Annals of Tourism Research, Vol. 30, No. 4, pp. 906­926, 2003
 2003 Elsevier Ltd. All rights reserved
Printed in Great Britain
0160-7383/$30.00
www.elsevier.com/locate/atoures
doi:10.1016/S0160-7383(03)00059-8
INTERNATIONAL LEISURE
TOURISTS' INVOLVEMENT PROFILE
Dogan Gursoy
Washington State University, USA
Erdogan Gavcar
Mugla University, Turkey
Abstract: The involvement construct has received a great deal of attention in recent years.
However, the majority of research on this topic has focused on internal tourists, recreational
activities participants, or associated services and products. Using survey data collected from
European leisure tourists at an international destination, this study examined the underlying
dimensions of their involvement. Scale unidimensionality is first revealed by an exploratory
factor analysis and validated by a confirmatory factor analysis. The data suggested that this
is a three dimensional construct: pleasure/interest, risk probability, and risk importance.
Construct validity, including discriminant, convergent and nomological validity, and
reliability are also satisfactorily established. Keywords: international tourists, involvement,
confirmatory factor analysis, scale development.  2003 Elsevier Ltd. All rights reserved.
Résumé: Profil de la participation des touristes internationaux de loisirs. La construction
théorique de la participation a eu beaucoup d'attention dans ces dernie`res années. Le plus
grand nombre de recherches a` ce sujet ont cependant surtout porté sur les touristes domes-
tiques, la participation aux activités de loisirs ou les services et les produits associés. En
utilisant des données d'enque^te recueillies de voyageurs européens de loisirs a` une desti-
nation internationale, cette étude a examiné les dimensions sous-jacentes de leur partici-
pation. L'unidimensionalité d'échelle est révélée d'abord par une analyse factorielle explora-
toire et puis validée par une analyse factorielle confirmatoire. Les données sugge`rent que
cette participation est une construction a` trois dimensions : plaisir ou intére^t, probabilité de
risque et importance de risque. On établit de manie`re satisfaisante la fiabilité et la validité
de la construction, y compris la validité discriminante, convergente et nomologique. Mots-
clés: touristes internationaux, participation, analyse factorielle confirmatoire, développement
d'échelle.  2003 Elsevier Ltd. All rights reserved.
INTRODUCTION
Involvement is a widely used concept in the consumer behavior
literature and within the last decade has found increasing followers
among leisure, recreation, and tourism researchers (Havitz and Diman-
che 1999). The subject is regarded as one of the most important deter-
Dogan Gursoy is Assistant Professor in the School of Hospitality Business Management at
Washington State University (Pullman WA 99164-4742, USA. Email <dgursoy@cbe.wsu.edu>).
His research areas include tourists' information search behavior, community support for
tourism development, cross-cultural studies, and involvement. Erdogan Gavcar is Professor
at Mugla University. His research interest includes pricing, destination attractiveness, motiv-
ations, and involvement.
906
907GURSOY AND GAVCAR
minants of consumer behavior (Broderic and Mueller 1999) and a cen-
tral point in explaining decision-making processes (Dimanche, Havitz
and Howard 1993). It is considered to be the degree to which con-
sumers engage in different aspects of the consumption process: pro-
duct, advertising, information search, information processing,
decision-making, and the act of purchasing (Broderic and Mueller
1999).
While many theoretical models for the conceptualization of involve-
ment have evolved, researchers criticized the lack of agreement on
what the term actually means (Havitz and Dimanche 1999; Rothschild
1984). Concerns with measurement have resulted in the development
of several scales intended to operationalize the construct. Two of these
have captured most researchers' interest. They are Zaichkowsky's
(1985) "personal involvement inventory" (PII) and Laurent and Kap-
ferer's (1985) "consumer involvement profile" (CIP). Both scales have
been used and tested extensively (Havitz and Dimanche 1997). The
former's PII has been supported in its unidimensional structure by
many and evidence has been provided for a two-factor structure
(Broderic and Mueller 1999; Mittal 1989). However, the latter's CIP
has received much more attention due to its multidimensional struc-
ture (Broderic and Mueller 1999). This construct has been used and
tested extensively (Jamrozy, Backman and Backman 1996). Most of the
tests were conducted on tangible products such as beer (Mittal and
Lee 1988), cameras, jeans (Mittal 1995), television sets, shampoo, mat-
tresses (Laurent and Kapferer 1985), cars, clothing, consumer elec-
tronic equipment, audio recording (Rodgers and Schneider 1993),
food items (Broderic and Mueller 1999), or shopping and makeup
(Bloch 1993). Several researchers examined involvement in the con-
text of recreation and leisure (Dimanche et al 1991; Jain and Sriniva-
san 1990), and in the context of tourism (Jamrozy et al 1996; Kim,
Scott and Crompton 1997).
Even though involvement and its importance in tourists' decision-
making processes are well articulated, the same related to international
destinations has received little attention. Examinations of the con-
struct, in tourism and recreation, were mostly conducted on North
American tourists, recreational activities participants, or associated ser-
vices and products (Havitz and Dimanche 1997). Previous research
indicates that involvement scales at the activity, product, and desti-
nation levels are likely to result in a different number of dimensions.
For example, the original version of the CIP was developed at the pro-
duct level and resulted in five dimensions. However, when it was
applied to leisure setting and activities, it most consistently yielded four
dimensions. It is not clear how many dimensions the CIP scale may
yield if it is applied to tourists at international destinations.
The purpose of this study is to examine the applicability of the CIP
scale on international leisure tourists to better understand the concept
of their involvement. Using a three-step process, this paper develops
and validates this involvement scale based on Laurent and Kapferer's
CIP scale. This analysis includes a conceptual discussion of involvement
and of their CIP scale. It utilizes empirical data gathered from leisure
908 TOURISTS' INVOLVEMENT
tourists who were vacationing in an outbound destination to test the
applicability of the scale.
TOURISTS' INVOLVEMENT
Involvement has been examined from three main perspectives: pro-
duct-centered, subject-centered, and response-centered orientations
(Finn 1983). Although these orientations are important in understand-
ing consumer behavior, a subject-centered perspective is used in this
study to provide a framework for understanding the influence of the
individual tourist's involvement on his/her decision-making and inter-
national leisure destination choice behavior (Celsi and Olson 1988).
