a consistent meaning in research being reported, quali?cations such as
pleasure person-trip of 16 þ etc. are only needed when, for example, segments are being
compared. However, saying too little in specifying the population to which data apply can be a
problem. If data collectors approach parties (e.g. in departure areas of an airport) and the
party selects who responds to a questionnaire, does a problem arise? Well, if attitude
questions are asked of the respondent, one has responses that are meaningful for person-visit
data but the data are collected for arbitrary members of parties. One does not have a sample
from the population of person-visits and cannot correct to that population by weighting.
Now, consider that a questionnaire being used was developed to collect party data. What
does that mean? In the CTS the party unit is household-party. However, if different clients for
information are interested in different types of parties (e.g. accommodation-party being
people who stay together), reason suggests collecting person-visit data with information that
allows analysis for different kinds of parties. Nevertheless, data collection for person-visits
can involve problems. Prorating party expenditures to a person can be appropriate because
much spending by a household can be for the party. If party expenditures are collected in
person-visit data, adding the expenditures or taking an average mixes units of party-visit
and person-visit. Prorating expenditures based on party size can give correct aggregate
expenditures but amounts (e.g. on gifts) do not necessarily apply to the respondent.
Analysis of amounts using respondent demographics is mixing units of analysis. Maybe,
someone has a use for the average party expenditure associated with a person-visit.
However, an application is not obvious to the authors. In another context, consider that for
the people of a country one international person-trip occurs per 2 residents of the country
(e.g. for residents/people 16 þ ). Writing that 50 percent of people of the country over 15
traveled internationally is not valid unless each traveler made only one international trip. The
problem with calculating 50 percent is that units of measurement are not properly
considered. Person trips divided by persons is average person-trips per person. The
number of persons who made any international trip (e.g. in 2008) divided by population size
is the percent of the population traveling internationally. Units matter.
Origin data collection on international travel
Consider formulating theory about travel behavior of and collecting data froma population in a
particular geographic area (e.g. a country). As for the CTS, the population could be
non-institutionalized Canadians 16 þ . To formulate theory, the units to which the theory
applies should be unambiguously identi?ed. Consider the theory is about international
pleasure trip taking people by people of a country. To test the theory, assume a representative
sample of the population is obtained. Representative means that data collected when
weighted can be used to make estimates for populations (e.g. population of the country,
person-trips, party-trips and person- and party-trip-days). If sampling and weighting are
perfect, estimates are unbiased. Data collection and weighting can involve sophisticated
considerations, as one can see from documentation for the CTS (Statistics Canada, 1997).
When one starts to think about international travel behavior of individuals, the unit being
considered can move from person to person-trip, party-trip or some other unit. Regarding
individuals and their behavior, one may bene?t from articles that examine behavior based on
long interviews with individuals (e.g. see Martin, 2008; DeCrop, 2001; March and Woodside,
2005, chs. 1-3). However, here assume concern is with international pleasure trips that are not
mainly devoted to visiting friends and relation (VFR) and are made in a year, Y (e.g. trips
VOL. 6 NO. 1 2012
j
INTERNATIONAL JOURNAL OF CULTURE, TOURISM AND HOSPITALITY RESEARCH
j
PAGE 75
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
)
ending in Y). Think about structured survey data being collected. Think of data being collected
that applies to a respondent, person data. Think of these data in a table with one row for each
person. Also, for those who travel internationally, think about data for trips being in another
data set. Think of trip data as a table with a rowfor every trip. For a person, typically data would
include age, gender and family membership information. Person data could include attitudes
and preferences information. For trips data can include information about trip planning, travels
mode(s), destination(s), distance, duration, activities, services used; and ratings (e.g. of
destinations, services and the trip). Consider that for each person a person-weight is de?ned
and the person-weights of all people sumto the total people in the population sampled. For all
people reporting an international trip or trips, you can visualize a link between the person’s
data in one table and her or his trip data in the trip-data table. Mathematically, a person is
associated with (e.g. linked by an ID) to each international trip he or she reported.
The last paragraph has presented a relational model of persons taking trips (i.e. taking is a
word that describes the link between tables). The structure described is similar to the
structure used by Statistics Canada (1997) in storing trip data for Canadians. Now, if you put
a person’s weight with each trip (i.e. in the person-trip table), you can call it a person-trip
weight. In analysis of trip data, the weight of the trip indicates how many trips like the given
one the survey indicates were made in a time frame (Y) being considered. Add the
person-trip weights up for all trips and you get total person-trips for Y. If trip-weights are
added for a segment for trips with certain attributes, one obtains an estimate of the number
of trips with the given attributes made by the segment members in Y. For CTS party size is
used to create a party-trip weight. As for person-trips, estimates can be made for segments
or the population.
Several matters already raised should be noted. Adding person-weights of people who
traveled internationally for pleasure in Y (adding weights for unique people who traveled) is not
the same as adding trip-weights for trips (i.e. using data in the person-trip table) given some
people took one trip while others took two or more trips. Therefore, if you have a theory and/or a
model about people taking international trips, do you test your theory usingperson data (e.g. 1
took an international trip and 0 did not) to test your theory? How do trip attributes of particular
trips and number of trips enter into theory and/or model? Given your theory is about a person
taking trips, can you justify using attributes of a person and associating these with information
about person-trips? With such modeling, you are presumably getting some aggregate idea
about attributes of people in?uencing trip taking. Many years ago Cicchetti et al. (1969; see
also CORDS, 1976, ch. 1) proposed that for the kind of modeling being described, two models
are needed. One is for the prediction of participation-nonparticipation. The other model is
estimated for those that traveled with number of trips as the dependent variable. However, if
number of trips is the dependent variable, where do attributes of trips enter into modeling?
Party travel
Commentary so far has focused on the person. Consider you are interested in party travel. In
generating information about party travel using person-trip data, person-trip weights must
be adjusted. Because of how Statistics Canada de?nes party (household-party) and
because numbers in the household going on trips is recorded, party-trip weights can be
calculated fromperson-trip weights (e.g. see Statistics Canada, 1997). However, if you are in
Taiwan, you know that a lot of international party travel is by people who do not live together
(are not in the same household). What good is the unit party if it is de?ned in a way that is not
appropriate for many trips. Party has a meaning to a particular tourist for a trip. Furthermore,
party can have different meaning for tour operators, accommodation providers and tour
guides. Outside the con?nes of a strict de?nition, for example by an agency or in a research
project or component of a project, party is ambiguous. Nevertheless, one can recognize
party members of different types of parties playing a critical role in decision making about
international travel by having such quali?ed terms such as household-party and
friends-party. Using such terms can clarify literature on people in?uencing decisions (for
such literature see part 4 of Crouch et al., 2004). Anyway, one of the tourism research terms
that is ambiguous is party. Can party be a unit of analysis when it means something different
for different trips?
PAGE 76
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
)
Use of terms and their meaning
The ideas already provided in this section are adequate to showanalysis of data about travel
obtained from people at their usual place of residence can involve different units
(e.g. person, person-trip and party trip). However, units such as person-trip are not found in
abundance in the travel research literature. Articles about trip decision making are common
and typically focus on a person and a single trip (e.g. see Crouch et al., 2004, part 4; or
Mazanec et al., 2001, part 3). A search of the tourism literature did not produce any article
devoted to explaining the diversity of people’s person-trips taken in a time period. The subtle
matter that is arising with the use of words is ambiguity. For some clients for information such
as airlines, person-trips may be desirable because person-trips relates to passenger
volume. For hotels party-visit data may seem desirable because occupancy relates to
party-trips. However, do Asian groups that call themselves a party occupy one room? In
Taiwan some accommodation is set up so Japanese groups of 10 þ people sleep in a room.
