INFORMATION TO USERS This manuscript has been reproduced from the microfilm master. UM I films the text directly from the original or copy submitted. Thus, some thesis and dissertation copies are in typewriter face, while others may be from any type of computer printer. The quality of this reproduction is dependent upon the quality of the copy subm itted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleedthrough, substandard margins, and improper alignment can adversely affect reproduction. In the unlikely event that the author did not send UM I a com plete m anuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion. Oversize m aterials (e.g., maps, drawings, charts) are reproduced by sectioning the original, beginning at the upper left-hand corner and continuing from left to right in equal sections with small overlaps. Each original is also photographed in one exposure and is included in reduced form at the back of the book. Photographs included in the original manuscript have been reproduced xerographically in this copy. H igher quality 6" x 9" black and white photographic prints are available for any photographs or illustrations appearing in this copy for an additional charge. Contact UM I directly to order. University Microfilms International A Beil & Howell Information C o m p a n y 3 0 0 North Z e e b R o a d . A n n Arbor, Ml 4 8 1 0 6 - 1 3 4 6 USA 313/761-4700 800/521-0600 O rder N u m b er 9208867 Causal m odeling o f tourist satisfaction: A pplication to M ichigan’s northw estern coastal tourism region Yokoyama, Fumito, Ph.D. Michigan State University, 1991 Copyright © 1991 by Yokoyam a, Fum ito. A ll rights reserved. UMI 300 N. ZeebRd. Ann Arbor, MI 48106 CAUSAL MODELING OF TOURIST SATISFACTION: APPLICATION TO MICHIGAN'S NORTHWESTERN COASTAL TOURISM REGION By Fumito Yokoyama A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Park and Recreation Resources 1991 ABSTRACT CAUSAL MODELING OF TOURIST SATISFACTION: APPLICATION TO MICHIGAN'S NORTHWESTERN COASTAL TOURISM REGION By Fumito Yokoyama This research was designed to investigate the process underlying tourists' degree of satisfaction with Michigan's northwestern coastal resort region. study included: (1) a comparison The objectives of the and evaluation of the discrepancy and the performance-based change models of tourist satisfaction, and (2) the identification of the key tourism constructs used to explain levels of tourist satisfaction. The data used for this study were collected in a separate study, the Michigan Travel Perception summer of 1987 and early winter of 1988. was a longitudinal panel study. Study, between the The original study The design of the study was a pre-post-post survey used to examine shifts in tourists' perceptions of Michigan's northwestern coastal resort area before and after their travel. who completed the pre-travel The sample was 107 respondents and the first post-travel surveys. Correlation analysis and path analysis were employed to evaluate the two alternative satisfaction models. the results of the analyses, Based on the performance-based change model was found to be the better model and was used to explain tourist satisfaction. The performance-based change model fit the data well, based on the results of path analysis. Fumito Yokoyama Confirmatory factor analysis and multiple regression analysis were used to investigate key tourism constructs which explained levels of tourist satisfaction. The confirmatory factor analysis identified and refined the following three tourism constructs: hospitality, and (1) (3) local outdoor tourism distinction, recreation. The (2) multiple regression analysis of tourist satisfaction on these three tourism constructs revealed that hospitality (Beta=0.52) explained the largest amount of tourist satisfaction compared to the other two constructs. The analysis also showed that these three tourism constructs together predicted levels of tourist satisfaction with some degree of accuracy regression coefficients.78; R-square=0.60). (multiple Based on these findings, implications for tourism planning and management and future research planners and are presented managers should and focus hospitality issues and tool to evaluate discussed: alternative upon (1) the tourism quality of puuu cnidxybib is a useiui causal models, and (3) confirmatory factor analysis is an effective technique for the measurement of underlying factors. Copyright by FUMITO YOKOYAMA 1991 ACKNOWLEDGEMENTS I would like to thank those who have helped, encouraged, and supported me for my entire student life. Without them, the completion of my doctoral degree would not have possible. To each of them, been I would sincerely like to say thank you for everything. Foremost, to Dr. Joseph D. Fridgen, my major professor, guidance committee chairman, and dissertation chairman, who challenged and encouraged me to complete my degree with his prompt feedback and intellectual comments on drafts of this dissertation. Dr. To the members of my dissertation committee,. Donald Holecek, Leefers, they dissertation. were Dr. Bonnie J. most helpful Knutson, and and Dr. supportive Larry of this The practical comments and guidance regarding the dissertation from both Dr. Holecek and Dr. Leefers are truly appreciated. I would especially Knutson for fields, providing an opportunity to do joint research, supporting opening my like to thank Dr. and eyes to the hospitality marketing encouraging me, both professionally and and personally. To inspire, my previous teach, and Japanese advisors encourage me to who were complete my able to doctoral degree: Professor Masaru Ikeda (National Institute of Fitness iii and Sports, Woman Kagoshima, College, Tokyo, Japan) and Dr. Seiji Osawa Japan), I would sincerely (Otsuma like to acknowledge them. To my friends and colleagues who shared the time of suffering through the comprehensive examination as well as sharing fun and intellectual conversations and a few drinks: Mr. Ju Hee Lee, Mr. Stephen Carlson, Mr. Sheng Jung Ou, and Mr. George Karlis, forward to I offer many thanks continuing out to them. friendship and I look professional relationship beyond graduate school. To my best friends and host family who have supported, encouraged, and taken care of me, in a very personal way: Mr. Masahiro and Mrs. Fumie Tomita, and Mrs. Lillian Kumata, I am truly pleased with their hearty caring and kindness. Finally, to my parents, Kozo and Teruko Yokoyama who have encouraged and supported me for my entire student life; without their understanding, encouragement, and caring, this dissertation would not have been coxupleted. offer special thanks to them. I would like to TABLE OF CONTENTS LIST OF TABLES .......................................... LIST OF F I G U R E S ......... Page vii xi Chapter I. INTRODUCTION .................................... 1 Significance of Tourism ...................... Tourism in Michigan .......................... Tourism Marketing ............................ Theoretical Background: Satisfaction Models .. 1 5 6 8 Discrepancy Model ....................... Performance-based change Model ......... Problem Statement ............................ Study Objectives ............................. Definitions Relevant to the Study ............ II. LITERATURE REVIEW ............................. Satisfaction Models and Theories ............. 8 9 10 12 13 14 14 Discrepancy Model ....................... 14 Theory ............................. Measurement cf Construct .......... 15 23 Comparison standard .......... Perceived performance ........ Disconfirmation ............... Satisfaction .................. 23 25 25 26 Application to the Current Study ... 27 Performance-Based Change Model ......... 28 Reinforcement Theory ............... 29 Dimensions of Tourism ........................ 32 Empirical Studies ....................... 35 Satisfaction Studies in Tourism .............. 40 v Chapter Page S u m m a r y ....................................... 44 III. METHODS ....................................... 47 Data Source ................................... Sample Characteristics ....................... Research Variables and Constructs ............ 47 52 62 Definition, Description, and Measurement .................... 62 Tourism attributes variables ...... Tourism constructs ................. Psychological constructs relevant to satisfaction models ....... 62 63 Expectation ................... Perceived performance ........ Comparison .................... Satisfaction .................. 67 67 68 68 Research Models ............................... 70 Discrepancy Model ....................... Performance-Based Change Model .......... 70 73 IV. ANALYSES TECHNIQUES AND PROCEDURES ............ 79 Data Analysis Techniques ..................... 67 79 Error of Measurement and Correction for Attenuation ........ 80 c a u n ................. ....................... . O j Confirmatory Factor Analysis ............ 89 Analysis Procedures .......................... Comparison of Satisfaction Models ...... Identification of Key Tourism Constructs Causing Satisfaction .... 97 97 99 Limitations of the S t u d y ..................... 100 V. FINDINGS AND DISCUSSION ........................ 104 Tourism Attribute Variables and Psychological Constructs ............ 104 Tourism Attribute Variables ............. 104 vi Chapter Page Psychological Constructs for the Satisfaction Models ........ 107 Comparison of Two Alternative Models for Satisfaction ................. 110 Correlation Analysis .................... Path Analysis ........................... Problems of Discrepancy Model and Control Variables ....... Summary of Findings Relative to Comparison of Satisfaction Models .. Identification of Key Tourism Constructs Related to Satisfaction ................. 110 115 120 123 123 Confirmatory Factor Analysis ............ Second-order Confirmatory Factor Analysis .................... Multiple Regression Analysis ............ Summary— Identification of Key Tourism Construct Relative to Tourist Satisfaction ............... 141 VI. CONCLUSIONS AND IMPLICATIONS .................. 143 Summary of the S t u d y ......................... Findings and Conclusions ..................... Implications for Tourism Planning and Management .... Future Research .............................. 143 147 124 132 134 153 155 BIBLIOGRAPHY ............................................ 153 APPENDICES .............................................. 166 Appendix A. PRE TRIP QUESTIONNAIRE OF MICHIGAN TRAVEL PERCEPTION STUDY ....... B. FIRST POST TRIP QUESTIONNAIRE OF MICHIGAN TRAVEL PERCEPTION STUDY ....... vii 166 169 LIST OF TABLES Page Reinforcement Theory: Attitude Change .... 31 Age Characteristics: Comparison of Research Sample and Regional Population ............ 53 Marital Status: Comparison of Research Sample and Regional Population ............ 53 Gender: Comparison of Research Sample and Regional Population ........... 54 Education: Comparison of Research Sample and Regional Population ........... 54 Income: Comparison of Research Sample and Regional Population ............ 55 Age Characteristics: Comparison of Research Sample and Michigan Travelers ............. 58 Marital Status: Comparison of Research Sample and Michigan Travelers ............. 58 Gender: Comparison of Research Sample and Michigan Travelers ............ 59 Education: Comparison of i\c^cuioil Sample and Michigan Travelers ............. 59 Income: Comparison of Research Sample and Michigan Travelers ............ 60 Tourist's Expectations and Perceived Performance of Tourism Attributes in the Target Area ........................... 105 Descriptive Statistics for Expectation, Perceived Performance, Comparison, and Satisfaction Constructs ................... 108 Frequency of Satisfaction Construct ...... 108 viii Table Page 5.4 5.5 Kurtosis and Skewness for Expectation, Perceived Performance, Comparison, and Satisfaction Constructs ...................... Ill Standardized Observed Correlation Matrix of Expectation, Perceived Performance, Comparison, and Satisfaction Constructs ..... 114 5.6 Corrected Correlation Matrix of Expectation, Perceived Performance, and Satisfaction Constructs in the Performance-Based Change Model .................................. 117 5.7 Path Coefficient, Standard Error of Estimates, and 95 Percent Confidence Interval of the Performance-Based Change Model .............. 117 5.8 Comparison between the Corrected Correlations and the Reproduced Correlations of Expectation, Perceived Performance, and Satisfaction Constructs in the Performance-based Change Model .............. 5.9 119 First Run of Confirmatory Factor Analysis on Tourism Constructs for Perceived Performance . 126 5.10 Second Run of Confirmatory Factor Analysis on Tourism Constructs for Perceived Performance . 128 5.11 Third Run of Confirmatory Factor Analysis on Tourism Constructs for Perceived Performance . 131 5.12 Fourth Run of Confirmatory Factor Analysis on Tourism Constructs for Perceived Performance . 133 5.13 Corrected Correlation Matrix due to the Error of Measurement of Tourism Constructs for Perceived Performance .................... 135 5.14 Second-order Confirmatory Factor Analysis on the Selected Tourism Constructs for Perceived Performance ........................ 136 5.15 Corrected Correlation Matrix of Tourism Constructs due to the Error of Measurement for Perceived Performance and Tourist Satisfaction Construct ....................... 138 ix Table Page 5.16 Multiple Regression Analysis of Tourist Satisfaction on Tourism Constructs for Perceived Performance ......................... x 138 LIST OF FIGURES Figure Page 2.1 Discrepancy Model ............................ 22 2.2 Performance-Based Change Model .............. 33 3.1 Map of Target Area ........................... 49 3.2 Sampling Structure ........................... 