Most hospitality and tourism products are purchased, consumed,
and evaluated in the form of services such as vacation trips. Their pro-
duction, consumption, and evaluation differ from those of tangible
goods in four fundamental ways: being mostly intangible, hetero-
geneous, inseparable, and perishable (Zeithaml, Parasuraman and
Berry 1990). Therefore, the purchase process of services and con-
sumers' involvement is likely to differ from that of durable goods. For
example, consumers, in reality, purchase and consume most services
at locations other than where they live (Sirakaya, Mclellan and Uysal
1996). Their decision-making process used to purchase tourism ser-
vices takes much longer than for many other products such as tele-
vision sets (Gursoy 2001). In addition, most of the time, they do not
get any tangible return for their investment, except maybe souvenirs
and a receipt. They also deal with a high-level of perceived risk because
of high personal investment of time, effort, and money (Teare 1992).
Consumers are likely to plan and save money to purchase tourism ser-
vices over a longer time period than many other product purchases
(Mountinho 1987) due to the high level of perceived risk involved.
Consequently, they are likely to be more involved in the decision-mak-
ing, selection, and purchase processes of tourism product.
The initial research in involvement in consumer behavior can be
traced back to the early work of Sherif and Cantril (1947) and the
social judgment theory developed by Brunswik (1956) and Sherif and
Hovland (1961). Since the early work of Sherif and Cantril (1947) and
the development of the social judgment theory, involvement has been
subjected to extensive definitional, conceptual, theoretical, and empiri-
cal examination (Rodgers and Schneider 1993). In general, it has been
identified with interest, excitement, and enthusiasm for product class,
activities, or information. It has also been related to personal values,
ego-involvement, and importance and risk perceptions (Jamrozy et al
1996). Researchers have also identified "affective" and "cognitive" types
of involvement (Park and Young 1986). Further, it has been observed
that consumers assign both emotional and rational values to products
(Zaichkowsky 1987), measuring the "hedonic" and "utilitarian" aspects
of consumption (Babin, Darden and Griffin 1994).
However, the general view of involvement has been one of "personal
relevance" (Zaichkowsky 1985). That is, consumers' level of involve-
ment with an object, situation, or actions is determined by the degree
909GURSOY AND GAVCAR
to which they perceive that concept to be personally relevant. This is
represented by the perceived linkage between their needs, goals, and
values, and their product knowledge. To the extent that product
characteristics are associated with personal goals and values, consumers
will experience strong feelings of personal relevance or involvement
with the product (Celsi and Olson 1988).
Throughout the years many theoretical models for the conceptualiz-
ation of involvement have evolved. They resulted in development of
several different definitions (Havitz and Dimanche 1999; Rothschild
1984) and scales that are intended to operationalize the construct.
However, as already noted, only two of them captured the interest of
most researchers (Jamrozy et al 1996). Both Zaichkowsky's PII and
Laurent and Kapferer's CIP have been used and tested extensively
(Havitz and Dimanche 1997, 1999).
The former scale assumes that involvement is concerned primarily
with the personal relevance to the consumer of a particular product,
advertisement, or a situation (Rodgers and Schneider 1993). The PII
construct has been supported in its unidimensional structure by many,
but evidence has also been provided for a two factor structure
(Broderic and Mueller 1999; Mittal 1989). Several researchers utilized
PII in tourism and leisure studies and concluded that it is a useful
tool in examining tourists' vacation behavior and their participation
in leisure activities (Backman and Crompton 1989, 1991; Havitz and
Crompton 1990).
However, CIP has received much more attention due to its multi-
dimensional structure (Broderic and Mueller 1999; Havitz, Dimanche
and Howard 1993). Laurent and Kapferer (1985) suggested five dimen-
sions of involvement: the perceived importance of the product; the
pleasure value of the product; the symbolic or sign value attributed by
the consumer to the product, its purchase, or its consumption; the
perceived importance of negative consequences in case of a poor
choice; and the perceived importance of making such a choice. Several
researchers have tested different involvement scales in recreation, leis-
ure, and tourism contexts (Havitz and Dimanche 1990; Havitz et al
1993) and determined that involvement has a multidimensional struc-
ture as suggested by the original CIP scale. Dimanche et al (1991)
translated this scale from French into English and tested it in the con-
text of recreational and touristic activities.
However, their findings contrasted with Laurent and Kapferer's that
CIP scale has five dimensions. Dimanche et al (1991) found that
involvement consisted of four dimensions in the recreational and tour-
istic context: the interest in, or perceived importance of the
product/service, as well as the perceived pleasure value attributed to
the product (importance/pleasure); the perceived sign value attri-
buted by the consumer to the product (sign); the perceived impor-
tance of negative consequences in a case of a poor choice (risk
consequences); and the perceived probability of making such a mistake
(risk probability). After the study by Dimanche et al, several
researchers in the field of leisure and tourism utilized the multidimen-
sional CIP scale. For example, Park (1996) investigated the relation-
910 TOURISTS' INVOLVEMENT
ships among fitness participants' attitudinal loyalty profiles and involve-
ment profiles utilizing CIP scale. Results indicated a strong relationship
between attitudinal loyalty and involvement. Park concluded that utiliz-
ation of involvement profiles is useful in determining consumers'
short-term usage patterns. However, he suggested that the attitudinal
loyalty concept is more useful for attracting long-term membership to
fitness programs, arguing that high involvement is one of the precon-
ditions for consumer loyalty, but not the only one.
Madrigal et al (1992) examined married couples' involvement with
family vacations. They identified similar underlying dimensions of their
involvement in family vacations to those reported by Dimanche et al
(1991). However, they were able to identify only two strong factors:
pleasure/importance and sign dimensions. Their study reported posi-
tive relationships between pleasure/importance dimension and gender
role and education and a negative relationship between
pleasure/importance and parental status. The relationship between
sign dimension and parental status was negative, but that between age
and sign dimension was positive.
Dimensions of Involvement
Although opinions about the dimensions of involvement are mixed,
most researchers agree the concept has a multidimensional structure
(Havitz, Dimanche and Howard 1993), while others argue that per-
ceived interest/importance of the product or activity alone represent
involvement (Mittal 1989, 1995). It is commonly agreed that interest/
importance is an important dimension; thus, every current concep-
tualization of involvement includes this dimension (Havitz and Diman-
che 1999).
Opinions about the number of dimensions of involvement are also
mixed. While Laurent and Kapferer (1985), developers of the CIP
scale, argue that it has five dimensions--interest/importance; hedonic
(pleasure); sign; risk probability; and risk importance--leisure and
tourism researchers argue that involvement consists of four:
importance/pleasure; sign; risk consequences; risk probability
(Dimanche et al 1991; Madrigal et al 1992). The latter group consist-
ently suggest that for recreation and touristic activities, interest/
importance, and hedonic (pleasure) dimensions of the CIP scale are
present in a single scale (Dimanche et al 1991, 1993; Jamrozy et al
1996). Laurent and Kapferer argue that the interest/importance and
hedonic (pleasure) dimensions are conceptually different.