In closing this consideration of units in relation to origin surveys, estimating percent traveling
is important since destination surveys only yield information about people who do not visit by
statistical inference (see, e.g. Phaneuf and Smith, 2006). One reason for origin surveys
measuring travel by members of a population or a segment. As introduced above, even with
data letting one know the number of people returning from international travel (e.g. from
Taiwan’s of?cers recording all returns), the ratio of international person-trips in a year (Y) to
persons in Taiwan is not the percent of Taiwanese traveling internationally in Y. Also, with
origin data on travel one may ?nd an average of 33 percent of persons (e.g. 16 þ ) travel
internationally in a given year. International person-trips per person in the country could be
0.66. The average number of person-trips per person is 0.66. As 66 percent this is
person-trips as a percent of the population. However, 0.33 is the proportion of the population
traveling internationally. Expressing units correctly matters. By month one may ?nd in
February only 5 percent of a population travels internationally while in July the percent is 15
percent. Since the same 0.05 percent that traveled in February may travel in other months,
monthly percent-traveling ?gures only set a lower limit for the percent traveling annually.
Time frame being a month or being a year matters in specifying a rate. Furthermore, while
person-trips like dollars can be totaled over time, adding numbers traveling internationally
for months does not yield total numbers who travel internationally. Valid operations depend
on units. Just as using the correct numerator in making a computation matters, one needs to
pays attention to the units in determining is a computation is appropriate.
Hypothesis acceptance/rejection and discussion
Repeatedly commentary in this paper has shown that units matter. Data collected frompeople
can be used to produce such estimates as the number traveling, the number of people in a
population traveling internationally, their person-trips made, their number of party-trips, their
number of party- or person-trip-days, etc. What a number speci?es being clear is important.
Because the matter the need for careful use of units is addressed by examples, reference to
papers with ambiguous use of terms and or other problems with units are not given. Therefore,
accepting or rejecting H1 comes down to whether the paper has built a solid case for broader
use of units such as person-visit and person-visit-days and more careful use of terms such as
visitor or traveler. The authors believe the case is made and accept H1.
H2 expresses why having and using an appropriate units (i.e. of observation, measurement
and analysis) is necessary. Stating ‘‘components of formulations of theory and parts of data
based research are meaningless when units that statements apply to are not clearly
de?ned’’ may seem to be exaggeration. However, making a strong statement is reasonable.
The Beaman and Redekop (1990) example that a statement about a majority of visitors being
foreign being true or false depending on the unit, the way of measuring, shows units must be
speci?ed. Nevertheless, logic also provides proof. Consider a planning application that cites
person-visit statistics when person-trips or person-visit days relate to the demand for
service. Change in person-visit statistics may give an idea of what change in demand is.
However, if person-visits are going up and length of stay is going down, person-visit trend
data can suggest growing demand when demand is declining. In other words, logic as well
VOL. 6 NO. 1 2012
j
INTERNATIONAL JOURNAL OF CULTURE, TOURISM AND HOSPITALITY RESEARCH
j
PAGE 77
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
)
as observation supports recognizing that person-visit and person-visit-day can have
differing trends (e.g. one growth and the other decline). Intuitively, person visits related to
what you see at entry and exit from a country and (2) person-visit-days relate to who you see
while party-visit-days relate to encounter with parties being served (e.g. in accommodation).
An example facilitates understanding issues. Assume people coming to a resort are
dominantly from the UK and from Spain. Assume UK visitors on average have a party size of
three and stay a week but half as many UK party-visits come as for Spain. Assume Spanish
visitors have an average party size of three and spend about the same per day as UK
visitors. Are the Spanish by far the most important group supporting the resort? If Spanish
visitors stay an average of two nights, visitors from UK rather Spanish visitors are the
day-to-day majority. The UK visitors should tend to be encountered more than twice as often
in daily business with visitors. Being the majority, in the sense presented, has clear
implication for planning and management. People responsible for marketing should
certainly not be basing plans strictly on volume of entry/exit (i.e. on person-visits). Starting to
get a valid picture of the role of visitors in the economic and social activity of an area is a step
in good research. For practical work, policy, planning or management justi?cation of action
should be rational. If numbers are used without knowing their units, logic is ?awed.
Though meaningless is a strong word to use, research is not scienti?c unless logic justi?es
what is done. Accepting H2 just acknowledges that components of tourism research have
no scienti?c meaning when what statements are about is not clear. Therefore, H2 is
accepted recognizing that meaningless applies to those aspects of any tourism research in
which statements about who or what are not clear, units are not clearly de?ned. In other
words, practical research and theory can be valid in some aspects and ?awed in others.
Need for best practices
Given conclusions reached, speci?cation of units is important to tourism research being
correctly interpreted or even valid. Practical or applied research involves observation.
Cooper and Schindler (2003, p. 659) maintain that a technical report should contain
‘‘suf?cient procedural information to allow others to replicate the study.’’ Kerlinger and Lee
(2000, p. 570) in commenting on the experimental versus the nonexperimental methods
state: ‘‘replication is always desirable, even necessary.’’ The Public Opinion Quarterly (POQ)
in its Notice to Contributors makes clear the importance that is placed on documentation that
allows for survey replication by making ‘‘at least approximate replication’’ a requirement
(www.oxfordjournals.org/jnls/list/poq/instauth/). However, replication can not occur without
correct interpretation of data. Using correct units of analysis in practical or applied research
is about describing things accurately enough so concepts, estimates, etc. are used and
interpreted correctly. Tourism research has no practical value unless application to
observing and describing things exists. If, for example, a theory about trip decision making
is unclear about units of data to use to con?rm the theory, clari?cation of the theory is
needed. Given person-visits can increase while person-visit-days decrease, theory,
propositions based on theory and analysis should be speci?c about units. As discussed,
treating a theory on deciding about taking a particular trip as a theory for predicting an
individual’s person-visits (e.g. in a year) is not be valid when causal factors relate to type of
trip. If one is formulating an aggregate model, the model should be described as such.
Bene?ts can arise from examining various trip decision theories (e.g. see chapters 1 to 3 of
March and Woodside, 2005). Worthwhile research would be examining models and related
theories to determine which are truly for individuals and which are aggregate (over people or
trips). For aggregate a concern would be identifying models/theories described, for
example, as if they apply to individuals. Furthermore, do some theories just apply to an
individual’s particular trip while some are more general? Classi?cation would facilitate
improving research.
Given tourism research based on vague units yields questionable theory and applied
research, having a best practice is desirable. Based on what has been presented, a tourism
research best practice could be:
PAGE 78
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
)
Best practice: units used in formulating theory or data collection and units used in analysis and
presenting results should be precisely de?ned.
Appropriate use of units depends on professional researchers and academics being
leaders in correct use of units. Promoting use of clearly de?ned units can occur in different
ways. Material being taught in universities and in professional development courses having
correct units is necessary in building a community that understands the need for and use of
units such person-visit and person-visit day. Fostering good practice depends on
constructive critique of the use of vague or invalid terms or faulty description of populations
and data collection in presentations at meeting and in articles/chapters of professional
books and journals. Also, encouraging rigorous use of units in practical (e.g. agency and
consultant reports) reports is necessary to get people to abandon dispositions to use terms
inappropriately or vaguely. The best practice suggested can have immediate
consequences if academic tourism research journals take a lead in causing change by
requiring appropriate use of units. An important step will be taken if journals such as the one
containing this article enforce rules preventing ambiguity whether in use of it or this or in use
of unit related terms vaguely or inappropriately.
Conclusion
This research introduces the importance of better use of units in applied and academic
tourismresearch. When units that apply to a statement, whether in an equation or formulation
of theory are not consistent, a problem exists. Terms like trip, visits, travelers or visitors can
be quite adequate for informal conversation. However, clear speci?cation of units is needed
when not doing this can possibly result in confusion. Yes, in some cases, context can make
clear what is meant by trips, tourists or visitors. However, as illustrated above, in some cases
accurate de?nition of populations, units of observation, measurement and analysis is critical.