51 3.3 List of Tourism Attributes ................... 64 3.4 Proposed Tourism Constructs ................. 66 3.5 Cognitive Model of Discrepancy ............... 71 3.6 Path Diagram of Discrepancy ................... 74 3.7 Cognitive Model of Performance-Based Change .. 75 3.8 Path Diagram of Performance-Based Change ...... 78 4.1 Measurement Model of True Score and Observed Value ................................ 82 4.2 Path Diagram of a Hypothetical Causal Model .. 87 4.3 Exploratory Factor Model ..................... 91 4.4 Confirmatory Factor Model ................... 93 4.5 A Flow Chart of Procedural Steps for Comparison of Satisfaction Models ............ 101 A Flow Chart of Procedural Steps for Investigation of Key Tourism Constructs to Predict Satisfaction ...................... 102 5.1 Histogram of Expectation Construct ........... 112 5.2 Histogram of Perceived Performance Construct ..................................... 112 5.3 Histogram of Comparison Construct ............ 113 5.4 Histogram of Satisfaction Construct .......... 113 4.6 xi Figure Page 5.5 Path Analysis of Performance-Based Change 5.6 Type I and Type II Errors ..................... xii .... 118 140 CHAPTER I INTRODUCTION Significance of Tourism Tourism can be defined as a composite of activities, services, and industries that delivers (McIntosh and Goeldner, 1984) destination. tourism Types of travel experiences to people visiting a travel facilities include parks, resorts, ski areas, camps, hotels and motels, restaurants, and entertainment centers. Many activities and behaviors tourists are included in tourism such as fishing, of boating, swimming, sightseeing, and the art of traveling itself. In addition, tourism involves the providers and their activities. Providers travel include offices, researchers. facility owners, hospitality advertising services, companies, consultants, and Related activities include magazine advertising, travel guide book, promotion, sponsoring local traditions and publishing culture festivities, and developing regulations and standards related to tourism. Thus, tourism is a complex, comprehensive, and dynamic phenomenon related to the temporary movement of people away from their normal places of work and residence (Mathieson and Wall, 1982). Tourism industry, has become an important and significant both domestically and internationally. In 1990, total worldwide domestic and international travel and tourism revenue was estimated to be $2 trillion; total travel and tourism revenue spending in the U.S. was estimated to be $377 billion. Forty million foreign travellers visited the U.S., spending $52.8 billion, directly supporting over 700,000 U.S. jobs, and generating $6 billion in federal, state and local tax revenue in 1990 (U.S. Travel & Tourism Administration, 1991). Tourism is associated with economic, social, cultural, and environmental impacts— both positive and negative (Gunn, 1988). First, increasing positive tourism economic income within a community, impacts area include: or region, generating business and government tax receipts, and producing jobs. These can be expressed as both direct and secondary benefits. Direct benefits occur as a direct conseguer.es of tourism activity in the local area. For instance, tourist expenditures become business receipts which in turn are used to pay wages, taxes, and purchase goods. The benefits include indirect and induced benefits. secondary The indirect benefits can be generated by primary business outlays. Local businesses spend part of their receipts on necessary goods and services to investments. serve customers, including new equipment Their suppliers in turn have to purchase certain items and additional supplies from other distributors and suppliers, some local, others not. The induced benefits may be generated by the spending of primary income by tourists. Some of the wage and salary income directly generated by tourist expenditures are spent by local people on goods and services produced in the local area. These benefits also include government expenditures in the local area which were induced by the tax revenue generated by tourist spending (Frechtling, 1987). Second, various associated with cultural tourism. social tourism. and Gunn cultural (1988) impacts stated that are cross- exchange is probably thegreatest social value of People can be exposed to different political, religious, economic, cultural, and environmental systems by traveling to another country or region. Tourists can enrich a local area by providing new ideas, new social norms, and inducing a greater respect for local culture and traditions. Therefore, tourism can facilitate international understanding and broaden the views of both local residents and tourists. Third, tourism can contribute improvements within the local area. to environmental Local residents can take advantage of improvements in transportation systems, water and sewage systems, health services, and recreation facilities generated by the need to serve tourists. Local farmers may access markets more easily through new roads and may even develop new markets serving tourist needs directly. Thus, tourism may result in improving the living standard of local residents (Crandall, 1987). Additionally, some tourists exposed to cultural heritage and natural beauty have become friends of conservation and provide financial support for preservation of these precious local traditions and natural environments to the local area (Gunn, 1988). For instance, some of the people visiting Yosemite National Park in California may be impressed by its natural beauty and wildlife, and they start to recognize the importance of preserving these precious natural resources for generations to come. They may provide financial donations for the conservation of Yosemite, or for their own local parks. If not planned correctly, tourism can produce negative impacts— economic, social, and environmental (Gunn, 1988). With respect to economic impact, tourism development requires more costs present for constructing supplies— for facilities, new instance, and services; facilities new thus, increase tourism expanding attractions, expansion cf water supply, disposal, electrical power, and so on. development may and local waste Additionally, tourism land prices significantly; local residents cannot buy land and houses which may have been affordable before tourism development occurred. As for social impact, masses of tourists can destroy existing local traditions and cultures; produce congestion and competition for local services; and result in more crime or prostitution, values and between increase the potential local residents and for conflicts tourists. in Regarding environmental impacts, tourism development can produce environmental damage such as air and water pollution, and the destruction of a fragile ecosystem, or the degradation of the quiet environment of a small community. Tourism in Michigan Tourism has Michigan. to be brought significant economic impacts to In 1987, total travel expenditures1 were estimated $8,620 million, travel-generated employment2 was estimated to be approximately 150 thousands jobs, and travel­ generated tax revenue3 was estimated to be (U.S. Travel income, Data Center, 1989). employment opportunities, Thus, tourism provides and tax revenues for the country and individual states like Michigan. are rough estimates, but $1,080 million provide some These measures indication of the magnitude and value of tourism in Michigan. Within Michigan, the northwestern coastal area is known to be a popular tourism destination. This area ranges from Mackinac Island, a famous sightseeing spot, to the Traverse City area, a popular and beautiful resort destination region. 1Travel expenditure is defined as "the exchange of money or the promise of money for goods and service while traveling, including any advance purchase of public transportation tickets, but which may be purchased in advance of the trip". 2Travel-generated employment is defined as "the number of jobs attributable to travel expenditures in an area". 3Travel-generated tax revenue is defined as "these federal, state and local tax revenue attributable to travel in an area". These two regions have been found to be very popular among travelers to Michigan (Fridgen, 1987). many fine restaurants, attractions, stores, The entire area has resorts, natural and recreation opportunities. beauty, It is apparent that this region is a significant element of tourism within Michigan. Tourism is most often thought to produce positive economic, social, and environmental growth within states and localities. However, as stated earlier, tourism can generate negative impacts values, environmental properly. or (e.g., land price increases, destruction, etc.) conflicts in if not planned Fortunately, most negative impacts can be avoided eliminated through adequate tourism planning. Additionally, in order to maximize the benefits from tourism, proper management and planning is important in all stages of development and delivery of tourism opportunities and e v"4 n.e-— ri n r p s-. —e ---- Tourism Marketing The American Marketing Association's (AMA) revised definition of marketing is as follows: Marketing is the process of planning and executing the conception, pricing, promotion, and distribution of ideas, goods and services, to create exchanges that satisfy individual and organizational objectives (AMA, 1985, p. 1) . This definition states that marketing can apply not only to manufactured physical goods but also to non-physical ideas and services. Considering current trends of increasing competition and diverse consumer needs and wants, marketing becomes an essential tool for management and planning in both public and private organizations and institutions. Tourism is no exception. Tourism planning, marketing which is is the a useful process of adjunct to tourism preparing for tourism development. Tourism marketing differs from the marketing of manufactured goods in that the tourism experience is "intangible and incapable of being stored or transported" (Uhl and Upah, 1983, p. 231) . Tourists cannot examine the tourism experiences and quality of the site unless they have been there before. Thus, experience is one associated with of quality the generating most control over important tourist the tourism marketing satisfaction tasks (Mahoney, 1987). In tourism, marketing can directly contribute to tourist satisfaction (Fridgen, 1991). Satisfied tourists are more likely to return to the tourism destination and generate even more potential tourists through word of mouth promotion. Both repeat visitors and new tourists consequently contribute to the bottom tourism. line objectives of maximizing the benefits of Perception of the quality of tourism experiences will influence tourist satisfaction and this is crucial for the success of any tourism enterprise. Therefore, it is necessary to understand how tourist satisfaction is generated and to use this knowledge to better manage, plan, and market tourism experiences. Theoretical Background: Satisfaction Models Satisfaction, years in the as a concept, has been studied for many field of leisure, recreation, and tourism. Researchers have borrowed from such related disciplines as sociology, psychology, economics, marketing, and consumer research because very few theories or models of recreation or leisure satisfaction exist. Propst and Lime (1981) suggested that more emphasis should be placed on using concepts and methods from marketing and consumer satisfaction research in order to more fully explain the psychological process of satisfaction. Researchers theories taken sciences. useful to have employed from consumer a broad behavior and range of models and related social Several models have evolved and could be considered the study of tourism. Two models that appropriate to the study of tourism and this study are: the discrepancy model and are (1) (2) the performance-based change model. Discrepancy Model Individuals destinations are hold a variety who need or want to travel potential travelers. of pre-images or to selected Potential travelers expectations of travel destinations based upon a range of information from a number of sources such as travel-related magazines and journals, advertising, promotional campaigns, word-of-mouth promotion, and past experiences. the travelers' These images and expectations become subjective basis for selecting travel destinations and evaluating actual future travel experiences. Potential travelers may decide to actually make a trip to the selected destination and may enjoy their travel experiences. If the perceived expectations, travelers. this This travel should sense of experiences lead to more satisfaction surpass their satisfaction for should develop an improved perception of the destination in the minds of the travelers. If the perceived experiences fall below their expectations, this should result in less satisfaction and possibly in negative perceptions of the destination. perceived experiences meet their expectations, lead to satisfaction initial images or and travelers expectations, should If the this should maintain presumably their positive. Operationally, the model requires the following psychological constructs to performances, constructs are be present: disconfirmation, model expectations, and specific, satisfaction. based on perceived These discrepancy and performance-based change models developed in the literature. Performance-Based Change Model In this model, the first steps are similar to those in the discrepancy model travelers, (i.e., individuals become potential and they hold various expectations about travel 10 destinations). However, in this model the comparison process between expectations and account for satisfaction. are assumed to perceived does not Rather, only perceived performances influence