Jamrozy et al (1996) examined the relationship between involve-
ment and opinion leadership in tourism. They utilized both PII and
CIP scales. Their findings indicated that both scales were strong predic-
tors of opinion leadership in tourism. They also revealed that only four
of the five dimensions of involvement, as measured by Laurent and
Kapferer's CIP scale, were present. However, only importance/
pleasure and risk probability dimensions were significant. They sug-
gested that the importance/pleasure dimension of either the CIP scale
or the PII scale might be sufficient to predict opinion leadership.
911GURSOY AND GAVCAR
Several other studies support the idea that consumer involvement is
unidimensional and can be assessed by measuring only the
interest/importance dimension of the CIP scale (Mittall 1989, 1995).
Others argue that involvement is a multidimensional construct and
cannot be measured by only interest/importance dimension (Havitz
and Dimanche 1990, 1997). Most of the recent studies in tourism and
leisure have utilized involvement as a multidimensional scale. Havitz
and Dimanche (1997) reported that six of the nine leisure involvement
data sets published through 1990 used unidimensional scales. How-
ever, excluding comparison research, only nine of thirty-three studies
published between 1991 and 1997 used unidimensional scales (Havitz
and Dimanche 1997).
Sample and Data Collection
The study aimed to address theoretical and measurement issues
raised by the contradictory findings of previous studies, by assessing
Laurent and Kapferer's CIP scale on international leisure tourists.
Their scale is assessed utilizing the data collected from international
leisure tourists to Turkey. Respondents were asked to answer questions
related to CIP scale and their knowledge of Turkey.
Tourists who were staying at selected hotels were interviewed using
a structured survey during the summer of 2001. Respondents were
screened regarding the purpose of their trip before being personally
interviewed. Only international leisure tourists were surveyed for this
study. All interviews were conducted in English, which was not the first
language of some of the respondents. As suggested by Dimanche
(1994), assumption of English fluency is a major limitation of current
tourism research. In order to overcome this limitation, all interviews
were conducted face-to-face with care that each question was thor-
oughly understood by each respondent.
A total of 460 tourists were interviewed. Most of them were English
(57.6%) followed by Irish (14.7%), German (14.2%), Dutch (7.7%),
and Scottish (2.3%). Only 3.6% were from other countries. Average
age of respondents was 34.2 and the average annual income was $
37,622. Most were female (62.2%). The largest portion was single
(45.4%), the remainder married (41.3%), divorced (7.7%), separated
(3.6%), and widowed (1.9%).
Measurement of Variables
The structured survey questionnaire used to collect data contained
a modified version of the Laurent and Kapferer's CIP scale to fit the
purpose of this study. Table 1 presents the questions used to measure
international leisure tourists' involvement. As shown in this table, the
scale consisted of 16 items, all measured on a 5-point Likert type scale
with "totally agree" at the low end and "totally disagree" at the high
end.
Tourists' knowledge was measured by a three-item scale--adopted
from Park, Mothersbaugh and Feick (1994)--measuring consumers'
912 TOURISTS' INVOLVEMENT
Table 1. Results of the EFA (N = 230)
Items Factor Factor Factor Factor Factor
1 2 3 4 5
When I chose a vacation destination, it is .434 .404
not a big deal if I make a mistake.a
It is really annoying to purchase a vacation .815
that is not suitable.
If, after I bought a vacation, my choice .710
proves to be poor, I would be really upset.
Whenever one buys a vacation, one never .494
really knows whether it is the one that
should have been bought.
When I face a variety of vacation choices, I .599
always feel a bit at loss to make my choice.
Choosing a vacation destination is rather .693
complicated.
When one purchases a vacation, one is .594
never certain of one's choice.
You can tell a lot about a person by the .465 .475 -.499
vacation destination he or she chooses.
The vacation I buy gives a glimpse of the .441 .435 -.616
type of man/woman I am.
The vacation you buy tells a little bit about .491 -.587
you.
It gives me pleasure to purchase to .669
purchase a vacation.
Buying a vacation is like buying a gift for .686
myself.
A vacation is somewhat of a pleasure to .746
me.
I attach great importance to a vacation. .722
One can say vacation destinations interests .673
me a lot.
A vacation destination is a topic that leaves .414 .526
me totally indifferent.a
Eigenvalue 3.454 2.253 1.68 1.464 1.24
% of Total Variance 21.587 14.082 10.499 9.147 7.002
Cronbach's Alpha 81.15 68.21 59.22
The Kaiser-Meyer-Olkin measure of sampling adequacy .724
The Bartlett's test of sphericity (significance level) .000
Note: All items are measured on a five-point Likert scale with the word "totally diagree"
at the low end and the word "totally agree" at the high end.
a
Reverse coded.
self-assessed knowledge. Park et al (1994) reported that the stan-
dardized Cronbach's Alpha for these items was 0.91. A seven-point
Likert type totally agree/totally disagree scale is used for these items.
Three statements that are proposed to measure tourists' familiarity are:
"compared to average person, I am very familiar with a wide variety of
vacation destinations"; "compared to my friends, I am very familiar with
913GURSOY AND GAVCAR
a wide variety of vacation destinations"; and "compared to people who
travel a lot, I am very familiar with a wide variety of vacation desti-
nations".
Study Methods
A three-step procedure was used in this study for assessment of CIP
scale. First, underlying constructs measuring it were identified by using
an exploratory factor analysis (EFA). Second, CIP scale attributes were
examined by using Cronbach reliability. Third, underlying constructs
measuring its attributes were validated by using a confirmatory factor
analysis. For the purpose of this study and confirmation of results, the
sample was randomly divided into two subsamples (N = 230 for each).
Both were examined for outliers and no outliers were found.
To detect scale dimensionality, an exploratory factor analysis (EFA)
with a principal component method was conducted for CIP using the
first subsample. The appropriateness of factor analysis was determined
by examining the Kaiser­Meyer­Olkin measure of sampling adequacy
and the Bartlett's test of sphericity. A value of .60 or above from the
Kaiser­Meyer­Olkin measure of sampling adequacy test indicates that
the data is adequate for exploratory factor analysis (Tabachnick and
Fidel 1989). A significant Bartlett's test of sphericity is also required.