Good science should involve precision rather than linguistic variety and vagueness. That
some readers are uncomfortable with terms that accurately state units of measurement and
observation shows a problem with what the reader has learned not a problem of literary
ability of the writer. A best practice, such as introduced above, is necessary so better
tourism research will occur.
References
Beaman, J.G. and Redekop, D. (1990), ‘‘Some special consideration in weighting survey data’’,
Les Cahiers du Tourisme, Centre des Hautes Etudes Touristiques, Aix-en-Provence, (Aix-En-Province),
France Se´ rie C # 118.
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’’, Consumer Psychology of Tourism,
Hospitality and Leisure, Vol. 2, pp. 49-65.
Burton, S. and Blair, E. (1991), ‘‘Task conditions, response formulation processes and response
accuracy for behavioral frequency questions in surveys’’, Public Opinion Quarterly, Vol. 55 No. 11,
pp. 50-79.
Cooper, D.R. and Schindler, P.S. (2003), Business Research Methods, 8th ed., McGraw-Hill, NewYork, NY.
CORDS (1976), ‘‘Canadian outdoor recreation demand study: volume 2 technical notes (provisional
copy)’’, available at: www.cals.uwaterloo.ca/pubs/ (accessed December 2009).
Crouch, G.I., Perdue, J.R.R., Timmermans, H. and Uysal, M. (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, New York, NY, pp. 333-47.
Fourier, J.B.J. (18221955), The´ orie Analytique de la Chaleur, ‘‘The Analytical Theory of Heat’’, Dover
Publications, NewYork, NY, replication of the English translation that ?rst appeared in 1878 with previous
corrigenda incorporated into the text, by Alexander Freeman.
VOL. 6 NO. 1 2012
j
INTERNATIONAL JOURNAL OF CULTURE, TOURISM AND HOSPITALITY RESEARCH
j
PAGE 79
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
)
Johnston, R.J. and Tyrrell, T.J. (2003), ‘‘Estimating recreational user counts American’’, Journal of
Agricultural Economics, Vol. 85 No. 3, pp. 554-68.
Kerlinger, F.N. and Lee, H.B. (2000), Foundations of Behavioral Research, 4th ed., Harcount, Singapore.
Krosnick, J.A. (1999), ‘‘Survey research’’, Annual Review of Psychology, Vol. 50, pp. 537-67.
Langhaar, H.L. (1951), Dimensional Analysis and Theory of Models, John Wiley, New York, NY.
Lucas, R.C. (1963), ‘‘Bias in estimating recreationists’ length of stay form sample interviews’’, Journal of
Forestry, Vol. 61 No. 2, pp. 912-3.
March, R. and Woodside, G.A. (2005), Tourism Behavior: Travelers’ Decisions and Actions, CABI,
Wallingford.
Martin, D. (2008), ‘‘Grounded theory of international tourism: building systematic propositions fromemic
interpretation of Japanese travellers visiting the USA’’, in Woodside, G.A. and Martin, D. (Eds), Tourism
Management: Analysis, Behavior and Strategy, CABI Publishing, Wallingford.
Mazanec, J.A., Crouch, G.I., Brent Richie, J.R. and Woodside, A.G. (2001), Consumer Psychology of
Tourism, Hospitality and Leisure, Vol. 2, CABI Publishing, Wallingford.
Medlik, S. (2003), Dictionary of Travel, Tourism and Hospitality, Elsevier, Amsterdam.
Phaneuf, D.J. and Smith, V.K. (2006), ‘‘Recreation demand models’’, in Ma¨ ler, K.G. and Vincent, J.R.
(Eds), Handbook of Environmental Economics, Elsevier, Amsterdam, p. 2.
Rayleigh, L. (1915), ‘‘The principle of similitude’’, Nature, Vol. 95, pp. 66-8.
Schreuder, H.T., Tyer, G.L. and James, G.A. (1975), ‘‘Instant- and interval-count sampling: two new
techniques for estimating recreation use’’, Forest Science, Vol. 21 No. 1, pp. 40-4.
Scheaffer, R.L., Mendenhall, W. and Ott, R.L. (1996), Elementary Survey Sampling, 5th ed., Wadsworth
Publishing Co., Belmont, CA.
Sheaffer, A. and O’Leary, T. (2005), ‘‘Noncommercial ?sh consumption and anglers at risk’’, Human
Dimensions of Wildlife, Vol. 10 No. 4, pp. 229-38.
Smith, V.K. and Desvouges, W.H. (1985), ‘‘The generalised travel cost method and water quality
bene?ts: a reconsideration’’, Southern Economic Journal, Vol. 52, pp. 371-81.
South Grow (2009), ‘‘South Grow Regional Initiative’’, available at: www.southgrow.com/_data/
datadocs/southgrowtourismindustry4.pdf (accessed July 2009).
Statistics Canada (1997), Canadian Travel Survey Microdata User’s Guide, SC document identi?cation
number 87M0006GPE, Statistics Canada, Ottawa.
Sudman, S. (1967), Applied Sampling, Academic Press, New York, NY.
Taiwan Tourism Bureau (2008), 2007 Annual Survey Report on Visitors Expenditure and Trend in Taiwan,
Taiwan Tourism Bureau, Taipei.
Tyrrell, T.J. and Johnston, R.J. (2002), ‘‘Estimating regional visitor numbers’’, Tourism Analysis, Vol. 7
No. 1, pp. 33-41.
Vaske, J., Beaman, J.G. and Miller, C. (2005), ‘‘Cognitive processes in hunters’ recall of participation and
harvest estimates’’, Journal of Wildlife Management, Vol. 69 No. 3, pp. 967-75.
Corresponding author
Ming-Yang Wang can be contacted at: [email protected]
PAGE 80
j
INTERNATIONAL JOURNAL OF CULTURE, TOURISM AND HOSPITALITY RESEARCH
j
VOL. 6 NO. 1 2012
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. Stefano De Cantis, Anna Maria Parroco, Mauro Ferrante, Franco Vaccina. 2015. Unobserved tourism. Annals of Tourism Research
50, 1-18. [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_864823900.pdf
This paper aims to show the need for tourism researchers to identify clearly units of
observation and measurement.
International Journal of Culture, Tourism and Hospitality Research
Units, populations, and valid analyses
Chia-Huh J oy Liang Hung-Bin Chen Ming-Yang Wang
Article information:
To cite this document:
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 - 80
Permanent link to this document:http://dx.doi.org/10.1108/17506181211206261
Downloaded on: 24 January 2016, At: 22:18 (PT)
References: this document contains references to 29 other documents.
To copy this document: [email protected]
The fulltext of this document has been downloaded 308 times since 2012*
Users who downloaded this article also downloaded:
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. 54-69http://dx.doi.org/10.1108/17506181211206252
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
Daniel Briggs, Tim Turner, (2012),"Understanding British youth behaviors on holiday in Ibiza", International J ournal of Culture, Tourism and
Hospitality Research, Vol. 6 Iss 1 pp. 81-90http://dx.doi.org/10.1108/17506181211206270
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
)
Units, populations, and valid analyses
Chia-Huh Joy Liang, Hung-Bin Chen and Ming-Yang Wang
Abstract
Purpose – This paper aims to show the need for tourism researchers to identify clearly units of
observation and measurement.
Design/methodology/approach – Using examples focusing on international tourism, discussion
shows how terms commonly used in tourism research can be vague, ambiguous or invalid for
formulating theory or analysis or for generating and presenting research results.
Findings – A tourism research best practice is needed regarding identifying populations and
observation and measurement units so ambiguous or invalid use of terms like person, visits, visitor and
travelers does not occur and speci?c terms, for example, person-visit, person-visit-day and party-visit
are used to communicate clearly.