travelers expectations. performances satisfaction, irrespective of Expectations can be thought of an initial attitude toward tourism experiences anticipated at the destination, and the initial attitude can be changed only by a message or input gained through experiencing the performance of the site or destination. performance. This can be called perceived Thus, individuals who perceived high levels of performance associated with a tourism experience should also experience a high degree of satisfaction. Conversely, the individual with lower perceived performance evaluations would also have lower satisfaction evaluations associated with travel to this particular destination. There are two alternative models for explaining tourist satisfaction which leads to the following question: model best explains tourist satisfaction? Which One way to answer this question is to empirically evaluate both the discrepancy model and the performance-based change model in terms of their ability to explain and predict tourist satisfaction. In order to carry out this comparison, a causal modeling approach is used and this technique enables the researcher to better understand tourist satisfaction and how tourist satisfaction 11 is developed. Beyond comparing the two models, it is important to investigate the applicability of consumer satisfaction to the study of tourism. Tourism satisfaction should be investigated empirically and systematically. This type of research has not been completed. Furthermore, once the tourist satisfaction process is clarified by comparing the alternative satisfaction models, the following question related to tourism planning and management can be addressed: What kinds of tourism constructs1 best explain and predict tourists' satisfaction? To conduct this investigation of tourism constructs, the first step is to identify a set of reliable tourism constructs and examine the relationship between resultant findings them and tourist satisfaction. The can then be used for strategic tourism marketing and planning by public and private agencies, tourism 4 M ^ 15 ^ m 4 A xiiu u 0 U 4 .^ cia/ To *3 aiiVA date, 4m w 4 aa ^ 4 a m m X iiau iu u L iu iib * causal modeling of tourist satisfaction is particularly lacking in the field of leisure, recreation, and tourism. A few satisfaction related studies have been reported; for instance, Pizam (1978) identified dimensions of 1A tourism construct represents some dimensions of tourism, for instance, sightseeing, hospitality, outdoor recreation, and so on. The construct can be measured by multiple items which would describe one aspect of the construct. Operationally, the term "construct" has the same meaning as the term "factor" in an exploratory factor analysis. Tourism constructs are content specific whereas psychological constructs are model specific based on the discrepancy models used for this study. 12 tourist satisfaction, relationship of Pearce pre-travel (1980) investigated favorability satisfaction with post-travel evaluation, and the post-travel and Maddox (1985) developed satisfaction measurements. Study Objectives The purpose of this study is to analyze the process of tourist satisfaction, as observed, measured, and reported by tourists visiting the northwestern coastal region of Michigan. A causal modeling approach is to be used. The study includes two phases: (1) an evaluation of the discrepancy model and the performance-based change model for tourist satisfaction, and (2) an investigation of what tourism constructs are the strongest predictors of tourist satisfaction. Specifically, the study objectives are to: 1. Quantify psychological constructs relevant to both the discrepancy model and the performance-based change model. 2. Compare the two alternative models— the discrepancy and performance-based change models— and evaluate which best explains tourist satisfaction with the northwestern coastal resort area in Michigan. 3. Evaluate unidimensionality1 of the tourism constructs ^nidimensionality can be defined as the existence of one latent trait or construct underlying a set of measures (Anderson, Gerbing, and Hunter, 1987). That is, if measures for a construct are unidimensional, it is concluded that the construct exits and can be measured by these measures. 13 identified. 4. Investigate the key tourism constructs which explain and predict tourist satisfaction with the northwestern coastal resort area. Definitions Relevant to the Study For purposes of this study, the following definitions were considered relevant: Attitude— an affective, evaluative, or emotional response to some object. Construct— any concepts or integrated set of concepts of which one is conscious (The Macmillan dictionary of psychology, 1989) . Expectation— the most likely performance of products and services that will be used and experienced in the future. Perceived performance— an individual subjective perception of actual performance for products and services based on one's use and experience after purchase. Satisfaction— an individual subjective attitude toward product or service based on one's use and experience. CHAPTER II LITERATURE REVIEW In this chapter, three fields of literature are reviewed: (1) satisfaction model, (2) dimensions of tourism, satisfaction in tourism studies. to the first study objective and (3) The first review is related of evaluating the consumer satisfaction model and the attitude change model for tourist satisfaction. model and This section includes the consumer satisfaction the attitude change model. The second explains what kinds of factors constitute tourism. review The third review covers how satisfaction has been studied in the tourism field and what kinds of problems have emerged from previous tourism satisfaction studies. Satisfaction Models and Theories Discrepancy Model The concept of consumer satisfaction has been a continuing central issue of marketing research and practice. The U.S. Department Satisfaction valid (Praff, information of Agriculture's 1972) about Index of Consumer was the first study to provide consumer 14 satisfaction to policy 15 makers. Since research has then, the volume increased of consumer drastically. These satisfaction studies have proposed numerous theoretical structures and developed the means to measure these constructs (Churchill and Surprenant, 1982). Two key issues relevant to consumer satisfaction and to this study will be reviewed and discussed in this section. They include: (1) the theory of consumer satisfaction and (2) the measurement of constructs in the consumer satisfaction model. Theory Researchers definitions for and practitioners the concept of have proposed consumer many satisfaction. Examples include: satisfaction is the function of an initial standard and some perceived discrepancy from the initial reference point (Oliver, 1980); consumer satisfaction refers to an individual evaluative response of outcomes of products and services (Westbrook and Oliver, 1380); consumer satisfaction is perceived quality or the difference between what consumers expect and what they receive (LeBoeuf, 1987; Parasuraman et al, 1986); and, finally, consumer satisfaction occurs when one's experience of a service offering matches one's expectations (Cina, 1989). Within the discrepancy model, it is generally agreed that consumer satisfaction/dissatisfaction (CS/D) may be defined as a consumer's discrepancy response between to the prior evaluation expectations of the perceived and the actual 16 performance of the product or service as perceived after its consumption (Tse and Wilton, 1988). have various available expectations about information prior experience with the product For instance, consumers products based on Based on their to purchase. after use, they their evaluate it comparing their expectations with their perceptions of the product's performance. Customers derive the experience of satisfaction or dissatisfaction from their evaluation. The nature of this comparative process needs elaboration and within the field of consumer further satisfaction there exist several theories to explain the process. These include: (1) discrepancy theory, (2) adaptation level theory, (3) assimilation-contrast theory, and (4) equity theory. Discrepancy theory (or expectancy disconfirmation theory as it is called sometimes), originated organizational behavior (Lawler, theory suggest that 1973) . satisfaction is in the field of Proponents of this determined by the differences between the actual outcomes a person receives and some expected outcome level. In effect, it actually involves two processes consisting of the formulation of expectations and the (dis)confirmation of those expectations by comparing performances (Oliver and DeSarbo, 1988). Oliver (1980) stated that consumers are believed to form expectations of a product prior to purchase. Subsequent purchase and usage actual level of performances of the product. reveal Consumers make a judgment of the product based on the comparison between 17 their expectations and the actual performance. The judgement is labeled positive disconfirmation if the product is better than expected, negative disconfirmation if worse than expected, and just confirmation if it is as expected. positive disconfirmation and just confirmation satisfaction, but to different degrees or levels. lead The to Negative disconfirmation results in various level of dissatisfaction. Helson's (1964) adaptation level theory can be considered as a basis of the previous discrepancy theory (Oliver, 1980; Oliver and DeSarbo, 1988). According to Helson's (1964) adaptation theory: [The theory] posits that one perceives stimuli only in relation to an adapted standard. The standard is a function of perceptions of the stimulus itself, the context, and psychological and physiological characteristics of the organism. Once created, the "adaptation level" serves to sustain subsequent evaluations in that positive and negative deviations will remain in the general vicinity of one's original position. Only large impacts on the adaptation level will change the final tone of the subject's evaluation (p. 461). In consumer satisfaction/dissatisfaction (CS/D) research, one's expectations of products or services can be seen as an adaptation level standard. The degree to which the product or service performance deviates from the level standard refers to the disconfirmation, as mentioned previously. Positive or negative deviations from the base level are caused by the degree to which products or services exceed, meet, or fall short of individual's expectations. be seen as an additive function Satisfaction, then, can of the expectation 18 (adaptation) level (disconfirmation). and This the subsequent paradigm has been considerable consumer satisfaction research judgements supported (Oliver, by 1980; Swan and Trawick, 1981; Bearden and Teel, 1983; LaBarbera and MaZursky, Wilton, 1983; 1988), predictions Churchill and and explains implied also by Surprenant, some 1988; Tse and counterintuitive assimilation-contrast theory. For instance, Bearden and Teel (1983) examined the antecedents of consumer satisfaction based on the data of 375 respondents in a two-phase longitudinal study of consumer experience with automobile repairs and services. The results supported the paradigm that expectations and disconfirmation were plausible determinants of satisfaction. Assimilation-contrast theory (Sherif and Hovland, 1961) suggests that expectations provide an anchor for subsequent judgement. function Assimilation or contrast effects will arise as a of the degree of differences that expectations and perceived performances. exist between So far the theory is the same as adaptation level theory, but if the differences are "not assumes too the discrepancies large," an individuals assimilation are effect reluctant from pre-held positions to occurs. It acknowledge (expectations); thus they assimilate their judgement toward the initial position. That is, should perceived result in performances higher higher subsequent than expectations judgement (greater satisfaction) than if performances meet the expectations. 19 However, if the difference is "very large," a contrast effect should occur. the difference. direction of This refers to a tendency to exaggerate Individuals magnify the discrepancy in the the disconfirmation. That is, perceived performances lower than expectations should result in lower subsequent judgement (less satisfaction) than if expectations match the performances. Moreover, this theory produces further questions: how small is "not too large" for the assimilation effect to occur, and how large is "very large" for the contrast effect to take place. Since these questions are very subjective for each individual, it is very difficult to measure. there are individual differences, In other words, a discrepancy which one person judges to be "not too large" could be perceived to be "very large" by others. Equity theory is thought to apply to any exchange where a person invests inputs in a urcm&cioLiuii (Oliver and Swan, 1989). culci receives outcomes The person compares the balance of his input and output for products or services with others. Consumer inputs involve time, money, shopping effort, and the like. Outcomes include product performance, services surrounding the sale, prestige of the brand, and so forth. If the input-output balance is proportionate to that of others, satisfaction with the product or service results. input-output balance If the is disproportionately higher or lower than others, more or less satisfaction results, respectively. 