In order to make sure that each factor identified by EFA has only one
dimension and each attribute loads only on one factor, attributes that
had factor loadings of lower than .40 and attributes loading on more
than one factor with a loading score of equal to or greater than .40
on each factor were eliminated from the analysis (Hattie 1985).
After identifying the dimensions, a Cronbach's Alpha reliability test
was conducted for each measurement scale using the first subsample.
Any item with a below .50 item-to-total correlation was eliminated
(Chen and Hsu 2001; Zaichkowsky 1985). Further, a confirmatory fac-
tor analysis was performed by specifying the posited relationships of
the observed variables to the underlying five dimensions, with these
allowed to intercorrelate freely. The covariance matrix was used as the
input data for the confirmatory factor analysis procedure available in
LISREL 8 (Joreskog and Sorbom 1989), which was performed using
the second subsample (N = 230).
Construct Identification and Validation
As already noted, the CIP scale includes 16 items. Both the Kaiser­
Meyer­Olkin measure of sampling adequacy and the Bartlett's test of
sphericity tests indicated that it was appropriate to perform a factor
analysis (Table 1). The result of the principle component factor analy-
sis indicated that there are five underlying dimensions (factors) (Table
1). In order to make sure that each factor identified by EFA had only
one dimension and each attribute loaded only on one factor, attributes
that had factor loadings of lower than .40 and attributes loading on
more than one factor with a loading score of equal to or greater than
914 TOURISTS' INVOLVEMENT
.40 on each were eliminated from the analysis. This procedure resulted
in elimination of five items and two factors.
Table 2 shows the items and the factors that remained after the elim-
ination of those five items and two factors. As presented in this table,
the EFA generated three unidimensional factors. The results were
somewhat consistent with the previous involvement research in leisure
and tourism. Like previous studies, the interest/importance and pleas-
ure dimensions of the original CIP scale were combined under one
factor, interest/pleasure. However, results of the EFA procedure indi-
cated that sign dimension of the original CIP scale is not a unidimen-
sional construct in measuring international leisure tourists' involve-
ment, and it should be eliminated from the analysis. Using the
Cronbach's Alpha reliability test, the reliability of the remaining factors
measuring interest/pleasure, risk probability, and risk importance
were determined to be 81.15, 68.21, and 59.22, respectively. Results of
this test indicated that only the risk importance dimension has a
reliability score below .60 (Alpha = 59.22). However, it was deemed
close enough to be considered as a reliable measure.
Table 2. Remaining Factors and Items Loadings (N = 230)
Items Pleasure/ Risk Risk
Interest Probability Importance
It gives me pleasure to purchase to purchase a .669
vacation.
Buying a vacation is like buying a gift for .686
myself.
A vacation is somewhat of a pleasure to me. .746
I attach great importance to a vacation. .722
One can say vacation destinations interests me .673
a lot.
Whenever one buys a vacation, one never .494
really knows whether it is the one that should
have been bought.
When I face a variety of vacation choices, I .599
always feel a bit at loss to make my choice.
Choosing a vacation destination is rather .693
complicated.
When one purchases a vacation, one is never .594
certain of one's choice.
It is really annoying to purchase a vacation .815
that is not suitable.
If, after I bought a vacation, my choice proves .710
to be poor, I would be really upset.
Eigenvalue 3.454 2.253 1.464
% of Total Variance 21.587 14.082 9.147
Cronbach's Alpha 81.15 68.21 59.22
The Bartlett's test of sphericity (significance level) .724
The Kaiser-Meyer-Olkin measure of sampling adequacy .000
Note: Items with a factor loading of higher than .40 are shown.
915GURSOY AND GAVCAR
Confirmatory Factor Analysis
A confirmatory measurement model that specifies the posited
relations of the observed variables to the underlying constructs, with
the construct allowed to intercorrelate freely was tested as rec-
ommended by Sethi and King (1994). The use of confirmatory factor
analysis ensures the unidimensionality of the scales measuring each
construct in the model. Items identified through EFA procedure were
utilized in the factor analysis. However, before testing the overall
model, the unidimensionality of each construct was assessed individu-
ally (Sethi and King 1994). Constructs with unacceptable fits were
respecified by deleting the indicators that had not worked out as
planned, to preserve the potential to have a unidimensional measure-
ment scale (Anderson and Gerbing 1988). A measurement model for
each pair of constructs was then estimated, combining them two by
two (Joreskog 1993).
First, each construct's fit was measured. After making sure that the
fit of each construct was acceptable, the fit of two constructs (a pair)
was measured. All constructs were paired with each other. Assessing
each construct individually and deleting unacceptable indicators
resulted in elimination of one indicator from the "Risk Importance"
construct. The items that remained after this step are presented in
Table 3. All of the composite reliabilities were above .7 with the excep-
tion of the risk importance construct The composite reliability score
of the risk importance construct was .61, which was below the .70 guid-
eline. However, it was determined to be close enough to consider due
to the fact that reliability scores that are between .60 and .70 represent
the lower limit of acceptability (Hair, Anderson, Tatham and Black
1998).
The overall fit of this final involvement measurement model was the
chi-square value with 32 degrees of freedom (2
(32)) = 45.57 (p =
0.057); the goodness - of - fit index = .96; the adjusted goodness - of
- fit index = .94; the normed - fit index = .91; the non - normed - fit
index = .96; the comparative - fit index = .97; the incremental - fit
index = .97; and the parsimony goodness - of - fit index = 0.56; the
parsimonious normed - fit index = .65; and the critical sample size
(N) = 269.79. The fit statistics suggested that the fit of the model was
acceptable. The 2
value was found to be nonsignificant suggesting an
acceptable fit. In addition, values of the goodness-of-fit index, the
adjusted goodness-of-fit index, the normed-fit index, the non-normed-
fit index, the comparative-fit index and the incremental-fit index range
from zero to 1.00 with a value above .90 indicating good fit (Byrne
1989). The parsimony goodness-of-fit index and the parsimonious nor-
mal-fit index scores were also found to be acceptable. Further, the
indicators of residuals, the root mean square, standardized the root
mean square and the root mean square error of approximation were
.051, .044 and .040, respectively. Additionally, convergent validity was
established for all indicators of each construct. Details on the proper-
ties of the measurements are provided in Table 3, showing a three-
factor measurement model of a tourist's involvement was established.