Originality/value – The paper clari?es the necessity of using terms like person-visit, person-visit days
and nuclear-family party visit to give research clear meaning and, in some cases, to avoid propagating
questionable or invalid analysis.
Keywords Unit, Dimensional analysis, Person-visit, Party-visit, International travel, Weighting, Tourism,
Sampling theory, Data analysis
Paper type General review
Introduction
This paper is to prompt discussion by raising issues. The underlying thesis is that ?aws are
occurring in tourism research because of vague use of terminology about who or what is
being considered both in theoretical and practical work. Reading of Beaman and Redekop
(1990) regarding weighting and units of survey estimates prompted thoughts about using
dimensional analysis. Fourier (1822), who developed procedures for ?nding wave structures
associated with mathematical functions, introduced dimensional analysis. Dimensional
analysis is a simple and powerful technique for learning about the dependence of a
phenomenon on the dimensions and properties of a system (see, e.g. Langhaar, 1951;
Rayleigh, 1915). Dimensional analysis is based on the simple principle that the units on each
side of any valid physical equation must be the same.
Discussion with colleagues of the units of tourism research resulted in recognizing the value
of applying dimensional analysis ideas in tourism research. Confusion exists in some
documents and articles about theory, data collection and analysis because units of
observation, measurement or analysis are inadequately de?ned or speci?ed. Unit of analysis
has been de?ned as ‘‘the major entity that is being analyzed in a study. It is the ‘what’ or
‘whom’ that is being studied’’ (e.g. see unit of analysis at www.socialresearchmethods.net/
kb/unitanal.htm; see units of measurement and of sampling and/or observation at http://
davidakenny.net/u_o_a.htm). A problem with unit of analysis being ‘‘the major entity being
studied’’ is that what the major entity is need not be clear. When data on people and their
trips are being studied to learn about decision making, is the major entity the person, trip,
person-trip, party-trip or possibly a decision?
PAGE 70
j
INTERNATIONAL JOURNAL OF CULTURE, TOURISM AND HOSPITALITY RESEARCH
j
VOL. 6 NO. 1 2012, pp. 70-80, Q Emerald Group Publishing Limited, ISSN 1750-6182 DOI 10.1108/17506181211206261
Chia-Huh Joy Liang is an
Assistant Professor in the
Department of Tourism and
Leisure Management,
National Penghu University,
Makung, Taiwan. Hung-Bin
Chen is an Assistant
Professor in the Department
of Hospitality Management,
National Penghu University,
Makung, Taiwan.
Ming-Yang Wang 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
)
Consider terminology used for units in tourism research. A person who in providing survey
data from home reports international travel in a certain time frame is an international traveler.
Also, someone interviewed on reentering his or her home country is an international traveler.
However, for a period of time, say 2008, the number of people in a country who can report
traveling internationally with return in 2008 is less than the number of international travelers
returning, unless all international travelers in 2008 make only one trip in 2008. Ambiguity exits.
Furthermore, when data are collected from people at their homes, one can give demographic
distributions for unique people who travel internationally (i.e. people returning fromat least one
international trip in 2008). One may ?nd demographic information (e.g. age-gender-income
distributions) published so marketers understand the attributes of international visitors. Now
consider getting travel data for 2008 by collecting data from visitors returning from abroad or
from foreign visitors who are departing, for example, from Taiwan in 2008. Is giving
demographic distributions by tabulations for these international visitors the same as getting
demographics information from data collected when people are at home? The unit of
observation for people arriving back in a home country or leaving a country is person-visit and
tabulations of demographics for a randomsample froma population of person-visits is not the
same as for the observation unit of people of the population that traveled internationally in
2008. Using person-visit data (e.g. by weighting) to give distributions for unique people that
traveled abroad is not trivial unless, for example, one has data from customs and immigration
with people identi?ed (e.g. by passport number) so a person with multiple international trips is
only counted once. One can see Tyrrell and Johnston (2002) and Johnston and Tyrrell (2003)
for considerations in determining unique visitors. Unique is time frame dependent (e.g. for a
month, year or life time up to a point in time). Beaman and Redekop (1990) refer to considering
unique visitors in a year in allowing one vote or expression of opinion rather than letting
frequent users (say in 2008) have a vote per visit.
The ideas of the last paragraph show the need to be clear about units used in tourism
research. Traveler, visitor, tourist or some combination of these is the only terminology one
?nds in many articles. Person-visit is found in lists of de?nitions used by agencies like Tourism
Manitoba (http://ti.travelmanitoba.com/assets/pdf/tourism_de?nitions_and_concepts.pdf)
and Statistics Canada (e.g. see Statistics Canada, 1997). The term is also found in some
practically oriented material (e.g. see South Grow, 2009). Neither the term person-visit nor
person-visit-day is de?ned in the Dictionary of Travel, Tourism and Hospitality (Medlik, 2003).
Is tourism research science? If so, what does that imply? In the physical sciences if units are
not made clear, research is recognized as ?awed. If you are dealing with acceleration and
you report a result in meters per second or give a number and write that an object speeded
up, people know your result is invalid or the research or its reporting lacks rigor. If
researchers use terms like international traveler, party or visitor in ambiguous ways, is the
research scienti?c? Is using terms like person-visit or party-visit-days necessary? Is the use
of such undesirable because using them is not necessary? Do some researchers see being
precise as distracting from the readability of an article or only being unnecessary for survey
designers and statisticians? Beaman and Redekop (1990) show that international mountain
park visitors to Canada can be described as majority foreign or majority domestic by
changing the unit of measurement. Given their result, presumably, how you measure a
majority of visitors being foreign, impacts planning and management decisions. From
another perspective, while one can commend South GrowRegional Initiative for reporting on
person-visits to southern Alberta as a percent of person-visits to Alberta, if impact of the
visitors relates more closely to party-visit-days and person-visit-days than person-visits,
should reporting be different? While person-visits or party-visits to a country re?ect the work
load at entry/exit points, the case is made subsequently that party-visit days and person-visit
days spent circulating about a country relate to locals’ perceptions about what visitors are
like and to services used and thus services demanded.
Research context and literature
Unit of measurement mattering has been illustrated. However, to clarify the scope of this
article, consider information on problems with a service or reactions to an offering (e.g. see
VOL. 6 NO. 1 2012
j
INTERNATIONAL JOURNAL OF CULTURE, TOURISM AND HOSPITALITY RESEARCH
j
PAGE 71
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
)
Cooper and Schindler, 2003, p. 200). A unit of observation, member of a population, need
not be a person. All complaints ?led in 2006 by visitors to a country is a population (i.e.
population of complaints ?led). If complaints are on a form, a complaint form is the unit of
observation. Tabulations of such complaint information should not be treated as if one has
complaint information obtained by a survey of exiting visitors. Even relating the number of
complaint forms ?led in a month to, for example, total person-visits in the month can be
misleading unless all complaints from a person for a visit are counted as appearing on one
form. When a person can ?le multiple complaint forms, dividing numbers of complaints
forms by the number of person visits can give a false impression about complaining. Say 1
percent of person-visits result in any complaint, those who ?led multiple complaint forms
could result in complaint forms per person-visit being 2 percent. In a survey, one can have a
similar situation. While only 2 percent of visitors register one or more complaints, one can
have 70 complaints from 20 people. Most, say 60, complaints could come from ten people
who complain about a lot of things. Simply stating that complaints per person-visit was 7
percent does not convey the fact that only 1 percent of visitors had more than one complaint.
Raising issues about good reporting is not to suggest that complaints ?led about problems
with visiting a country or with services at any destination should not be investigated.
However, investigation is facilitated if complaints are collected with information that
facilitates recognizing how complaints relate to complainants and potential complainants.