20 Comparison rules for deciding what is equitable include fairness and preference (Oliver and Swan, 1989). Fairness is thought to be a function of the equality of the input-output combination; thus both positive and decreases the feeling of fairness. negative Whereas, inequality preference is thought to be a function of the positiveness of inequality, preference should increase as positive inequality increases. Positive inequality results in embarrassment and guilt (Anderson et al., 1969), whereas negative inequality results in distress, resentment, and vindication (Walster et al., 1978) . In effect, however, positive inequality seems to increase satisfaction levels (Brockner and Adsit, 1986) . This theory also emphasizes the importance of others' input-output balance when an individual judges his/her own equity balance. inputs are disproportionately higher for one When person, satisfaction should increase as the person's outcome increases compared to those of others. seen as the degree of The basis of comparison can be equity which individuals perceive between what they obtained and what the other person obtained. Oliver and Swan (1989) suggested that the equity concept is an additional factor in the post purchase response, as is (dis)confirmation. Thus, the equity and confirmation processes can work together to determine levels of consumer satisfaction. In summary, consumer satisfaction is considered the 21 subjective response of individuals to their (dis)confirmation resulting from expectations perceived performance services. and the judgement comparison of of between products and In consumer satisfaction paradigms, confirmation and disconfirmation play most important roles in determining the level of satisfaction. proposed a discrepancy Churchill and Surprenant (1982) model based on previous consumer satisfaction studies which can be seen in Figure 2.1. In this model, satisfaction results from disconfirmation and perceived performance, disconfirmation is determined by comparison standard and perceived performance, and perceived performance is influenced by comparison standard. Their model has been widely used in the field of consumer satisfaction research for various kinds of products and services. It is assumed here that consumer satisfaction theory and the related model can be applied to the process of tourist satisfaction. This is the central theme of this study. To build a model of tourist satisfaction, concepts from each of these theories will be used. First, discrepancy theory will be used as a major basis for this study along with adaptation level theory, in order to develop causal models of tourist's theory satisfaction. will be used as Second, a the assimilation-contrast further interpreting the causal model. explanation basis for Third, the equity theory will not be used because of the limitations found in the secondary data available for this study, i.e., the data did not permit 22 COMPARISON STANDARD DISCONFIRMATION SATISFACTION PERCEIVED PERFORMANCE Note. This figure is adapted and modified from Figure I in an article of Churchill and Surprenant (1982). This is a path diagram which is a formal m eans to describe a causal model. The arrows in this Figure indicate causal relationships am ong variables in th e model. Figure 2 .1: D iscrepancy Model 23 an analysis related to the equity theory. Measurement of Constructs The CS/D psychological performance, theory encompasses constructs: the comparison disconfirmation, and following standard, four perceived satisfaction. These constructs have been operationalized and measured in various ways. A review of each follows. Comparison standard A comparison standard refers to anticipated performance and serves as a base for determining disconfirmation related to products and service offerings. The comparison standard has been operationalized as expectation (Oliver, 1980), ideal (Sirgy, 1984), or equity (Woodruff et al., 1983). The following explanation of the construct is based on Tse and Wilton's review (1988). Expectation, which is the most commonly used comparison standard, represents the most likely performance of products and services that will be used and experienced in the future. The construct is measured as what performance "will (probably) be." It is influenced by previous experience, advertising, learning from promotion, word of mouth, and so forth. Ideal refers to the optimal performance of products and services the consumers would hope for. performance "can be." It is It is measured as what influenced by previous experience, learning from promotional materials, and word-ofmouth communication. 24 Equity depicts a proportional performance based on the balance between an individual's inputs and outputs. It is measured as what performance "ought to be" relative to the input. The measurement of the construct may be influenced by the price paid, effort invested, advertising, and previous experiences. Researchers have not agreed on a single best conceptualization and operationalization of the comparison standard. However, Tse and Wilton (1988) investigated how these three alternatives of expectation, ideal, and equity as comparison process standards based on work the in the Churchill satisfaction and Surprenant's formation consumer satisfaction model (Figure 2.1). In the consumer satisfaction model, is assumed the comparison standard to have direct effects on disconfirmation and performance and an indirect effect their on satisfaction results measures indicated through disconfirmation. that among of comparison standard, the three However, alternative equity variable fails to produce a direct effect on performance and disconfirmation. Whereas, the construct ideal produces both a direct negative effect on disconfirmation and performance negative effect on satisfaction, and an indirect the expectation construct produces both a direct and positive effect on disconfirmation and performance and an indirect effect on satisfaction. Therefore, considering the degree and direction of effect of the alternative comparison standards, Tse and Wilton concluded 25 that expectation seems to be the best conceptualization for the satisfaction formulation process. Perceived performance Perceived performance refers to an individual subjective perception of the actual performance of products and services based on usage and experiences after purchase. is on subjective performance evaluation of The emphasis products and services by individuals rather than any objective performance. In other words, it is concerned with what performance was perceived to be by a person, not by objectively testing and evaluating the product or service. It is reasonable to assume that increasing performance should increase satisfaction with the exception of the case within assimilation theory (Churchill and Surprenant, 1982). Disconf irmation Disconfirmation is an important intermediate construct in the consumer satisfaction model. discrepancies performance. been former the comparison standard and subtractive (Oliver, 1980; approach (Latour and Peat, 1978) Churchill and Surprenant, can be expressed as an approach refers to perceptions — an and 1982) . algebraic difference between the standard and the performance. latter actual Two approaches for measuring this construct have proposed: subjective The between It represents perceived The individual subjective judgements of differences between performances and previously developed expectations. 26 Tse and Wilton (1988) suggested that the subjective approach was superior to the subtractive approach in a study of new, hand-held, miniature record players. satisfaction model, disconfirmation is direct positive effect on satisfaction. In the consumer assumed to have a The result showed both subjective and subtractive approaches produced direct positive effects on satisfaction. However, the subjective disconfirmation, was found to have a larger predictive power for satisfaction. Satisfaction Satisfaction research. is the Churchill critical and line Surprenant construct (1982) in defined CS/D and operationalized the satisfaction construct as follows: Conceptually, satisfaction is an outcome of purchase and use resulting from the buyer's comparison of the rewards and costs of the purchase in relation to the anticipated consequences. Operationally, satisfaction is similar to attitude in that it can be assessed as the sum of the satisfactions with the various attributes of the product or service (p. 493). Thus, satisfaction may be operationalized as an individual's subjective attitude is attitude based toward on usage a product and or experience service. resulting actual performance compared to anticipated performance. This from In the consumer satisfaction model (Figure 2.1), satisfaction may be directly influenced by both disconfirmation and perceived performance, and indirectly affected by both the comparison standard and perceived performance. Regarding the measurement of satisfaction, Westbrook and 27 Oliver (1980) postulate: Most often, simple, single-item rating scales are employed ... multi-item rating scale measures have found application infrequently, despite their potential to reduce measurement error ... It is doubtful that the cognitive-valuative, affective, and conative elements of satisfaction can be adequately measured in a single 5- or 7-point scale "very satisfied — very dissatisfied" rating scale (p. 94). They tested five different satisfaction instruments using the multitrait-multimethod analysis: (1) verbal scale, (2) graphic scale, (3) Likert scale, (4) semantic differential (SD) scale, and (5) inferential scale. The results suggested that the scales with the highest reliability are the SD and Likert scales. Application to the Current Study Based on the previous review and available data related to tourist satisfaction formation process, the following measurements will be used in this study taking into account the fact that the data already existed in its original form. The comparison standard will be expectation, i.e, what are the anticipated performance levels of various tourism attributes at the destination, estimates made prior to travel. performance performance travel. is a related tourist's to the subjective same tourism Perceived perception attributes of after Disconfirmation, which is renamed comparison in this study, will be a mathematical difference between the scores of expectation and performance for each of the tourism attributes. Satisfaction will be an attitudinal response of 28 overall satisfaction destination. about tourism experiences at the All psychological constructs will be measured by multiple items with the exception of satisfaction construct, which is a one item measure. Performance-Based Change Model The term attitude has been used to describe emotional states, behavioral dispositions, beliefs, opinions, and perceived social distances (Hunter, Danes, and Cohen, 1984). Hunter et al. (1984) offered the definition of attitude as an affective, evaluative, or emotional response to some object. Attitudinal persons. object objectives For of an instance, may be a ideas, places, things, "tourism experience" attitude— "I am satisfied with experience" may be the attitudinal response. may be the or an tourism This definition of attitude will be used throughout of the study. Messages play a central role in stimulating a change in attitude, and there are two types of messages• excernai and internal (Hunter, Danes, and Cohen, 1984). External messages are those transmitted from a source like a speaker or mass media to a receiver like a listener or consumer. messages individual are those receiver; transmitted thus they within occur the when Internal mind the of an individual thinks about an attitudinal object. Attitude change theories can be grouped into three classes: (1) reinforcement theories, (2) affective consistency theories, and (3) cognitive consistency theories (Hunter, 29 Danes, and Cohen, 1984). Reinforcement theories assume that a message with positive or negative affect will positive or negative attitude change toward induce a an object. Affective consistency theories assume that attitude change depends on between the the Cognitive emotional feelings consistency compatibility of the or receiver theories predict incompatibility and the belief source. from other beliefs rather than predicting belief change from a message. Recent cognitive consistency theorists attitude is a special kind of belief. have argued that Within the context of this study, the reinforcement theory is the most appropriate among the three theories. For instance, expectation of tourism experience (object) can be considered as an attitude; perceived performance can be viewed as an message; satisfaction with a tourism experience (object) can be seen as an attitude; thus, a high perceived performance (message) will result in experience. a high level of satisfaction with a tourism The reinforcement theory is, therefore, reviewed in this section. For review of the affective consistency and cognitive consistency theories of attitude change, the reader is directed to Petty and Cacioppo (1981). Reinforcement Theory Proponents of this theory assume that the message from the source either reinforces or weakens a receiver's initial attitude, which attitude change. causes either a positive or a negative A person's attitude is a response toward an 30 object and it can either be reinforced or weakened based on the direction, positive or negative nature of the message. The attitude change process is depicted in Table 2.1. A positive message has a positive impact on an initial positive attitude and a negative message has a negative impact on an initial negative attitude. On the other hand, "punish positive object attitude" means to weaken an initial attitude toward an object. That is, to weaken a positive attitude is to cause a negative impact on the initial positive attitude, which consequently leads to a negative attitude change. fact, the opposite is true. In Therefore, the algebraic sign of attitude change is predicted by the same direction of the message value in the reinforcement theory, that