916 TOURISTS' INVOLVEMENT
Table 3. Laurent and Kapferer's CIP Scale (N = 230)
Constructs and Indicators Completely Indicator Error
Standardized Reliability Variance
Loadings
Pleasure/Interest 0.70a
0.79b
Buying a vacation is like buying a gift for myself. 0.59 0.33 0.67
A vacation is somewhat of a pleasure to me. 0.66 0.50 0.50
I attach great importance to a vacation. 0.88 0.70 0.30
One can say vacation destinations interests me a 0.67 0.40 0.60
lot.
Risk Probability 0.70a
0.76b
Whenever one buys a vacation, one never really 0.42 0.16 0.84
knows whether it is the one that should have
been bought.
When I face a variety of vacation choices, I always 0.71 0.35 0.65
feel a bit at loss to make my choice.
Choosing a vacation destination is rather 0.96 0.58 0.42
complicated.
When one purchases a vacation, one is never 0.71 0.37 0.63
certain of one's choice.
Risk Importance 0.66a
0.61b
It is really annoying to purchase a vacation that is 0.74 0.55 0.45
not suitable.
If, after I bought a vacation, my choice proves to 0.58 0.34 0.66
be poor, I would be really upset.
a
Variance extracted estimate.
b
Composite reliability of each construct.
The three constructs were pleasure/interest, risk probability, and risk
importance. Each of the three constructs was measured by multiple
attributes/indicators.
Test of Reliability and Validity
In an overall measurement model, the adequacy of the individual
items and the composites are assessed by measures of reliability and
validity. Two types of reliability measures--composite reliability and
the estimated percentage of variance extracted by each construct--
were examined. The composite reliability, as calculated with LISREL
estimates, is analogous to a coefficient alpha, which shows the internal
consistency of the indicators assessing a given factor (Hatcher 1994)
and is calculated by the formula provided by Fornell and Larcker
(1981). A value higher than .70 is acceptable for a composite reliability.
However et al (1998) argue that reliability scores that are between .60
and .70 are also acceptable because they represent the lower limit of
acceptability. As shown in Table 3, the composite reliability scores of
all constructs measuring tourists' involvement exceeded the acceptable
levels. The variance extracted estimate measures the amount of vari-
917GURSOY AND GAVCAR
ance that is captured by a factor. The desirable level of variance cap-
tured is 50% or higher (Fornell and Larcker 1981). Table 3 shows
that the variance extracted estimate for each factor also exceeded the
acceptable levels.
Three types of validity measures (discriminant validity, convergent
validity and nomological validity) were examined. To ensure that the
constructs are not measuring the same concept or ideas, the discrimin-
ant validity was assessed for each construct in the measurement model
by examining the constructs in sets of two (Joreskog 1993). Two mod-
els were tested for every possible pair of estimated constructs. The first
model was the constrained one where the correlation parameter was
constrained between each pair of constructs to 1.00. The second was
the unconstrained or free model, where the correlation parameter
between two constructs was not manipulated (not fixed at 1.00)
(Joreskog 1971). The 2
value was generated for both constrained and
unconstrained models with the respective degrees of freedom. After-
wards, a 2
difference test was performed on the two models. A signifi-
cantly lower 2
value for the unconstrained (free) model demonstrated
that discriminant validity had been achieved (Anderson and Gerbing
1988). Table 4 indicates that all of the constructs possessed discrimin-
ant validity.
Convergent validity is the overlap between alternative measures that
are intended to measure the same construct, but that have different
sources of undesired variation (Judd, Smith and Kidder 1991). One
method often used to estimate convergent validity for structural equ-
ation modeling studies is to examine the standardized confirmatory
factor analysis parameters' estimated pattern coefficient. (Marsh and
Grayson 1995). Convergent validity can be assessed from the measure-
ment model by determining whether each indicator's estimated pat-
tern coefficient on its posited underlying construct factor is significant
(Anderson and Gerbing 1988). Statistically significant large factor load-
ings indicate convergent validity.
Table 5 presents the standardized parameter estimates for the pro-
posed three-factor measurement model produced by LISREL. There
are three lines of information for each observed indicator. The first
line represents the estimate, the value in parenthesis on the second
Table 4. Results of Discriminant Validity Tests
Path Correlation 2
df 2
df Change Change Sig.Level
Value w/Corr. w/Corr. in 2
in df
Fixed Free
1­2 0.03 202.30 20 35.87 19 1 166.43 0.00
1­3 0.12 52.50 9 11.56 8 1 40.94 0.00
2­3 0.12 56.29 9 13.66 8 1 42.63 0.00
Note: Corr = Correlation; df = Degrees of Freedom; 1 = Pleasure/Interest; 2 = Risk
Probability; 3 = Risk Importance.
918 TOURISTS' INVOLVEMENT
represents the standard error, and the third represents the t-value. An
examination of the standardized parameter estimates in Table 5 reveals
all estimates exhibit the correct sign and size and are consistent with
the underlying theory. The table also presents that all of the estimated
pattern coefficients on their posited underlying construct factors are
significant at the 0.05 (each had a t-value > 1.96). Therefore, the
convergent validity was achieved.
After making sure that the scale has convergent and discriminant
validity, a confirmatory assessment of nomological validity of constructs
is conducted (Anderson and Gerbing 1988). In assessing nomological
validity, the causal relationships between the derived and other con-
structs or variables of interest are further examined to see if the
Table 5. Parameter Estimates for the Proposed Measurement Model (N = 230)
LAMDA X Pleasure/ Risk Risk
Interest Probability Importance
Buying a vacation is like Estimate 0.59
buying a gift for myself. SD (0.07)
t-value 8.30
A vacation is somewhat of a Estimate 0.66
pleasure to me. SD (0.06)
t-value 10.90
I attach great importance to Estimate 0.88
a vacation. SD (0.07)
t-value 13.29
One can say vacation Estimate 0.67
destinations interests me a SD (0.07)
lot. t-value 9.56
Whenever one buys a Estimate 0.42
vacation, one never really SD (0.08)
knows whether it is the one t-value 5.41
that should have been
bought.
When I face a variety of Estimate 0.71
vacation choices, I always SD (0.09)
feel a bit at loss to make t-value 8.14
my choice.