The ideas and issues introduced do not just apply to tourism or leisure researchers. In the
business oriented text of Cooper and Schindler (2003, p. 179) one reads: ‘‘a population
element is the subject on which measurement is being taken.’’ Subsequently, they state, ‘‘the
de?nition of the population may be apparent from the management problem or research
question(s) but often it is not (p. 186).’’ The text gives practical examples and acknowledges
‘‘confusion about whether the population should be persons, households, or families, or a
combination of these’’ should be considered in analysis. Cooper and Schindler are
acknowledging that measurement and analysis can involve multiple units. Does a single
appropriate unit of analysis exist for survey data collected from, for example, international
leisure travels returning home or exiting a given foreign country (e.g. Taiwan)? From material
earlier, the answer is that in general no single unit exists. One may need multiple units to
achieve research objectives.
To amplify on units in observation and analysis, consider real data. In the Canadian Travel
Survey (CTS) households are selected and a person 16 þ in the household is selected so
that the unit of observation for weighted data is a member of the non-institutional population
of residents of Canada 16 þ (Statistics Canada, 1997). Now, this article is not endorsing
continual repetition of information like ‘‘non institutional population of residents of Canada
16 þ ’’ or providing further details on why the sample is not perfect. Rather the concern is
that at some point a clear statement is made. For Taiwan’s survey similar to the CTS, phone
numbers are selected and a person 18 þ associated with the phone number is selected to
report on travel (Taiwan Tourism Bureau, 2008). The population for the survey is reported as
Taiwanese residents 18 þ . With weighting the unit of observation is a Taiwanese resident.
For both Canada and Taiwan, person weights must be calculated to relate a person
appropriately to the population sampled. For Canada, data on all trips taken in a month are,
in principal, recorded (re accurate reporting see Beaman et al., 2001). Therefore, Statistics
Canada (e.g. Statistics Canada, 1997) views weighted trip data as approximating a random
sample of person trips by the CTS population. Given the person-trip sample, the person
weight and a zero-one variable (i.e. one is traveled) can be used to estimate the number of
people in the population or segment who are international travelers in a given month (i.e. time
period for monthly data are one month). Also for Canada, data on the number of international
trips taken by a person can be weighted by a person’s weight and added over a group of
people (e.g. people in a segment) to calculate the number of international person trips taken
by segment members. One might think that in Taiwan one could make estimates like those
made using the CTS. If one could and the estimates agreed with the return records kept by
of?cials who record every return by residents, one could check on survey accuracy or bias.
However, from survey documentation one learns that to prevent data collection from taking
too long, data are only collected on up to three trips. If trips to be reported were selected at
PAGE 72
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
)
random from some preliminary list of all trips, after weighting maybe Taiwan would have a
sample representative of person-trips of the survey population. As the survey is executed,
one has a sample of an ill-de?ned population. One should not make estimates as if trip data
represented person-trips of the population.
Now, consider that research often reports on travels with other people, on travel in a party.
should one report person-trips or party-trips? For parties that travel in one vehicle and
occupy one hotel room, service providers concerned with vehicle transportation and
accommodation presumably want party-trip information. Recognizing that different clients
for statistics ?nd one or the other unit useful, Statistics Canada provides weights for
computing person-trips and party-trips. However, for the CTS parties are only made up of
household members. Friends from different households traveling together for CTS are not
traveling in a party. Knowing what party means in for data or for a theory matters. Units matter
for other than technical reasons. They matter because different units are appropriate in
different contexts.
The picture emerging is that more attention should be paid to units for which theory is
formulated, models are built, analyses are run, etc. Thinking about issues and concepts
provided shows the need for and value of units. A person can be at an international
destination and return home several times in a year. However, just as length of stay can affect
the probability of selecting a person or party (Lucas, 1963; Scheaffer et al., 1996; Sheaffer
and O’Leary, 2005), length of stay can affect the probability of selection of people returning
from abroad. A person who comes back from international trips several times in a year has
more chance of reporting her or his travel than a person who return once. When people who
make a lot of trips are treated as unique people, one gets a distorted picture of travel by
unique people in a country. For trips ending, for example in a month or year, total expenditure
calculated from person-visit data gives a valid estimate of expenditures abroad. However,
average expenditure per person-trip is not the same as average expenditure per person that
travels. One is back to using the right unit for the right purpose.
Problem statement and research strategy
While Beaman and Redekop (1990) focus on response weighting in their discussion of units,
this paper takes a general and non-technical approach to the need for speci?cation of units.
Given the theme of the special issue in which this article appears, examples and discussion
tend to address international tourism. However, hypotheses are general. The generality is
appropriate since material presented is easily altered to apply to other aspects of tourism,
for example, to domestic tourism. The general thesis is that the terminology used in tourism
research should appropriately identify units. As already introduced, evidence exists that
terms for units of observation, measurement and analysis are being used vaguely and
inappropriately. H1 and H2 formalize ideas implicit in the thesis:
H1. For tourism research to be credible science, researchers need to use precise
terminology for units of observation, measurement and analysis in theory
development, in planning of research, in analysis of data and in presentation of
results.
H2. Components of formulations of theory and of data based research are meaningless
when units that statements are about or apply to are not clearly de?ned.
Sampling caveat
A matter that has not been raised is that survey results can be biased because of sampling,
recall or other bias (Burton and Blair, 1991; Krosnick, 1999). A way to ?nd out if a survey
estimate is biased is to compare the estimate with an accurate benchmark (Vaske et al.,
2005). Such measurement of bias has been done for international travel. Statistics Canada
has caused a bias in a person-visit ?ow. The bias, caused by changes in data collection
methods Statistics Canada considered innocuous is estimated at 10 percent (e.g. see
Beaman et al. 2001). To avoid making qualifying about bias in research results a ‘‘no bias
assumption’’ is accepted:
VOL. 6 NO. 1 2012
j
INTERNATIONAL JOURNAL OF CULTURE, TOURISM AND HOSPITALITY RESEARCH
j
PAGE 73
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
)
No bias assumption: all potential respondents to a sample survey respond (no non response)
and, respond accurately; and the sample is, given weighting, representative of its population
within the limits of statistical variability.
Making the no bias assumption may seem inappropriate or unnecessary. The assumption is
made because the authors feel that without it, some of the statements in following sections
should be quali?ed or clari?ed. In fact, making the assumption does not compromise this
research. Making the assumption simply acknowledges that in much tourism research the
no bias assumption is accepted. As implied above, for the CTS Statistics Canada assumes
bias is negligible, though statistically signi?cant violation of the no bias assumption is
established. In other words, accepting the assumption is not an assertion that the
assumption is correct, or a good approximation, where adopted. Rather, just as making a
?awed measurement of a magnitude in the physical sciences does not affect the unit that
should be used, recognizing correct units that apply to tourism theory or statistics is a
separate matter from, for example, unbiased measurement.
Research strategy
The research strategy of this paper is simple. Introductory material has provided some
ideas. The analysis and results section develops those ideas. The Discussion pursues
issues and implications of matters covered in Analysis and results. In concluding, discussion
turns to the need for a best practice for tourism research of appropriately identifying units of
observation, measurement and analysis.
Analysis and results
Based on what has been presented earlier in this paper, issues exist regarding units to use
and being used in tourism research. From a dimensional analysis perspective, a key matter
is to whom or to what statements refer. Consider a statement like a ‘‘sample of visitors was
obtained as the visitors entered ?ight departure areas’’ or ‘‘the table gives demographic and
expenditure information for tourists’’. Referring to travelers, tourists or visitors instead of
person-trips may make some readers comfortable. However, do the statements given make
clear who is in the sample (i.e. the population) or if demographic information is for persons or
person visits? If linguistic variety and readability yields questionable or invalid interpretation
of results or hides ?awed formulation of research, tourism research is not good science.
General considerations
Thinking about units and people’s tourism related activities is facilitated by considering data
collection about travel. A type of data collection introduced and the one for which a subsection
occurs in this section is collection of person and trip data frompeople at home. Another type of
data collection already commented on is for complete destination/trip experiences.