is, sign(Aa) -sign(m) where: delta a = attitude change; m = message. The theory argues that all receivers should react in the same direction as the message's direction. cause positive change; thus, each A positive message will receiver will be more favorable or less unfavorable toward the object after the message. A negative message will produce negative change; thus, each receiver will be more unfavorable or less favorable toward the object as a consequence of the message. In order to apply this theory to tourist satisfaction, constructs in the model are conceptualized as follows: 1. Expectation is an initial attitude toward an 31 Table 2.1 Attitude toward obj ect Reinforcement Theory: Attitude Change Message value Reinforcement process Attitude change Positive Positive Strengthen the attitude, more positive toward object Positive Positive Negative Weaken the attitude, less positive toward object Negative Negative Positive Weaken the attitude, less negative toward object Positive Negative Negative Strengthen the attitude, more negative toward object Negative Note. From Mathematical Models of Attitude Change (p. 11) by J. E. Hunter, J. E. Danes, and S. H. Cohen, 1984, Oriando, FL: Academic Press, Inc. Attitude toward object = Expectation Message = Perceived performance Attitude change = Satisfaction Object = Tourism experience 32 anticipated tourism experiences (object); 2. Perceived performance is a message from actual tourism experiences; 3. Satisfaction tourism is an attitude experiences as a change toward consequence of the the message. The performance-based change model of the reinforcement theory is provided in Figure 2.2. directly determined by In this model, perceived satisfaction is performance. Perceived performance is influenced by expectation. Dimensions of Tourism In this third review section, dimensions of tourism are reviewed and clarified. According to McIntosh and Goeldner (1984), tourism is a composite of activities, services, and industries that deliver a travel experience. transportation, hotels and motels, It may include restaurants and bars, shops, entertainment and/or recreation activity facilities, and other hospitality services. This suggests that tourism has multidimensional constructs. Lew (1987) comprehensively reviewed past studies of tourism attractions and developed a framework to guide tourism attraction research. He identified three broad perspectives based on his review of the literature: organizational, and (3) cognitive (1) ideographic, perspectives. (2) These perspectives are not independent of each other, and do have 33 EXPECTATION SATISFACTION PERCEIVED PERFORMANCE Figure 2.2: Perform ance-B ased C hange Model 34 some overlap, but it provides a useful framework to use in this study. First, the ideographic perspective refers to descriptions of attraction types in a tourism area, i.e., natural beauty, recreation facilities, organizational tourism i.e., infrastructure. perspective has three region: spatial, unstructured vs. capacity, structured, growth, short-term vs. long term. Second, the characteristics of a and temporal slow growth features, vs. rapid This perspective has often been used for tourism planning studies (e.g., Gunn, 1980; Peck and Lepie, 1977; Rodenberg (1980). Finally, the cognitive perspective represents tourists' perceptions and experiences in the tourism site, i.e., expensive, friendly people, quiet, relaxed. This ideographic perspective perspective. sometimes For overlaps example, with "tennis" is the an ideographic category (description), but it becomes a cognitive category (experience) through participation in the activity. Many tourism image studies can be categorized into Lew's cognitive perspective, ideographic perspective, or combination of the two perspectives (e.g., Britton, Crompton, 1986; 1979; Haahti, Ritchie and Zinns, 1978; 1979; Woodside et al., concerned with tourism empirical studies using Shih, 1986). 1979; Cohen, 1979; Henshall and Roberts, 1985; 1986; Since this study is mainly perceptions the Thompson and Cooper, and ideographic images, and/or several cognitive perspectives will be reviewed in terms of tourism constructs 35 and these attraction categories. Empirical Studies Gearing, Swart, and Var (1974) developed a group of seventeen criteria to judge tourism attractiveness of tourism sites in Turkey. constructs: (1) These criteria were organized into five natural historical factor, and (5) factor their factor, social factor, (3) (4) recreational and shopping facilities, infrastructure, food and includes two criteria: social (2) factor includes shelter. Their natural natural beauty and climate; four criteria: artistic and architectural features, festivals, distinctive local features, and fairs and exhibits; their historical factor includes three criteria: ancient ruins, historical prominence; facilities includes religious their five significance, recreational criteria: and sports and shopping facilities, educational facilities, facilities conducive to health, rest, and tranquility. nighttime recreation, and shopping facilities; and their infrastructure, food, and shelter factor includes two criteria: infrastructure facilities above minimal tourism quality. and food & lodging A weight (relative importance) was assigned to each criterion based on interviews with twenty-six tourism experts unidimensionality1 of these criteria. without testing Then, a weighted total, which is a quantified index of tourism attractiveness, was 1Unidimensionality may be defined as the existence of one latent construct underlying a set of measures. 36 calculated for each tourism area in order to compare and rank all tourism areas. attractiveness Columbia, was This Canada applied (Var, same to approach tourism Beck, and of regions Loftus, tourism in British 1977). They concluded that the procedure mentioned previously provides a means for establishing a quantified index of touristic attractiveness to be used as a basis for making comparisons among touristic regions and districts. Pizam, Neuman, and Reichel (1978), applying an exploratory factor analysis to evaluations by 685 tourists on a site, identified eight factors associated satisfaction with Cape Cod, Massachusetts. they found were: (1) beach (4) eating and drinking accommodation facilities, (6) campground (2) cost, facilities, facilities, environment, and (8) extent of commercialization. factor (beach opportunities) contains tourist The eight factors opportunities, hospitality, with beach (3) (5) (7) The first facilities, cleanliness of beach areas, parking availability, and space availability on the beach. The second factor (cost) includes the cost of vacationing, general goods and services, and the quality of goods and services relative to price. The third factor (hospitality) consists of willingness of residents (or employees) to aid tourists, general friendliness of employees, courtesy of residents (or employees) general hospitality of residents. toward tourists, and The fourth factor (eating and drinking facilities) includes availability of restaurants, 37 cafeterias, and bars; quality of eating and drinking places; and quality of service in eating and drinking places. fifth factor (accommodation facilities) facilities (or service) The contains quality of in hotels/motels. The sixth factor (campground facilities) consists of availability (or quality) of campground facilities. includes scenery environment. and The seventh factor natural attractions, (environment) and quality of The eighth factor (extent of commercialization) contains a single category of extent of commercialization. They concluded that these factors are not universal, but rather depend on the characteristics of the destination area such as facilities, attractions, weather and so forth. Pearce (1982), by comparing tourists' images of Greece and Morocco, pre- and post­ investigated whether or not tourists change their images and perceptions following travel to a destination. He developed 13 constructs of holiday travel environments summarizing 270 conscruccs produced by 10 subjects. include: The thirteen constructs (measured with one item) (1) cheap shopping, (2) adventurous holiday, contact with local peasant people, (5) interesting tourist sights, (3) (4) exotic local customs, (6) spectacular scenery, (7) appealing food, (8) swinging social life, (9) absence of other tourists, (12) (10) good sun and beaches, interesting politics personal attraction. and (11) good winter sports, society, and (13) strong He concluded that tourists may change their perception of the holiday environment after they visit 38 the country. Furthermore, the perceptual changes of the destination country may reflect on one's home country as well. He also suggested the need for more research on tourists' attitudinal changes in the field of tourism research. Klenosky (1985) investigated the stability of underlying tourism dimensions across samples for the same destination and across destinations for the same sample, by applying an exploratory factor analysis to 27 tourism-related dichotomous adjectives. He found three stable dimensions: (1) environmental excitement, (2) undeveloped tranquility, and (3) service orientation. excitement The first dimension of environmental includes: sandy, fun, enjoyable, colorful, appealing, delightful, outdoor-oriented, and interesting. The second of: dimension of undeveloped tranquility consists secluded, peaceful, pleasant, unspoiled, quiet, natural, and restful. The final dimension of service orientation contains: friendly, courteous, forested1, middle-class-oriented, familyoriented, and delightful. In summary, multidimensional destination area. these studies nature of empirically tourism supported characteristics the in a Although most of the identified constructs fall into Lew's ideographic perspective framework, there is no agreement of a single universal tourism construct tourism destinations. 1This dimension. for all Recalling an implication by Pizam et item should appear in the first or second It may be a typographical error in the article. 39 al. (1978), the identified tourism factors are not universal, rather they are destination specific. Tourism characteristics largely depend on tourism related attributes offered destination area. in a For example, if a tourism destination is located in a coastal area, the tourism characteristics should reflect the tourism attributes specific to that destination, e.g. , water related activities like boating, fishing, swimming and so forth. Therefore, it is essential that local tourism related attributes be considered in order to establish tourism constructs. Furthermore, no studies tested unidimensionality within each construct or reported reliability coefficients of each construct. Without testing the unidimensionality, there is no guarantee that all items in the identified factor, which are supposed to measure one construct, actually measure the same underlying construct. In terms of reliability coefficients, suppose construct A has a high reiiauxxiuy coefficient and construct B has low reliability. A quantitative relationship (i.e., correlation) between these two constructs A and B will be attenuated depending on both reliability coefficients. A further explanation and correction of this problem will be presented in Chapter IV. These unidimensionality and reliability issues should be considered in developing a valid and reliable measurement for underlying tourism constructs. 40 Satisfaction Studies in Tourism Little was known about tourism satisfaction, its components, measurement, determinants, and consequences in the 1970s. Pizam et al. (1978) applied the consumer satisfaction approach1 to tourist satisfaction with a destination area— Cape Cod, Massachusetts. They tried to empirically identify the components of tourism satisfaction and suggested methods to measure them (i.e., exploratory factor analysis). Dann (1978) criticized their article on theoretical and methodological grounds. For example, he suggested that satisfaction is not a function of specific satisfaction with the service offered by the destination, but a function of overall life satisfaction (a person's health, current marital happiness, degree of anomie, and so on). He also argued that questionnaires or interviews using Likert-type scales were not appropriate instruments to measure tourist satisfaction (because their results vers highly skewed and did not result in normal distributions), and he suggested that researchers use unstructured observational methods instead. Pizam et al. (1979) replied to Dann's critique in terms of concept confusion and faulty methodological assumptions. They responded that Dann's comment, which expressed that one's attitudes are only a function of overall life satisfaction, was contrary to the findings of modern psychological research. 1As reviewed earlier, tourist satisfaction is the result of the comparison between tourists' perceived performances of a destination and their expectations about the destination. 