Choosing a vacation Estimate 0.96
destination is rather SD (0.09)
complicated. t-value 10.17
When one purchases a Estimate 0.71
vacation, one is never certain SD (0.09)
of one's choice. t-value 8.30
It is really annoying to Estimate .74
purchase a vacation that is SD (0.28)
not suitable t-value 2.62
If, after I bought a vacation, Estimate 0.58
my choice proves to be poor, SD (0.24)
I would be really upset. t-value 2.58
919GURSOY AND GAVCAR
derived constructs behave in the way stipulated by the underlying
tenets guiding the tourists' involvement scale development (Chen and
Hsu 2001). Nomological validity is determined by examining the causal
relationship between the derived construct and the construct that mea-
sures tourists' knowledge of the vacation destination.
Generally, there is strong support for the relationship between
involvement and knowledge (Havitz and Dimanche 1999). Kim et al
(1997) suggest that as the level of involvement increases, a tourist is
more likely to pay attention to any incoming information about the
destination, because high involvement indicates (approximately) per-
sonal relevance and importance. People are more likely to pay more
attention to incoming information about something that has personal
relevance and high importance (Celsi and Olsen 1988; Foxall and
Bhate 1993).
According to the elaboration likelihood model (Petty and Cacioppo
1983), consumers can follow two routes to persuasion. Central route
processing occurs when involved consumers seek product related infor-
mation intentionally. An involved consumer pays more attention to
incoming information and thoroughly processes the incoming infor-
mation (Petty, Cacioppo and Schumann 1983). This results in high
product knowledge. This type of cognitive response to incoming infor-
mation has been shown to mediate subsequent brand attitudes (Petty
and Cacioppo 1983; Petty et al 1983).
Peripheral route processing occurs when uninvolved customers lack
sufficient motivation to pay close attention to incoming information.
Since uninvolved customers pay less attention and fail to process
incoming information thoroughly, the level of knowledge they acquire
through incidental learning tends to be lower than involved customers'
knowledge. The literature suggests that involved consumers are more
likely to have better product knowledge than uninvolved consumers.
Using this thesis, a structural model was constructed to further test
the construct validity of the three-factor representing tourists' involve-
ment. In the structural model, three dimensions of involvement were
exogenous variables and the destination knowledge was an endogen-
ous variable (Figure 1). To determine the causal relationship between
involvement and knowledge constructs, three path coefficients were
estimated. Fit indices provided by LISREL indicated that model had
an acceptable fit (Table 6).
The structural equation modeling results revealed that all of the
"three path" coefficients were significant. Results of the structural equ-
ation modeling are shown in Table 6. Results indicated that all three
dimensions of involvement, interest/pleasure (r = .70), risk probability
(r = .13), and risk importance (r = .14), have positive direct impact
on tourists' destination knowledge. Three dimensions of involvement
explained 38% of the variance in tourists' destination knowledge. Con-
sistent with the information search and knowledge literature, results
of the structural equation modeling revealed that all three dimensions
of involvement had nomological validity.
920 TOURISTS' INVOLVEMENT
Figure 1. Structural Model
CONCLUSION
The research reported in this paper applies the multidimensional
consumer involvement profile (CIP) developed by Laurent and Kapf-
erer (1985) to leisure tourists at international vacation destinations.
Although this scale had been previously used in the recreation, leisure,
and tourism context, the objective of this study was to test the CIP
scale in an international tourism context with respect to reliability and
construct validity.
The study results supported the multidimensional nature of tourists'
involvement (Dimanche et al 1991; Park 1996). However, this did not
reveal the exact factors discussed by Laurent and Kapferer in their
Table 6. Structural Equation Modeling of Involvement and Knowledge
Variables Coefficients t statistics
Interest/Pleasure 0.70 6.79
Risk Probability 0.13 2.66
Risk Importance 0.14 2.41
Knowledge R2 = .45
2
= 92.54; df = 71; P = 0.44; the goodness - of - fit index = .95; the adjusted
goodness - of - fit index = .92; the parsimony goodness - of - fit index = .64; the non -
normed fit index = .96; the parsimonious normal - fit index = .69; the comparative -
fit index = .97; the incremental - fit index = .97; standardized root mean square =
.049; critical sample size = 250.02.
921GURSOY AND GAVCAR
study with consumer goods. Whereas Laurent and Kapferer's CIP pro-
duced five dimensions, the present research produced three dimen-
sions. The data also did not reveal the exact dimensions discussed by
Dimanche et al (1991) and other recreation, leisure, and tourism
researchers (Jamrozy et al 1996; Park 1996). Most have suggested
involvement has four dimensions because in the recreation, leisure,
and tourism context pleasure and importance dimensions are synony-
mous (Dimanche et al 1991, 1993; Jamrozy et al 1996).
Contrary to the foregoing, the findings of this study indicated that
international leisure tourists' involvement is rather a three dimensional
construct. The first dimension, pleasure/interest, was composed of a
mix of four importance and pleasure items. This result was consistent
with those reported in other studies in the leisure and tourism litera-
tures (Dimanche et al 1991, 1993; Jamrozy et al 1996). The second
dimension consisted of four items measuring risk probability. This
result was also consistent with previous studies (Dimanche et al 1991,
1993; Jamrozy et al 1996). The third dimension, risk importance, con-
sisted of two items. Again, this finding was consistent with earlier
research.
Previous studies in leisure, recreation, and tourism context also
identified a fourth or "sign" dimension. Results of the exploratory fac-
tor analysis indicated that the three sign items significantly loaded on
factor 3, as presented in Table 1, held together well in a conceptual
sense. However, the present data also suggested that the sign dimen-
sion identified by previous studies is not a unidimensional construct
in the context of international leisure tourism. Results of the explora-
tory factor analysis indicated that the variables proposed to measure
the sign dimension loaded on pleasure/interest and risk probability
in addition to that of sign. Based on Hattie's (1985) heuristic for elimi-
nating items with mixed loadings, these items were excluded from the
analysis. This contrasted with the findings of some other studies
(Dimanche et al 1991; Siegenthaler and Lam 1992). However, other
studies reported the sign dimension of involvement as being relatively
less important in terms of variance explained (Havitz and Dimanche
1997; Jamrozy et al 1996; Kim et al 1997). One reason that sign is not
distinct might be the simplicity of items used to measure sign value,
which fails to capture the possible multidimensional nature of sign
(Havitz and Dimanche 1997). Previous studies indicate that both per-
sonal identity and social identity components are present in sign
(Dimanche and Samdahl 1994).
In the international tourism context, there might be an overlap
between interest/pleasure and sign dimensions of involvement. Several
studies examined the relationship between personality (sign) and
interest and they reported overlaps and shared dimension between
them (Larson and Borgen 2002; Schneider, Ryan, Tracey and Rounds
1996). For example, researchers who studied personality and interest
by examining Holland's (1997) "Big Six model of interest" and the
"Big Five model of personality" reported a substantial overlap (Larson
and Borgen 2002). Schneider et al (1996) also reported a shared
dimension between personality and interest. The shared variance
922 TOURISTS' INVOLVEMENT
between pleasure/interest and sign might be explained by the overlap
between interest and personality.