Statements above suggest serious problems in making inferences about unique people at
an origin traveling from complete destination/trip experiences (e.g. see truncation bias in
Smith and Desvouges, 1985; or Phaneuf and Smith, 2006). The problemis unit related. People
who do not travel are not in the person-visit populations data collected. One can make
inferences about people who did not travel but are like those that did. For example, in
modeling one can infer that international travelers from a country who tend to travel
internationally with a given frequency (e.g. about every two years) are like an equal number of
people at home. Units problem exists when research treats person-visit information as person
information. Collection of data on complete destination/trip experiences is not pursued in this
section because hypotheses can be accepted without further consideration of such sampling.
Intercepting people during an experience and observations such as instant counts are even
introduced in this article because conceptual matters regarding units of analysis and data
collection can be intricate (Schreuder et al., 1975).
Sampling to get data for making valid estimates for speci?c populations involves many
issues. Different general sources on sampling are useful. Though dated, Sudman’s Applied
Sampling (Sudman, 1967) is readable and covers a broad range of material. Scheaffer et al.
(1996) cover such matters as systematic sampling (Ch. 7), cluster sampling (Ch. 8-9) and
PAGE 74
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
)
estimating population size (ch. 10). Many other publications are listed with annotation at
www.bettycjung.net/Surveys.htm useful (accessed August 2009).
Sampling, party and being speci?c
Examples illustrate matters of concern. In the statement above about sampling, consider
interviews were with non institutionalized persons 16 þ with data collected on all trips of
certain types (e.g. over 50km from home and not work related) ending in 2008 and by
weighting data yield unbiased estimates for the population, for person-trips and for household
party trips. Given caveats on the population sampled and about observations representing
their population, the article is not suggesting continual repletion of caveats. If the term
person-trip has a consistent meaning in research being reported, quali?cations such as
pleasure person-trip of 16 þ etc. are only needed when, for example, segments are being
compared. However, saying too little in specifying the population to which data apply can be a
problem. If data collectors approach parties (e.g. in departure areas of an airport) and the
party selects who responds to a questionnaire, does a problem arise? Well, if attitude
questions are asked of the respondent, one has responses that are meaningful for person-visit
data but the data are collected for arbitrary members of parties. One does not have a sample
from the population of person-visits and cannot correct to that population by weighting.
Now, consider that a questionnaire being used was developed to collect party data. What
does that mean? In the CTS the party unit is household-party. However, if different clients for
information are interested in different types of parties (e.g. accommodation-party being
people who stay together), reason suggests collecting person-visit data with information that
allows analysis for different kinds of parties. Nevertheless, data collection for person-visits
can involve problems. Prorating party expenditures to a person can be appropriate because
much spending by a household can be for the party. If party expenditures are collected in
person-visit data, adding the expenditures or taking an average mixes units of party-visit
and person-visit. Prorating expenditures based on party size can give correct aggregate
expenditures but amounts (e.g. on gifts) do not necessarily apply to the respondent.
Analysis of amounts using respondent demographics is mixing units of analysis. Maybe,
someone has a use for the average party expenditure associated with a person-visit.
However, an application is not obvious to the authors. In another context, consider that for
the people of a country one international person-trip occurs per 2 residents of the country
(e.g. for residents/people 16 þ ). Writing that 50 percent of people of the country over 15
traveled internationally is not valid unless each traveler made only one international trip. The
problem with calculating 50 percent is that units of measurement are not properly
considered. Person trips divided by persons is average person-trips per person. The
number of persons who made any international trip (e.g. in 2008) divided by population size
is the percent of the population traveling internationally. Units matter.
Origin data collection on international travel
Consider formulating theory about travel behavior of and collecting data froma population in a
particular geographic area (e.g. a country). As for the CTS, the population could be
non-institutionalized Canadians 16 þ . To formulate theory, the units to which the theory
applies should be unambiguously identi?ed. Consider the theory is about international
pleasure trip taking people by people of a country. To test the theory, assume a representative
sample of the population is obtained. Representative means that data collected when
weighted can be used to make estimates for populations (e.g. population of the country,
person-trips, party-trips and person- and party-trip-days). If sampling and weighting are
perfect, estimates are unbiased. Data collection and weighting can involve sophisticated
considerations, as one can see from documentation for the CTS (Statistics Canada, 1997).
When one starts to think about international travel behavior of individuals, the unit being
considered can move from person to person-trip, party-trip or some other unit. Regarding
individuals and their behavior, one may bene?t from articles that examine behavior based on
long interviews with individuals (e.g. see Martin, 2008; DeCrop, 2001; March and Woodside,
2005, chs. 1-3). However, here assume concern is with international pleasure trips that are not
mainly devoted to visiting friends and relation (VFR) and are made in a year, Y (e.g. trips
VOL. 6 NO. 1 2012
j
INTERNATIONAL JOURNAL OF CULTURE, TOURISM AND HOSPITALITY RESEARCH
j
PAGE 75
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
)
ending in Y). Think about structured survey data being collected. Think of data being collected
that applies to a respondent, person data. Think of these data in a table with one row for each
person. Also, for those who travel internationally, think about data for trips being in another
data set. Think of trip data as a table with a rowfor every trip. For a person, typically data would
include age, gender and family membership information. Person data could include attitudes
and preferences information. For trips data can include information about trip planning, travels
mode(s), destination(s), distance, duration, activities, services used; and ratings (e.g. of
destinations, services and the trip). Consider that for each person a person-weight is de?ned
and the person-weights of all people sumto the total people in the population sampled. For all
people reporting an international trip or trips, you can visualize a link between the person’s
data in one table and her or his trip data in the trip-data table. Mathematically, a person is
associated with (e.g. linked by an ID) to each international trip he or she reported.
The last paragraph has presented a relational model of persons taking trips (i.e. taking is a
word that describes the link between tables). The structure described is similar to the
structure used by Statistics Canada (1997) in storing trip data for Canadians. Now, if you put
a person’s weight with each trip (i.e. in the person-trip table), you can call it a person-trip
weight. In analysis of trip data, the weight of the trip indicates how many trips like the given
one the survey indicates were made in a time frame (Y) being considered. Add the
person-trip weights up for all trips and you get total person-trips for Y. If trip-weights are
added for a segment for trips with certain attributes, one obtains an estimate of the number
of trips with the given attributes made by the segment members in Y. For CTS party size is
used to create a party-trip weight. As for person-trips, estimates can be made for segments
or the population.
Several matters already raised should be noted. Adding person-weights of people who
traveled internationally for pleasure in Y (adding weights for unique people who traveled) is not
the same as adding trip-weights for trips (i.e. using data in the person-trip table) given some
people took one trip while others took two or more trips. Therefore, if you have a theory and/or a
model about people taking international trips, do you test your theory usingperson data (e.g. 1
took an international trip and 0 did not) to test your theory? How do trip attributes of particular
trips and number of trips enter into theory and/or model? Given your theory is about a person
taking trips, can you justify using attributes of a person and associating these with information
about person-trips? With such modeling, you are presumably getting some aggregate idea
about attributes of people in?uencing trip taking. Many years ago Cicchetti et al. (1969; see
also CORDS, 1976, ch. 1) proposed that for the kind of modeling being described, two models
are needed. One is for the prediction of participation-nonparticipation. The other model is
estimated for those that traveled with number of trips as the dependent variable. However, if
number of trips is the dependent variable, where do attributes of trips enter into modeling?