41 These psychological findings indicated that a person will only rarely exhibit a large degree of unity of attitudes based on his single ideology or life philosophy. Thus, they argued, believing that every single satisfaction with a product or service is a function of one's overall life satisfaction, in spite of service, the quality and expectations is not reasonable. Likert-type scale and of the product or Dann's criticism of using the questionnaires approach ignores the whole range of literature on social science methodology and the field of satisfaction research. The Likert-type scale is the most widely used measurement technique applied to the assessment of social attitudes because it is highly reliable, simple, and responses. are has a wide range of possible alternative In contrast, unstructured observational methods difficult to quantify, unreliable, and subsequent to personal bias by the observer. Van Raaij and Francken (1984) defined satisfaction as the difference tourism, consumer between adopting expectations and actual performance the most widely used definition satisfaction literature. The in in the disconfirmed expectation or the unfair equity of costs and benefits create dissatisfaction. They applied discrepancy theory and equity theory to the study of the satisfaction and dissatisfaction in a tourism setting. However, they did not show any empirical support for applicability of their definition of satisfaction. Botterill (1987) questioned Van Raaij and Francken's 42 definition of satisfaction. If satisfaction can be achieved by diminishing the difference between tourist expectation and vacation performance, has a highly satisfied tourist made a perfect prediction of the outcome of the vacation? But, how did the traveler perfectly predict performance of tourism experiences prior to travel? This seems unlikely. More importantly, unpredictable tourism related events seems to lie at the heart of tourism experience. For instance, when a tourist goes fishing in a destination area and catches the largest salmon caught that day. Perhaps a salmon fishing contest was going on the area which the tourist did not know about, and the tourist entered and won first prize. Both events of catching a large salmon and getting the first prize increase satisfaction the tourist's experience relative to the destination. level of tourism Yet, these events and experiences were not predicted or expected. Van follows. Raaij It (1987) cannot replied be to denied Botterill's that comment tourists have as some expectations prior to travel, derived from advertising, travel guides, past experiences, and word of mouth. kinds of expectations: (1) general specific and detailed expectations. There are two expectation and (2) Some examples of general expectation include good restaurants, nice climate, beautiful leaches, and so on. Some examples of specific expectation are uncrowded no waiters. beaches, loss of luggage, and very friendly A general expectation is more likely to be confirmed 43 than a specific and detailed expectation. Therefore, tourists do not have to make perfect predictions of their vacations and they may have Fulfilled both general satisfaction general and/or level. and specific specific This does expectations. expectations not exclude the increase fact that unanticipated events may increase satisfaction as well. is, good unanticipated events may increase the That perceived performance of tourism experiences, which in turn result in higher satisfaction. Maddox (1985) tested the validity of three global satisfaction scales (i.e., graphics, delighted-terrible, and the face scale) using data (n=411) from a satisfaction study with tourism in Nova Scotia, Canada. A multitrait-multimethod (MTMM) approach was employed to test convergent validity1 and discriminant validity2. The results indicated that convergent and discriminant validity were confirmed. both Thus, all of the satisfaction scales were measuring the same thing, and satisfaction construct. with Within delighted-terrible tourism the scale is three showed a distinct, satisfaction the largest measurable scales, the convergent validity. Convergent validity means that a measure should positively correlated with other measures of the same construct. 2Discriminant validity means that a measure should not correlate with theoretically unrelated measures or constructs. 44 Summary Consumer satisfaction research has provided much evidence to support services. discrepancy theory for manufactured goods and The review of past studies of consumer satisfaction provides some clues to tourist satisfaction the in application terms of constructs, and research Likert is the among best the theory measurement psychological scale of design. to scale, First, other measurement a scales. Second, expectation, perceived performance, disconfirmation, and satisfaction constructs are major research constructs. Third, pre and post survey design is needed to test satisfaction models. Performance-based change theory has introduced a different view of satisfaction formulation relative to the discrepancy theory. The performance-based change theory does not include disconfirmation construct in the model, which is the most model important mediating consurucu (cf. Figure 2.1 and 2.2). in In ui8 discrepancy other words, the performance-based change model assumes that satisfaction is a function of perceived performance only, whereas the discrepancy model assumes that satisfaction is a function of both disconf irmation and perceived performance. This conflict between the models has to be empirically investigated. In the field of tourism research, the review of satisfaction in tourism has shown some conflicts of dialogue among tourism researchers in terms of tourist satisfaction 45 formulation. Pizam et al. (1978) and Van Raaij et al. (1984) support the application of the consumer satisfaction approach to tourist satisfaction. satisfaction is based They on the propose that differences tourist between their expectations and the actual performance of a destination area. Dann (1978) argues that tourist satisfaction is a function of overall life satisfaction, whether a highly and Botterill satisfied tourist prediction for the vacation. has However, neither made of the a perfect the (1979) and Van Raaij original studies responses to criticism were supported empirically. applicability questioned These arguments and questions have been responded to by Pizam et al. (1987). (1987) consumer satisfaction nor the Therefore, approach to satisfaction in tourism should be investigated scientifically. Few studies related to tourism satisfaction have been done because of the complicated nature of the construct of tourism satisfaction. Also, to study the satisfaction formulation process illustrated earlier requires panel data with a pre and post design. Some researchers attempted to investigate a portion of the satisfaction formulation process. Pizam et al. tourist (1978) identified eight factors (constructs) of satisfaction. Maddox satisfaction with construct. However, satisfaction formulation conducted. tourism no as (1985) an independent empirical process clarified study for of tourism that measurable the has whole been The satisfaction formulation process in tourism 46 should be investigated empirically. In addition to this need for empirical support, this kind of study should enhance an understanding of tourist behavior, including Further, satisfaction, perceptions, and travel patterns. this type of research should contribute to better tourism marketing and planning efforts. CHAPTER III METHODS The methods sources of data, variables and chapter (2) is sample constructs, divided into four parts: characteristics, and (4) research (3) (1) research models. The chapter begins with on explanation of the data source used for this study. Then, characteristics of the study sample are described and compared to (1) data from the north central census region and (2) travelers visiting Michigan. Finally, definitions and measurements of the research variables and constructs are introduced, and two alternative models of tourist satisfaction are presented and explained. Data Source This study is based on data originally obtained in the Michigan Travel Perception Study (Fridgen and Hsieh, 1989). Very few studies have examined tourists' perceptions before and after travel to a tourism destination. Thus, little was known about how tourists' perceptions change over the course of their travel experience. The goal of the original study was to explore how tourists' perceptions changed over time, 47 48 before and specific after vacation objectives: (1) travel. The identify study positive had and three negative changes in perceptions following visits to the northwestern coastal zone stability of in Michigan, perceptions (2) and determine images the over strength time and following vacation travel to Michigan's coastal zone, and (3) estimate what proportion of potential tourist to Michigan's coastal zone actually make the trip. The study was basically a longitudinal panel study (i.e., pre, post, post-post design) of travelers' perceptions of the northwestern coastal region of Michigan, the target region of study (Figure 3.1). of out-of-state tourists were asked to A group provide their perceptions of the target region before and after their trip to Michigan. There are several destination region for study, area is known restaurants, to be quality a reasons to select this including the following: popular tourism resorts, area extensive with the fine recreation opportunities and beautiful natural scenery; it is a popular tourism area within Michigan and within the central midwestern part of the U.S., sought out by residents and non-residents alike; and the area has been studied before (Fridgen, 1987). The population of the study was defined as individuals who were not Michigan residents who called the Michigan Travel Bureau requesting travel information about the target region between July and individuals were September, 1987. A identified as potential total of 2,265 travelers to the 49 C E D A *• w ill! ST IGNACE^es>“ ' CHtNEtu* UL->tN&C > Q hog JWACKINAC STRAITS i /C R O S S V ILLA G E CHEBOY GAN PELLSTO N ALOHA 2'2 BAY SH O R E BOYNE' C IT Y ELLSW ORTH l 20 EAST N O R TH .P O R T EA ST JO R D A I BOYNE 1FALLS PO R T C E N T R A L '' )lakeX^ o S W A N lT O U /N •e [B E LLA lRE GLEN ELK R A PIO S <3w wpireh!?JLE ELAN A U W /\ 13i M A N C E' Ibarker r^j^yrraEEK LA KE, TRAVERSE CITY i W il l ia m s b u r g [N OR IE N Z O N IA AVER 22. J15 IN T E R L O C H E N T H O M P S O N V IL L E I iC O P E M lS H Figure 3.1: Map of Target Area 50 target region during the summer of 1987. The sampling structure of the study is given in Figure 3.2. Self-administered questionnaires including a map of the region were sent to potential travelers summer of 1987. (2,265) during the A total of 1,292 questionnaires (57%) were returned; 840 respondents had already made their trip before they received the questionnaires, so these respondents were excluded. The second wave of questionnaires was distributed to the remaining 452 respondents in December, 1987. of 381 questionnaires (84%) were returned. A total Among this group, a total of 107 respondents actually traveled to the target area. Finally, the third wave of questionnaires was sent to these 107 individuals in November, 1988, and 101 (95%) questionnaires were returned. For the purpose of the current study, the second sample of 107 questionnaires was cample contained the analyzed. neccooaiy This portion of the liiiuiiuctuiuji regcircixiig expectation and perceived performance about the target area based upon the first and the second questionnaires. Of course, secondary sufficiency there data: (Tull are problems availability, and Hawkins, associated relevance, 1984) to with accuracy, name a using and few. Availability refers to whether or not the data you are looking for is available. Relevance refers to the extent to which the data fit the information you need for the research problem. Accuracy is a real problem when the secondary data is not 51 TOTAL SAMPLE Pre Trip Measure 2265 Hon Response Pre Trip Return 1292 57X No Travel 135 10X Plan To Travel 452 35X Already Traveled 687 53X Post Trip I Measure Non Response 73 16X Post Trip I Response 381 84X Visited MI 172 45X Target Region 107 62X No Trip 207 54X MI, Not Target 65 38X Post Trip II Measure Returned 101 95X Non Response 6 5X Figure 3.2: Sampling Structure 52 taken directly from the original source (e.g., using second­ hand report). Sufficiency refers to whether or not secondary data meet all the data requirements for the research problem at hand. Fortunately, the data for this study were available and accurate because the data were from the original source of the Michigan Travel Perception Study. The data were also relevant included and sufficient since they necessary information for the current study and meet almost all the data requirements for the research problem of the study. However, the will data still had selected limitations which be discussed in Chapter IV. Sample Characteristics The first comparisons between selected socio-demographic characteristics of the research sample and the population in the north central census region are provided through 3.5. Indiana, in Table 3.1 The north central census region includes; Ohio, Illinois, Michigan, Wisconsin, Minnesota, Iowa, Missouri, North Dakota, South Dakota, Nebraska, and Kansas. The reason for comparing the sample to this census region is that potential tourists to the target area from Illinois, Indiana, Ohio, Wisconsin, and Minnesota made up 81.2% of the inquires about the target area (Fridgen and Hsieh, 1989). Although this study was designed to test the applicability of satisfaction models, and not necessarily to extrapolate the findings to the population, the comparisons provide a basis to 53 Table 3.1: Age Characteristics Age 20 25 35 45 55 65 Sample - 24 - 34 - 44 - 54 - 64 and older N % 4 32 31 16 13 10 3.8 30.2 29.2 15.1 12.3 9.4 Regional population N 5,565 9,441 6,472 5,873 5,609 6,683 % 14.0 23.8 16.3 14.8 14.1 16.9 Source: U.S. Department of Commerce. (1983). 1980 Census of Population: General Social and Economic Characteristics (PC801-C1). Note. Valid cases are 106, and mean is 42.0 and S.D. is 13.1 for research sample. Regional population is in thousands. Table 3.2: Marital status Marital status Married Single Divorced/widowed/separated Sample N % 84 17 6 78.5 15.9 5.6 Regional population N 16,661 11,013 5,296 % 50.5 33.4 16.1 Source: U.S. Department of Commerce. (1983). 