Findings of this study provide an incremental contribution to inter-
national leisure tourists' involvement scale development. Dimanche et
al (1991) adopted Laurent and Kapferer's CIP scale and refined it to
be utilized in recreation, leisure, and tourism, with many using the CIP
scale version introduced by Dimanche et al. This study goes one step
further by applying the CIP scale to international leisure tourists.
Results presented here indicate that such involvement has a different
factor structure than that for structure of consumers of tangible goods,
and participants of leisure and recreational activities.
However, it would be foolhardy to claim that this paper has finalized
the international leisure tourists' involvement scale discussion in a
definite way. Rather, this study is considered as an initial step toward
the programmatic and systematic research into the examination of
international leisure tourists' involvement. Further investigation of the
subject is needed to conceptualize the understanding of the three
dimensions identified and their influences on international leisure
tourists' choice behavior and decision-making processes. More
research with other populations and destinations is required to con-
firm this factor structure of involvement in international tourism con-
text.
The current study results have limitations and caution should be
used in any generalization of the findings to other areas, populations,
and activities. Specifically, data were collected from international tour-
ists to Turkey and it is likely that they answered the survey questions
based on their motivations to visit this destination, with the outcome
being specific to this group. Therefore, involvement of tourists to other
international destinations, countries, and geographic regions should
be explored to see if the proposed measurement scale and the model
hold. In addition, tourism motivations of leisure tourists may influence
their level of involvement, which is likely to be different from those
wanting to escape. Understanding the motivations of tourists may
enable one to better understand the involvement of tourists.
Another limitation of the study was the interview language, since
English was not the first language of some of the respondents. As sug-
gested by Dimanche (1994), the assumption of English fluency is a
major limitation, especially when latent constructs such as involvement
are to be measured. For future studies, it is suggested that the inter-
views be conducted in the first language of respondents to eliminate
this limitation. Further, this study examined only international leisure
tourists' involvement. If the data collection and interviews were
expanded to include international tourists who travel for other pur-
poses (such as business, convention, and meeting) then there could
be different levels of involvement. Examining purposes other than leis-
ure may enhance the understanding of international tourists' involve-
ment. A
923GURSOY AND GAVCAR
REFERENCES
Anderson, J., and D. Gerbing
1988 Structural Equation Modeling in Practice: A Review and Recommended
Two-Step Approach. Psychological 103:411­423.
Babin, B., W. Darden, and M. Griffin
1994 Work and/or Fun: Measuring Hedonic and Utilitarian Shopping Value.
Journal of Consumer Research 20:644­656.
Backman, S., and J. Crompton
1989 Discriminating Between Continuers and Discontinuers of Two Public Lei-
sure Services. Journal of Park and Recreation Administration 7(4):56­71.
1991 The Usefulness of Selected Variables for Predicting Activity Loyalty. Leis-
ure Sciences 13:205­220.
Bloch, P.
1993 Involvement with Adornments as Leisure Behavior: An Exploratory
Study. Journal of Leisure Research 25:245­262.
Broderic, A., and R. Mueller
1999 A Theoretical and Empirical Exegesis of the Consumer Involvement Con-
struct: The Psychology of the Food Shopper. Journal of Marketing Theory
and Practice 7(4):97­108.
Brunswik, E.
1956 Perception and the Representative Design of Experiments. Berkeley CA:
University of California Press.
Byrne, N.
1989 A Primer of LISREL: Basic Applications and Programming for Confirma-
tory Factor Analytic Models. New York: Springer-Verlag.
Celsi, R., and J. Olson
1988 The Role of Involvement in Attention and Comprehension Processes.
Journal of Consumer Research 15:210­224.
Chen, J-S., and C.H-C. Hsu
2001 Developing and Validating a Riverboat Gaming Impact Scale. Annals of
TourismResearch 28:459­476.
Dimanche, F.
1994 Cross-cultural Tourism Marketing Research: An Assessment and Rec-
ommendations for Future Studies. Journal of International Consumer Market-
ing 6(3/4):123­134.
Dimanche, F., M. Havitz, and D. Howard
1991 Testing the Involvement Profile Scale in the Context of Selected Rec-
reational and Touristic Activities. Journal of Leisure Research 23(1):51­66.
1993 Segmenting Recreationists and Tourists Using Involvement Profiles. Jour-
nal of Travel and Tourism Marketing 1(4):33­52.
Dimanche, F., and D. Samdahl
1994 Leisure as Symbolic Consumption: A Conceptualization and Prospectus
for Future Research. Leisure Sciences 16:119­129.
Finn, D.
1983 Low Involvement isn't Low Involving. In: Advances in Consumer
Research 10:419-424.
Fornell, C., and D. Larcker
1981 Evaluating Structural Equation Models with Unobservable Variables and
Measurement Error. Journal of Marketing Research 18:39­50.
Foxall, G., and S. Bhate
1993 Cognitive Styles and Personal Involvement as Explicators of Innovative
Purchasing of Healthy Food Brands. European. Journal of Marketing
27(2):5­16.
Gursoy, D.
2001 Development of Travelers' Information Search Behavior Model. Unpub-
lished doctoral dissertation. Blacksburg VA: Virginia Polytechnic Institute and
State University.
Hair, J., R. Anderson, R. Tatham, and W. Black
1998 Multivariate Data Analysis (5th ed.). Upper Saddle River NJ: Prentice
Hall.
924 TOURISTS' INVOLVEMENT
Hatcher, L.
1994 A Step-by-Step Approach to Using the SAS System for Factor Analysis and
Structural Equation Modeling. Cary NC: SAS.
Hattie, J.
1985 Methodology Review: Assessing Unidimensionality of Tests and Terms.
Applied Psychological Measurement 9(2):139­164.
Havitz, M., and J. Crompton
1990 The Influence of Persuasive Messages on Propensity to Purchase Selected
Recreational Services from Public or from Commercial Suppliers. Journal of
Leisure Research 22:71­88.
Havitz, M., and F. Dimanche
1990 Propositions for Guiding the Empirical Testing of the Involvement Con-
struct in Recreational and Tourist context. Leisure Sciences 12:179­196.