Party travel
Commentary so far has focused on the person. Consider you are interested in party travel. In
generating information about party travel using person-trip data, person-trip weights must
be adjusted. Because of how Statistics Canada de?nes party (household-party) and
because numbers in the household going on trips is recorded, party-trip weights can be
calculated fromperson-trip weights (e.g. see Statistics Canada, 1997). However, if you are in
Taiwan, you know that a lot of international party travel is by people who do not live together
(are not in the same household). What good is the unit party if it is de?ned in a way that is not
appropriate for many trips. Party has a meaning to a particular tourist for a trip. Furthermore,
party can have different meaning for tour operators, accommodation providers and tour
guides. Outside the con?nes of a strict de?nition, for example by an agency or in a research
project or component of a project, party is ambiguous. Nevertheless, one can recognize
party members of different types of parties playing a critical role in decision making about
international travel by having such quali?ed terms such as household-party and
friends-party. Using such terms can clarify literature on people in?uencing decisions (for
such literature see part 4 of Crouch et al., 2004). Anyway, one of the tourism research terms
that is ambiguous is party. Can party be a unit of analysis when it means something different
for different trips?
PAGE 76
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
)
Use of terms and their meaning
The ideas already provided in this section are adequate to showanalysis of data about travel
obtained from people at their usual place of residence can involve different units
(e.g. person, person-trip and party trip). However, units such as person-trip are not found in
abundance in the travel research literature. Articles about trip decision making are common
and typically focus on a person and a single trip (e.g. see Crouch et al., 2004, part 4; or
Mazanec et al., 2001, part 3). A search of the tourism literature did not produce any article
devoted to explaining the diversity of people’s person-trips taken in a time period. The subtle
matter that is arising with the use of words is ambiguity. For some clients for information such
as airlines, person-trips may be desirable because person-trips relates to passenger
volume. For hotels party-visit data may seem desirable because occupancy relates to
party-trips. However, do Asian groups that call themselves a party occupy one room? In
Taiwan some accommodation is set up so Japanese groups of 10 þ people sleep in a room.
In closing this consideration of units in relation to origin surveys, estimating percent traveling
is important since destination surveys only yield information about people who do not visit by
statistical inference (see, e.g. Phaneuf and Smith, 2006). One reason for origin surveys
measuring travel by members of a population or a segment. As introduced above, even with
data letting one know the number of people returning from international travel (e.g. from
Taiwan’s of?cers recording all returns), the ratio of international person-trips in a year (Y) to
persons in Taiwan is not the percent of Taiwanese traveling internationally in Y. Also, with
origin data on travel one may ?nd an average of 33 percent of persons (e.g. 16 þ ) travel
internationally in a given year. International person-trips per person in the country could be
0.66. The average number of person-trips per person is 0.66. As 66 percent this is
person-trips as a percent of the population. However, 0.33 is the proportion of the population
traveling internationally. Expressing units correctly matters. By month one may ?nd in
February only 5 percent of a population travels internationally while in July the percent is 15
percent. Since the same 0.05 percent that traveled in February may travel in other months,
monthly percent-traveling ?gures only set a lower limit for the percent traveling annually.
Time frame being a month or being a year matters in specifying a rate. Furthermore, while
person-trips like dollars can be totaled over time, adding numbers traveling internationally
for months does not yield total numbers who travel internationally. Valid operations depend
on units. Just as using the correct numerator in making a computation matters, one needs to
pays attention to the units in determining is a computation is appropriate.
Hypothesis acceptance/rejection and discussion
Repeatedly commentary in this paper has shown that units matter. Data collected frompeople
can be used to produce such estimates as the number traveling, the number of people in a
population traveling internationally, their person-trips made, their number of party-trips, their
number of party- or person-trip-days, etc. What a number speci?es being clear is important.
Because the matter the need for careful use of units is addressed by examples, reference to
papers with ambiguous use of terms and or other problems with units are not given. Therefore,
accepting or rejecting H1 comes down to whether the paper has built a solid case for broader
use of units such as person-visit and person-visit-days and more careful use of terms such as
visitor or traveler. The authors believe the case is made and accept H1.
H2 expresses why having and using an appropriate units (i.e. of observation, measurement
and analysis) is necessary. Stating ‘‘components of formulations of theory and parts of data
based research are meaningless when units that statements apply to are not clearly
de?ned’’ may seem to be exaggeration. However, making a strong statement is reasonable.
The Beaman and Redekop (1990) example that a statement about a majority of visitors being
foreign being true or false depending on the unit, the way of measuring, shows units must be
speci?ed. Nevertheless, logic also provides proof. Consider a planning application that cites
person-visit statistics when person-trips or person-visit days relate to the demand for
service. Change in person-visit statistics may give an idea of what change in demand is.
However, if person-visits are going up and length of stay is going down, person-visit trend
data can suggest growing demand when demand is declining. In other words, logic as well
VOL. 6 NO. 1 2012
j
INTERNATIONAL JOURNAL OF CULTURE, TOURISM AND HOSPITALITY RESEARCH
j
PAGE 77
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
)
as observation supports recognizing that person-visit and person-visit-day can have
differing trends (e.g. one growth and the other decline). Intuitively, person visits related to
what you see at entry and exit from a country and (2) person-visit-days relate to who you see
while party-visit-days relate to encounter with parties being served (e.g. in accommodation).
An example facilitates understanding issues. Assume people coming to a resort are
dominantly from the UK and from Spain. Assume UK visitors on average have a party size of
three and stay a week but half as many UK party-visits come as for Spain. Assume Spanish
visitors have an average party size of three and spend about the same per day as UK
visitors. Are the Spanish by far the most important group supporting the resort? If Spanish
visitors stay an average of two nights, visitors from UK rather Spanish visitors are the
day-to-day majority. The UK visitors should tend to be encountered more than twice as often
in daily business with visitors. Being the majority, in the sense presented, has clear
implication for planning and management. People responsible for marketing should
certainly not be basing plans strictly on volume of entry/exit (i.e. on person-visits). Starting to
get a valid picture of the role of visitors in the economic and social activity of an area is a step
in good research. For practical work, policy, planning or management justi?cation of action
should be rational. If numbers are used without knowing their units, logic is ?awed.
Though meaningless is a strong word to use, research is not scienti?c unless logic justi?es
what is done. Accepting H2 just acknowledges that components of tourism research have
no scienti?c meaning when what statements are about is not clear. Therefore, H2 is
accepted recognizing that meaningless applies to those aspects of any tourism research in
which statements about who or what are not clear, units are not clearly de?ned. In other
words, practical research and theory can be valid in some aspects and ?awed in others.
Need for best practices
Given conclusions reached, speci?cation of units is important to tourism research being
correctly interpreted or even valid. Practical or applied research involves observation.
Cooper and Schindler (2003, p. 659) maintain that a technical report should contain
‘‘suf?cient procedural information to allow others to replicate the study.’’ Kerlinger and Lee
(2000, p. 570) in commenting on the experimental versus the nonexperimental methods
state: ‘‘replication is always desirable, even necessary.’’ The Public Opinion Quarterly (POQ)
in its Notice to Contributors makes clear the importance that is placed on documentation that
allows for survey replication by making ‘‘at least approximate replication’’ a requirement
(www.oxfordjournals.org/jnls/list/poq/instauth/). However, replication can not occur without
correct interpretation of data. Using correct units of analysis in practical or applied research
is about describing things accurately enough so concepts, estimates, etc. are used and
interpreted correctly. Tourism research has no practical value unless application to
observing and describing things exists. If, for example, a theory about trip decision making
is unclear about units of data to use to con?rm the theory, clari?cation of the theory is
needed. Given person-visits can increase while person-visit-days decrease, theory,
propositions based on theory and analysis should be speci?c about units. As discussed,
treating a theory on deciding about taking a particular trip as a theory for predicting an
individual’s person-visits (e.g. in a year) is not be valid when causal factors relate to type of
trip. If one is formulating an aggregate model, the model should be described as such.
Bene?ts can arise from examining various trip decision theories (e.g. see chapters 1 to 3 of
March and Woodside, 2005). Worthwhile research would be examining models and related
theories to determine which are truly for individuals and which are aggregate (over people or
trips). For aggregate a concern would be identifying models/theories described, for
example, as if they apply to individuals. Furthermore, do some theories just apply to an
individual’s particular trip while some are more general? Classi?cation would facilitate
improving research.