1980 Census of Population: General Social and Economic Characteristics (PC801-C1). Note. Valid cases are 107 population is in thousands. for research sample. Regional 54 Table 3.3: Gender Gender Sample N Male Female 47 60 Regional population N % 43.9 56.1 28,614 30,251 % 48.6 51.4 Source: U.S. Department of Commerce. (1983). 1980 Census of Population: General Social and Economic Characteristics (PC801-C1). Note. Valid cases are 107 population is in thousands. for research sample. Regional Table 3.4: Education Education No high school diploma High school diploma Some college College degree Some graduate school Graduate degree Sample N % 3 23 35 23 6 17 2.8 21.5 32.7 21.5 5.6 15.9 Regional ]population N % 5,099 5,649 2,385 31.8 35.2 14.9 2,909 18.1 Source: U.S. Department of Commerce. (1983). 1980 Census of Population: General Social and Economic Characteristics (PC801-C1). Note. Valid cases are 107 population is in thousands. for research sample. Regional 55 Table 3.5: Income Income Less than $10,000 $10,000 - $19,999 $20,000 - $29,000 $30,000 - $39,999 $40,000 - $49,999 $50,000 - $59,999 $60,000 - $69,999 $70,000 or more Sample N % 2 12 13 21 12 11 5 9 2.4 14.1 15.3 24.7 14.1 12.9 5.9 10.6 Regional population N % 2,753 4,447 17.8 28.9 7,374 47.8 844 5.5 Source: U.S. Department of Commerce. (1983). 1980 Census of Population: General Social and Economic Characteristics (PC801-C1). Note. Valid cases are 85 for population is in thousands. research sample. Regional 56 determine the representativeness of the sample. With respect to age distribution (Table 3.1), the sample appropriately represented the age groups of 45-54 and 55-64. However, the youngest and oldest age groups of 20-24 and 65 and older were underrepresented while the middle age groups of 25-34 and 35-44 were overrepresented. Marital status category was is given overrepresented in Table (about 80%) 3.2. The while married single and divorced/widowed/separated categories were underrepresented. Single people by themselves do not seem to travel to a resort area like the target complete the survey. area or, if they did, they did not Gender distributions were adequately represented (Table 3.3). in terms categories of of underrepresented, higher were education high school completed diploma (Table or 3.4), lower the were while the categories of some college or overrepresented. xc is noted that about 75 percent of the research sample received at least some college education. Considering income (Table 3.5), the sample appropriately represented income categories of $20,000-$49,999. underrepresented the lower income categories of However, it less than $20,000 while it overrepresented the higher income categories of greater than $50,000. This can be explained by higher education characteristics of the sample, i.e., people who have completed higher education are likely to earn more income than 57 others. In summary, individuals with the sample high in education the study and were high primarily income Almost eighty percent of the sample was married, levels. and about seventy-five percent of the sample was middle aged. The distributional differences between the research sample and the census region population could partially be accounted for by the following factors. First, the reporting years between the two distributions were different. The data of the research sample were collected during the summer of 1987 and the early winter of 1988, whereas the data of the census region population were obtained in 1980. Second, the research sample was not randomly selected from the population in the north central census region. Since the sample were potential travelers who called the Michigan Travel Bureau to inquire about travel and tourism information of the target area interest in travel in Michigan, and tourism. these individuals had an The distribution of the sample might be reflected by the characteristics of those who were interested in tourism activities. this possibility, additional In order to examine comparisons were made with domestic travelers to and through Michigan (Tables 3.6 through 3.10). Age distribution is provided in Table 3.6. In general, the sample appropriately represented all age group categories with the exception of the two age groups of 20-24 and 35-44. 58 Table 3.6 Age Characteristics Sample Age 20 25 35 45 55 65 - 24 - 34 - 44 - 54 - 64 and older Michigan travelers N % % 4 32 31 16 13 10 3.8 30.2 29.2 15.1 12.3 9.4 13.0 31.0 21.0 14.0 14.0 8.0 Source: U.S. Travel Data Center. 1985. Travel To and Through Michigan 1983-84. Special study prepared for the Michigan Travel Bureau. Washington, D.C.: U.S. Travel Data Center. Note. Valid cases are 106, and mean is 42.0 and S.D. is 13.1 for research sample. Table 3.7: Marital status Marital status Married Single D ivorced/widowed/separated Sample Michigan travelers N % % 84 17 6 78.5 15.9 5.6 70.0 20.0 10.0 Source: U.S. Travel Data Center. 1985. Travel To and Through Michigan 1983-84. Special study prepared for the Michigan Travel Bureau. Washington, D.C.: U.S. Travel Data Center. Note. Valid cases are 107 for research sample. 59 Table 3.8: Gender Gender Male Female Sample Michigan travelers N % % 47 60 43.9 56.1 45.0 55.0 Source: U.S. Travel Data Center. 1985. Travel To and Through Michigan 1983-84. Special study prepared for the Michigan Travel Bureau. Washington, D.C.: U.S. Travel Data Center. Note. Valid cases are 107 for research sample. Table 3.9 Education Education No high school diploma High school diploma Some college College degree Some graduate school Graduate degree Sample N % 3 23 35 23 6 17 2.8 21.5 32.7 21.5 5.6 15.9 Michigan travelers % r* o O • \J 28.0 22.0 25.0 17.0 Source: U.S. Travel Data Center. 1985. Travel To and Through Michigan 1983-84. Special study prepared for the Michigan Travel Bureau. Washington, D.C.: U.S. Travel Data Center. Note. Valid cases are 107 for research sample. 60 Table 3.10: Income Income Less than $10,000 $10,000 - $19,999 $20,000 - $29,000 $30,000 - $39,999 $40,000 - $49,999 $50,000 - $59,999 $60,000 - $69,999 $70,000 or more Sample Michigan travelers N % % 2 12 13 21 12 11 5 9 2.4 14.1 15.3 24.7 14.1 12.9 5.9 10.6 9.0 26.0 27.0 20.0 17.0 Source: U.S. Travel Data Center. 1985. Travel To and Throuoh Michiaan 1983-84. Special study prepared for the Michigan Travel Bureau. Washington, D.C.: U.S. Travel Data Center. Note. Valid cases are 85 for research sample. 61 The youngest age group, 20-24, was underrepresented because the sample did not include any respondents 18 and 19 years of age. The middle age group of 35-44 was However, overrepresented. this suggests that the research sample is better representative of Michigan travelers than it is of the regional population as a whole. With respect to marital status (Table 3.7), the sample a little overrepresented those married while the single and divorced/widowed/separated categories were slightly underestimated. For gender (Table 3.8), the sample distribution was almost the same as the distribution of Michigan travelers. Thus, the research sample was very similar to these Michigan travelers. As for education school diploma or (Table 3.9), lower were the categories of high underrepresented, while the categories: of some college or higher were overrepresented. However, these deviations are much less pronounced than those observed for the regional population comparison. In terms of income (Table 3.10), the sample appropriately represented income categories of $20,000-$39,999. The lower income categories of less than $20,000 were underrepresented whereas the higher income categories of greater than $40,000 were overrepresented. The differences in categories between the sample and the Michigan travelers are less than those in the same categories observed between the sample and the 62 regional population census data. Therefore, the sample used in this study was more comparable to travelers to and through Michigan than to the regional population as a whole. This implies that Michigan tourists are likely to be older, married, have more than high school education, and have a relatively high income. These findings are to be expected, in light of the findings of many studies which reveal that the modern day traveler is older, well situated in terms of jobs, education and income, and usually is a family person. Research Variables and Constructs Definition. Description, and Measurement In this measurement of section the definitions, research introduced and discussed. parts: (1) tourism variables descriptions, and and are constructs The section is divided into two attributes variables and (2) tourism constructs. Tourism Attributes Variables Tourism attributes may be defined as tourism-related physical, social, and behavioral characteristics that tourists would see, observe, destination area. feel, and experience while in a Based on the review of previous tourism attribute related studies (Gearing et al, 1974; Var et al, 1977; Pizam et al, 1978; Klenosky, 1985; Fridgen, 1987), the following 27 tourism attributes variables were developed for 63 the original study (Fridgen, 1989) and were used in this study as well. fall These attributes are: colors, uncrowded, swimming, natural access to boating, scenery, attractions, the lakes, fishing, camping, shopping, resorts, friendly people, pleasant climate, peaceful beaches, charter fishing, restaurants, and quite, clean water, golf, hotels/motels, inexpensive prices, tennis, nightlife, family fun, historical sites, sightseeing, and festivals (Figure 3.3). These 27 tourism attribute variables were measured by the use of agreement scores based upon a five-point Likert-type scale with "strongly agree" to "strongly disagree" used as anchors. The original questionnaire is provided in Appendix A. These variables were measured before travelers went to the destination area and again after travel to the destination. The scores for the attributes, before and after, became data the for expectation scores and perceived performance ecoree. Tourism Constructs Tourism construct may be defined as an underlying common concept composed of a set of the tourism attributes associated with a destination area. The underlying common concept can be considered the same as a factor in an exploratory factor analysis; thus, a tourism construct could be treated in the same manner. The target area is well known as a resort area within the midwestern region of the United States. The area provides 64 Peaceful and quiet Uncrowdedness Friendly people Family fun Inexpensive prices Sightseeing Scenery Natural attractions Festivals Fall colors Historical sites Shopping Restaurants Hotels/motels Nightlife Camping Boating Fishing Charter fishing Resorts Golf Tennis E a s v^ ar:r:fas«; +-n l a l c e s Beaches Clean water Swimming Pleasant climate Figure 3.3 List of Tourism Attributes 65 many varied and quality restaurants, stores, motels and hotels for tourists. It also has great outdoor recreation opportunities like boating, fishing, and camping. Moreover, the natural beauty of the area attracts travelers to a good sightseeing spot. Therefore, these characteristics associated with resorts, hospitality, outdoor recreation, and sightseeing can be the basis for establishing tourism constructs for the target area. based The tourism constructs were logically formed on previous studies and experts' judgements. Five tourism constructs are proposed for this study based upon the 27 tourism attribute variables used originally, and a review of previous tourism constructs developed in related studies (Pizam, 1978; Klenosky, 1985). (1) The proposed constructs are: local environment and atmosphere, hospitality, 3.4). (2) sightseeing, (3) (4) outdoor recreation, and (5) resorts (Figure These tourism constructs will be examined regarding validity and reliability. The attributes included in the first construct, local environment and atmosphere includes the following: peaceful and quiet, uncrowdedness, friendly people, inexpensive prices. The includes: sightseeing, festivals, fall construct of colors, hospitality second construct scenery, includes: of natural and historical hotels/motels, and nightlife. family fun, sites. shopping, and sightseeing attractions, The third restaurants, The fourth construct of outdoor recreation includes: camping, boating, fishing, and charter 66 Local environment & atmosphere Peaceful and quiet Uncrowdiness Friendly people Family fun Inexpensive prices Sightseeing Sightseeing Scenery Natural attractions Festivals Fall colors Historical sites Hospitality Shopping Restaurants Hotels/motels Nightlife Outdoor recreation Camping Boating Fishing VAAUX U C 1 JU Resorts Resorts Golf Tennis Easy access to lakes Beaches Clean water Swimming Pleasant climate Figure 3.4 Proposed Tourism Constructs 67 fishing. golf, The last construct of resorts includes: tennis, easy access to lakes, beaches, resorts, clean water, swimming, and pleasant climate. Psychological Constructs Relevant To Satisfaction Models Based on the review of satisfaction psychological models are: and constructs relevant to (1) expectation, (4) satisfaction. literature, the two (2) perception, to disconfirmation research. In this study, in satisfaction (3) comparison, The third construct, identical four comparison, consumer is satisfaction the psychological constructs are defined as: Expectation -- the most likely performance of products and services that will be used and experienced in the future. In this study, perception of expectation the most is a tourist's likely performance of the attributes in the target area prior to travel. from one's available advertising, promotion, Perceived perception information or performance word -- an of actual performance of of the In this study, it tourism It derives area mouth individual's (e.g., exchanges). subjective of products and services based on a tourist's use and experience with purchase. subjective is a tourist's these after subjective perception of actual performance of the tourism attributes in the target area experienced while visiting the destination and reported after the trip. It experience in the target area. derives from one's tourism 68 Comparison -- an individual's judgement of the discrepancies between expectation and perceived performance of products and services. In this study, comparison is the numerical differences between a tourist's expectation score for each tourism attribute and perceived performance score for each tourism attribute with respect to the target area. Satisfaction -- an individual's subjective attitude toward products or services based on use and experience. this study, In satisfaction is a tourist's subjective attitude toward their overall experience while visiting the target area. In terms of measurement, expectation and perceived performance were measured as agreement scores by a five-point Likert scale ranging from disagree" for each of mentioned previously. traveling, while the "strongly 27 agree" tourism-related Expectation was perceived performance was traveling to the target area. to "strongly attributes measured measured before after Then, these 27 agreement scores for expectation and perceived performance were summed to make a construct score for expectation and perceived performance, respectively1. Strictly speaking, the Likert scale is not an interval scale, however, many studies in the social sciences have treated the scale as an interval scale. Previous studies likert scale can be summed up to form a single score for each individual (Tull and Hawkins, 1984). Additionally, previous studies of consumer satisfaction (Westbrook and Oliver, 1980) and tourism measurement (Maddox, 1985) reported high reliability when using the Likert scale. 