1997 Leisure Involvement Revisited: Conceptual Conundrums and Measure-
ment Advances. Journal of Leisure Research 29:245­278.
1999 Leisure Involvement Revisited: Drive Properties and Paradoxes. Journal
of Leisure Research 31:122­149.
Havitz, M., F. Dimanche, and D. Howard
1993 A Two-sample Comparison of the Personal Involvement Inventory (PII)
and Involvement Profile (IP) Scales Using Selected Recreation Activities. Jour-
nal of Applied Recreation Research 17:331­364.
Holland, J.
1997 Making Vocational Choices: A Theory of Vocational Personalities and
Work Environments (3rd ed.). Odessa FL: Psychological Assessment
Resources.
Jain, K., and N. Srinivasan
1990 An Empirical Assessment of Multiple Operationalizations of Involvement.
Advances in Consumer Research 17:594­602.
Jamrozy, U., S. Backman, and K. Backman
1996 Involvement and Opinion Leadership in Tourism. Annals of Tourism
Research 23:908­924.
Joreskog, K.
1971 Statistical Analysis of Sets of Congeneric Tests. Psychometrka 36:109­134.
1993 Testing Structural Equation Models. In Testing Structural Equation Mod-
els, K. Bollen and J. Long, eds., pp. 294­316. Newbury Park CA: Sage.
Joreskog, K., and D. Sorbom
1989 LISTREL7: User&#8217;s Reference Guide (1st edition). Chicago: Scien-
tific Software, Inc.
Judd, C., E. Smith, and L. Kidder
1991 Research Methods in Social Relations (6th ed.). Forth Worth TX: Har-
court Brace Jovanovich College Publishers.
Kim, S., D. Scott, and J. Crompton
1997 An Exploration of the Relationships among Social Psychological Involve-
ment, Behavioral Involvement, Commitment and Future Intensions in the
Context of Bird Watching. Journal of Leisure Research 29:320­341.
Larson, L., and F. Borgen
2002 Convergence of Vocational Interests and Personality: Examples in an
Adolescent Gifted Sample. Journal of Vocational Behavior 60(1):91­112.
Laurent, G., and J. Kapferer
1985 Measuring Consumer Involvement Profiles. Journal of Marketing
Research 22(February):41­53.
Madrigal, R., M. Havitz, and D. Howard
1992 Married Couples' Involvement with Family Vacations. Leisure Sciences
14:285­299.
Marsh, H., and D. Grayson
1995 Latent Variable Models of Multitrait-multimethod Data. In Structural
Equation Modeling: Concepts, Issues and Applications, R. Hoyle, ed., pp. 177­
198. Thousand Oak CA: Sage.
Mittal, B.
1989 A Theoretical Analysis of Two Recent Measures of Involvement. Advances
in Consumer Research 16:697­702.
925GURSOY AND GAVCAR
1995 A Comparative Analysis of Four Scales of Consumer Involvement. Psy-
chology and Marketing 12:663­682.
Mittal, B., and M. Lee
1988 Separating Brand-choice Involvement from Product Involvement via Con-
sumer Involvement Profiles. Advances in Consumer Research 15:43­49.
Mountinho, L.
1987 Consumer Behavior in Tourism. European Journal of Marketing
21(10):5­25.
Park, S.
1996 Relationship between Involvement and Attitudinal Loyalty Constructs in
Adult Fitness Programs. Journal of Leisure Research 28:233­250.
Park, C., D. Mothersbaugh, and L. Feick
1994 Consumer Knowledge Assessment. Journal of Consumer Research
21:71­82.
Park, C., and S. Young
1986 Consumer Response to Television Commercials: The Impact of Involve-
ment and Background Music on Brand Attitude Formation. Journal of Mar-
keting Research 23:17­24.
Petty, R., and J. Cacioppo
1983 Central and Peripheral Routes to Persuasion: Application to Advertising.
In Advertising and Consumer Psychology, L. Percy and A. Woodside, eds. Lex-
ington MA: Lexington Books.
Petty, R., J. Cacioppo, and D. Schumann
1983 Central and Peripheral Routes to Advertising Effectiveness: The Moderat-
ing Role of Involvement. Journal of Consumer Research 10:135­146.
Rodgers, W., and K. Schneider
1993 An Empirical Evaluation of the Kapferer-Laurent Consumer Involvement
Profile Scale. Psychology and Marketing 10:333­345.
Rothschild, M.
1984 Perspectives on Involvement: Current Problems and Future Directions.
Advances in Consumer Research 11:216­217.
Schneider, P., J. Ryan, T. Tracey, and J. Rounds
1996 Examining the Relation between Holland's RIASEC Model and the Inter-
personal Circle. Measurement and Evaluation in Counseling and Develop-
ment 29:123­133.
Sethi, V., and W. King
1994 Development of Measures to Assess the Extent to which an Information
Technology Application Provides Competitive Advantage. Management
Science 40:1601­1624.
Sherif, M., and H. Cantril
1947 The Psychology of Ego Involvement. New York: Wiley.
Sherif, M., and C. Hovland
1961 Social Judgment. New Haven CT: Yale University Press.
Siegenthaler, K., and T. Lam
1992 Commitment and Ego-involvement in Recreational Tennis. Leisure
Sciences 14:303­315.
Sirakaya, E., R. Mclellan, and M. Uysal
1996 Modeling Vacation Destination Decisions: A Behavioral Approach. Jour-
nal of Travel and Tourism Marketing 5(1/2):57­75.
Tabachnick, B., and F. Fidel
1989 Using Multivariate Statistics (2nd ed.). New York: Harper Collins Pub-
lishers.
Teare, R.
1990 An Exploration of the Consumer Decision Process for Hospitality Ser-
vices. In Managing and Marketing Services in the 1990s, R. Teare, L. Mout-
inho and N. Morgan, eds., pp. 233­248. London: Cassell Educational.
Zaichkowsky, J.
1985 Measuring the Involvement Concept. Journal of Consumer Research
12:341­352.
926 TOURISTS' INVOLVEMENT
1987 IU Emotional Aspect of Product Involvement. In Advances in Consumer
Research, vol. 14, M. Wallendorf and P. Anderson, eds., pp. 32­35. Provo UT:
Association for Consumer Research.
Zeithaml, V., A. Parasuraman, and L. Berry
1990 Delivering Quality Service. New York: The Free Press.
Submitted 13 June 2002. Resubmitted 27 February 2003. Accepted 2 April 2003. Final
version 21 April 2003. Refereed anonymously. Coordinating Editor: Stephen L.J. Smith