Given tourism research based on vague units yields questionable theory and applied
research, having a best practice is desirable. Based on what has been presented, a tourism
research best practice could be:
PAGE 78
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
)
Best practice: units used in formulating theory or data collection and units used in analysis and
presenting results should be precisely de?ned.
Appropriate use of units depends on professional researchers and academics being
leaders in correct use of units. Promoting use of clearly de?ned units can occur in different
ways. Material being taught in universities and in professional development courses having
correct units is necessary in building a community that understands the need for and use of
units such person-visit and person-visit day. Fostering good practice depends on
constructive critique of the use of vague or invalid terms or faulty description of populations
and data collection in presentations at meeting and in articles/chapters of professional
books and journals. Also, encouraging rigorous use of units in practical (e.g. agency and
consultant reports) reports is necessary to get people to abandon dispositions to use terms
inappropriately or vaguely. The best practice suggested can have immediate
consequences if academic tourism research journals take a lead in causing change by
requiring appropriate use of units. An important step will be taken if journals such as the one
containing this article enforce rules preventing ambiguity whether in use of it or this or in use
of unit related terms vaguely or inappropriately.
Conclusion
This research introduces the importance of better use of units in applied and academic
tourismresearch. When units that apply to a statement, whether in an equation or formulation
of theory are not consistent, a problem exists. Terms like trip, visits, travelers or visitors can
be quite adequate for informal conversation. However, clear speci?cation of units is needed
when not doing this can possibly result in confusion. Yes, in some cases, context can make
clear what is meant by trips, tourists or visitors. However, as illustrated above, in some cases
accurate de?nition of populations, units of observation, measurement and analysis is critical.
Good science should involve precision rather than linguistic variety and vagueness. That
some readers are uncomfortable with terms that accurately state units of measurement and
observation shows a problem with what the reader has learned not a problem of literary
ability of the writer. A best practice, such as introduced above, is necessary so better
tourism research will occur.
References
Beaman, J.G. and Redekop, D. (1990), ‘‘Some special consideration in weighting survey data’’,
Les Cahiers du Tourisme, Centre des Hautes Etudes Touristiques, Aix-en-Provence, (Aix-En-Province),
France Se´ rie C # 118.
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’’, Consumer Psychology of Tourism,
Hospitality and Leisure, Vol. 2, pp. 49-65.
Burton, S. and Blair, E. (1991), ‘‘Task conditions, response formulation processes and response
accuracy for behavioral frequency questions in surveys’’, Public Opinion Quarterly, Vol. 55 No. 11,
pp. 50-79.
Cooper, D.R. and Schindler, P.S. (2003), Business Research Methods, 8th ed., McGraw-Hill, NewYork, NY.
CORDS (1976), ‘‘Canadian outdoor recreation demand study: volume 2 technical notes (provisional
copy)’’, available at: www.cals.uwaterloo.ca/pubs/ (accessed December 2009).
Crouch, G.I., Perdue, J.R.R., Timmermans, H. and Uysal, M. (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, New York, NY, pp. 333-47.
Fourier, J.B.J. (18221955), The´ orie Analytique de la Chaleur, ‘‘The Analytical Theory of Heat’’, Dover
Publications, NewYork, NY, replication of the English translation that ?rst appeared in 1878 with previous
corrigenda incorporated into the text, by Alexander Freeman.
VOL. 6 NO. 1 2012
j
INTERNATIONAL JOURNAL OF CULTURE, TOURISM AND HOSPITALITY RESEARCH
j
PAGE 79
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
)
Johnston, R.J. and Tyrrell, T.J. (2003), ‘‘Estimating recreational user counts American’’, Journal of
Agricultural Economics, Vol. 85 No. 3, pp. 554-68.
Kerlinger, F.N. and Lee, H.B. (2000), Foundations of Behavioral Research, 4th ed., Harcount, Singapore.
Krosnick, J.A. (1999), ‘‘Survey research’’, Annual Review of Psychology, Vol. 50, pp. 537-67.
Langhaar, H.L. (1951), Dimensional Analysis and Theory of Models, John Wiley, New York, NY.
Lucas, R.C. (1963), ‘‘Bias in estimating recreationists’ length of stay form sample interviews’’, Journal of
Forestry, Vol. 61 No. 2, pp. 912-3.
March, R. and Woodside, G.A. (2005), Tourism Behavior: Travelers’ Decisions and Actions, CABI,
Wallingford.
Martin, D. (2008), ‘‘Grounded theory of international tourism: building systematic propositions fromemic
interpretation of Japanese travellers visiting the USA’’, in Woodside, G.A. and Martin, D. (Eds), Tourism
Management: Analysis, Behavior and Strategy, CABI Publishing, Wallingford.
Mazanec, J.A., Crouch, G.I., Brent Richie, J.R. and Woodside, A.G. (2001), Consumer Psychology of
Tourism, Hospitality and Leisure, Vol. 2, CABI Publishing, Wallingford.
Medlik, S. (2003), Dictionary of Travel, Tourism and Hospitality, Elsevier, Amsterdam.
Phaneuf, D.J. and Smith, V.K. (2006), ‘‘Recreation demand models’’, in Ma¨ ler, K.G. and Vincent, J.R.
(Eds), Handbook of Environmental Economics, Elsevier, Amsterdam, p. 2.
Rayleigh, L. (1915), ‘‘The principle of similitude’’, Nature, Vol. 95, pp. 66-8.
Schreuder, H.T., Tyer, G.L. and James, G.A. (1975), ‘‘Instant- and interval-count sampling: two new
techniques for estimating recreation use’’, Forest Science, Vol. 21 No. 1, pp. 40-4.
Scheaffer, R.L., Mendenhall, W. and Ott, R.L. (1996), Elementary Survey Sampling, 5th ed., Wadsworth
Publishing Co., Belmont, CA.
Sheaffer, A. and O’Leary, T. (2005), ‘‘Noncommercial ?sh consumption and anglers at risk’’, Human
Dimensions of Wildlife, Vol. 10 No. 4, pp. 229-38.
Smith, V.K. and Desvouges, W.H. (1985), ‘‘The generalised travel cost method and water quality
bene?ts: a reconsideration’’, Southern Economic Journal, Vol. 52, pp. 371-81.
South Grow (2009), ‘‘South Grow Regional Initiative’’, available at: www.southgrow.com/_data/
datadocs/southgrowtourismindustry4.pdf (accessed July 2009).
Statistics Canada (1997), Canadian Travel Survey Microdata User’s Guide, SC document identi?cation
number 87M0006GPE, Statistics Canada, Ottawa.
Sudman, S. (1967), Applied Sampling, Academic Press, New York, NY.
Taiwan Tourism Bureau (2008), 2007 Annual Survey Report on Visitors Expenditure and Trend in Taiwan,
Taiwan Tourism Bureau, Taipei.
Tyrrell, T.J. and Johnston, R.J. (2002), ‘‘Estimating regional visitor numbers’’, Tourism Analysis, Vol. 7
No. 1, pp. 33-41.
Vaske, J., Beaman, J.G. and Miller, C. (2005), ‘‘Cognitive processes in hunters’ recall of participation and
harvest estimates’’, Journal of Wildlife Management, Vol. 69 No. 3, pp. 967-75.
Corresponding author
Ming-Yang Wang can be contacted at: [email protected]
PAGE 80
j
INTERNATIONAL JOURNAL OF CULTURE, TOURISM AND HOSPITALITY RESEARCH
j
VOL. 6 NO. 1 2012
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. Stefano De Cantis, Anna Maria Parroco, Mauro Ferrante, Franco Vaccina. 2015. Unobserved tourism. Annals of Tourism Research
50, 1-18. [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_864823900.pdf