69 of consumer satisfaction also treated the scale as interval. Thus, for this study, Likert scale is assumed to have a properties of an interval scale. Expectation score was formulated as: 27 i-i where E was an expectation score for each tourism attribute i. Perceived performance score was formulated as: 27 E pi i-l where P was a perception score for each tourism attribute i. The comparison was measured by subtracting the score of the perceived performance from that of the expectation. This construct is the only endogenous variable in the satisfaction model. Comparison score was formulated as: 27 E i-l 70 Satisfaction was measured by a five-point Likert-type scale for an overall satisfaction measure. Research Models There are two proposed satisfaction models for explaining tourist satisfaction: the discrepancy performance-based change model. model and the In the following section, both models are described and discussed. Discrepancy Model A cognitive model of a theory visually illustrates how the theory presumes the human mind works, the order of events, and the nature of interactions of cognitive functioning. The cognitive model of the discrepancy is demonstrated in Figure 3.5. Selected relationships can be explained as follows. When a person has high expectations and perceives or finds high performance, then 'performance meets satisfaction is attained. expectation, and However, if the same person with high expectation perceives either medium or low performance, they will be less dissatisfaction. expectations satisfied Similarly, and perceives or when low a even experience person performance, has here low again performance matches expectation, and there will be a feeling of satisfaction. expectations there will Yet, perceives be even if either more the same medium person or high satisfaction with low performance, experienced by the 71 EXPECTATION PERFORMANCE COMPARISON HIGH MATCH MEDIUM BELOW LOW — ► EXCEED HIGH HIGH SATISFACTION MEDIUM LO W HIGH MEDIUM MEDIUM — ► MATCH — ► MEDIUM LO W ! , / \ L O W J^> — ► BELOW — ► LOW EXCEED — ► HIGH HIGH MEDIUM LO W — ► MATCH — ► MEDIUM Figure 3.5: C ognitive Model o f D iscrepancy 72 observer. perception If of expectations high are performance, medium the and there observed is the performance exceeds expectation and this will lead to more satisfaction as well. If a person's perception of performance is low, then performance is below expectation and the observer will be less satisfied or even dissatisfied. If the person's perception of performance is medium where performance meets expectations, satisfaction should result. Let us apply the discrepancy model to tourism. Before travel, potential tourists hold subjective expectations about a destination — they have expectations about tourism attributes such as dining, shopping, natural resources, etc. based on their available sources of information. Then, during a visit to the destination and after the visit, they formulate their perception of the tourism destination based upon actual experiences with the area. If these perceptions surpass expectaticns, this positive uisurepdncy will increase their level of satisfaction. Conversely, if perceptions fall below their expectations, this negative discrepancy will decrease their level of the satisfaction and most likely will become dissatisfaction. expectations, medium levels If their perceptions match their this represents confirmation and will lead to of satisfaction. This model assumes that satisfaction is a function of perceived performance and its comparison with expectations. On the other hand, a path diagram illustrates causal 73 relationships between variables or constructs and forms a causal model; thus an arrow represents causality or a linkage between the two relationships variables among the in the variables diagram. or constructs If the are all linear, then the path diagram can be interpreted in a series of structural equations. Further explanation of this path analysis will be presented in Chapter IV. The cognitive model of discrepancy can be rewritten into the path diagram and this is depicted in Figure 3.6. This path diagram is adapted from the Figure discrepancy model of 2.1 in Chapter II. The applicability of the original discrepancy model to tourist satisfaction can thus be examined. discrepancy model, this path Similar to the original diagram assumes that the expectation construct has a direct impact on both comparison and perceived performance constructs and indirect impact on satisfaction construct. construct mediated thorough the comparison The perceived performance construct has a direct impact on both comparison and satisfaction constructs and an indirect impact on satisfaction construct mediated through the comparison construct. direct impacts The comparison construct, from both expectation and receiving perceived performance, has direct impact on satisfaction. Performance-Based Change Model The depicted medium, cognitive model of in 3.7. A or Figure low performance-based person expectations. can have This model change either assumes is high, that 74 EXPECTATION COMPARISON SATISFACTION PERCEIVED PERFORMANCE Figure 3.6: P ath D iagram of D iscrepancy 75 EXPECTATION PERFORMANCE HIGH HIGH —► MEDIUM — ► HIGH MEDIUM LO W —► LOW HIGH —► HIGH MEDIUM — ► MEDIUM MEDIUM LO W i | LOW SATISFACTION — ► ■««^-*« * n i u n | ---- p*- MEDIUM — ► LO W LOW ■ ■ 1 nr i v a n | MEDIUM LOW Figure 3.7: Cognitive Model of Perform ance-Based Change 76 regardless of a person's initial attitude (i.e., expectation in this study), that person will react in the same direction as the result message (i.e., in attitude an Therefore, if performance, a perceived change person performance), (i.e., perceives which will satisfaction). that there is high a high degree of satisfaction will result; if perceptions are medium, a modest amount of satisfaction will result; if perceptions are low, then there will be less satisfaction or dissatisfaction. Let us apply the tourist satisfaction. performance-based change model to The first few steps of tourist behavior are the same as in the discrepancy model, i.e. the potential tourists hold their expectation of the destination area before travel and they form subjective perceived performance opinions of the area based on their actual experience when they travel to the area. expectation However, and the performance does explaining tourist satisfaction. expectations, comparison play not Rather, between a key their role in in spite of their only perceived performance of the destination area influences their level of satisfaction. Therefore, the higher the perceived performance of the area the higher the level of overall satisfaction will be. perceived performance will result In contrast, in lower lower levels of satisfaction. This model assumes that satisfaction is a function of perceived performance only. That is, the level of 77 satisfaction is directly influenced only by one's perception of performance. However, expectations have an indirect impact on satisfaction through perceived performance. This causal process is illustrated in a path diagram of performance-based change (Figure 3.8). The path diagram is adapted from the original performance-based change model seen in Figure 2.2. 78 EXPECTATION SATISFACTION PERCEIVED PERFORMANCE Figure 3.8: P ath D iagram o f Perform ance-B ased C hange CHAPTER IV ANALYSES TECHNIQUES AND PROCEDURES In this chapter, the procedures for the analyses used in this study are presented. The limitations of the study are also discussed.The chapter is divided into three parts: data analyses techniques, (2) procedural steps (1) of the analyses, and (3) limitations of the study. Data Analysis Techniques In any study, the researcher first must measure the research variables before analyzing the relationship between these variables. However, observed measurements are never perfect; error of measurement is always present. Thus, in the first section the theory of error of measurement is discussed and means for correcting it are introduced. In order to meet the study objectives, two statistical analysis techniques were employed: Both path analysis analyses are and briefly confirmatory described following section. 79 factor analysis. and discussed in the 80 Error of Measurement and Correction for Attenuation Measuring the sciences true is difficult value of variables in the social since the observed measurements indirect estimates of the constructs to be measured. the late 1890s, it has been known that the error of measurement systematically Since error measurement attenuate the correlation coefficient. words, are of In other lowers the correlation between measures in comparison to the correlation between the variables themselves (Hunter and Schmidt, 1990). Thus, observed correlation of the measures should not be used to analyze the relationship among the variables unless these correlations are corrected for attenuation. Let us denote T and U as the true scores that would have been observed on the independent and dependent variables using corresponding perfect measurements, respectively. observed variables x and y are: x-T+E1 y-U+E2 where E, and E2 are the error of measurement in the independent and dependent variables, respectively. The reliabilities of x (ra a ) and y (ryy ) are defined as: rxx~PxT Then 81 where T and U are the true scores; rho^. represents the population correlation between perfectly measured variable i and j . The observed correlation of x and y (ryJ *y includes: the systematic attenuation of the population correlation (pTU) by error of measurement and the unsystematic variation produced by sampling error. The systematic attenuation can be calculated by considering the causal paths from x to T to U to y (Figure 4.1): Thus, at the level of population correlations the systematic attenuation can be corrected as follows: P This is called the classic formula for attenuation or the Spearman-Brown formula. correction for However, at the level of observed correlations, a correlation (ryu) includes xy the true score of the correlation (pxy) and sampling error (e) : rxy-Pxy+e 82 TU Tx Uy y T=True score of construct T U=True score of construct U x=Observed score of construct T y=Observed score of construct U R=Correlation coefficient p=Path coefficient (In this case, correlation coefficient) Figure 4 .1: M easurem ent Model of T rue Score and O bserved Value 83 The unsystematic variation caused by sampling error cannot be corrected, but it can be expressed as a confidence interval. If the observed correlation is corrected using the above formula, the following equation yields: r c r*y I-- 1 i"■ y^xx y^ yy _ /^xxVryyP ru+e __ j y r x x y x yy #'" e r ii/ .. i y^xxy^yy where rc represented corrected correlations between observed variables x and y. Thus, the first term of the above equation clarifies that observed correlation can be corrected for the attenuation. However, the corrected correlation still includes the sampling error which is the last term of the above equation. Since the sampling error can be expressed by the confidence interval, we must apply the above correction formula (i.e., the last term) to both lower and upper endpoints of the confidence interval for the confidence instance, uncorrected interval the 95% correlation for a in order corrected confidence interval to form correlation. for the the For observed correlation of 0.20 between variable x and y with standard error1 of 0.096 is given by 0.20 ± 1.96 population correlation < 0.39. (0.096) or 0.01 < Assume that variables x and.y have reliabilities of 0.45 and 0.55, respectively. 1Standard error is a standard deviation estimate (i.e., correlation in this case). In order of parameter 84 to correct the confidence interval, the Spearman-Brown formula must be applied to each endpoint of the above confidence interval. Thus, the corrected confidence intervals for both endpoints are calculated as follows: Lower endpoint: ■■ 0,01 - 0 •01 -0. 02 VO.45 VO.55 °-50 Upper endpoint: 0 ■39 _ Oj.39 _q 7 g VO.45 VO.55 °-50 Therefore, the corrected and uncorrected confidence intervals can be compared as follows: Corrected: 0.02