INFORMATION TO USERS This manuscript has been reproduced from the microfilm master. UMI 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 submitted. 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 UMI a complete manuscript 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 materials (e.g., maps, drawings, charts) are reproduced by sectioning the original, beginning at the upper left-hand comer 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. Higher 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 UMI directly to order. A Bell & H owell Information C o m p an y 300 N orth Z e e b R o a d . A nn Arbor. Ml 4 8 1 0 6 -1 3 4 6 USA 3 1 3 /7 6 1 -4 7 0 0 8 0 0 /5 2 1 -0 6 0 0 EVALUATION O F AN INQUIRY PROJECT AT FRANKENM UTH , M ICHIGAN By Chao-Cheng W ang A DISSERTATION Submitted to M ichigan State University in partial fulfillment o f the requirements for the degree o f DOCTOR OF PHILOSOPHY Departm ent o f Park, Recreation, and Tourism Resources 1996 UMI Number: 9 631356 UMI Microform 9631356 Copyright 1996, by UMI Company. All rights reserved. This microform edition is protected against unauthorized copying under Title 17, United States Code. UMI 300 North Zeeb Road Ann Arbor, MI 48103 ABSTRACT EVALUATION OF AN INQUIRY PROJECT AT FRANKENM UTH, MICHIGAN By Chao-Cheng Wang Expansion o f a tourism market depends on tourism advertising. A brochure is one o f many tourism advertising tools designed to attract potential visitors' attention and to generate additional destination visits. Evaluation o f the effectiveness and efficiency o f brochure advertising is essential in understanding the performance o f a brochure. Since advertising budgets are limited, the development o f forecasting models is essential for cost effective advertising to potential tourists. Frankenmuth, Michigan was selected as the focus for this research. The objectives o f this study were to (1) identify factors that influence different tourist market segment's decision to make trips to Frankenmuth after requesting information from the Frankenmuth Chamber o f Commerce and Visitor Bureau (FCCVB) and (2) develop models that estimate the probability that different market segments will visit Frankenmuth. Information useful for developing marketing and/or advertising strategies was obtained through a cross-section mailed survey which resulted in 595 usable questionnaires for: 1) descriptive analyses that produced profiles o f different market segments, 2) testing for significant differences between segments, and 3) developing models that predict the propensity o f different segments to visit Frankenmuth. Stepwise logistic regression analyses (LRA) were applied to develop two models to predict the propensity to visit Frankenmuth. Model 1 results indicate that travel distance to Frankenmuth, in-state vs out-of-statc, perceptions o f the quality o f the brochure received in response to inquiries, elapsed time between the inquiry and receipt o f the information, and level o f education most influenced whether an inquiry was followed by a visit. The overall correct prediction rate for Model 1 is 68.32%. Model 2 was developed to predict the probability o f repeat visits to Frankenmuth The results indicate that the likelihood o f a repeat visit is influenced by travel distance to Frankenmuth, satisfaction with previous Frankenmuth (travel) experiences, and ranking o f Frankenmuth as a travel destination. The overall correct prediction rate for Model 2 is 78.13%. In conclusion, this study presents operative approaches to predicting visits to Frankenmuth among inquiries to FCCVB and revisits to Frankenmuth among current visitors. The major conclusion drawn form the results o f this study is that, with only a m odest additional investment, conversion studies ( a standard approach to evaluating advertising effectiveness) can yield for more useful marketing information than is commonly obtained. ACKNOWLEDGMENTS I would like to thank Dr. Donald Holecek, my academic advisor and Chairperson o f my doctoral dissertation for his guidance, insightful comments, and understanding. W orking with him contributed greatly to my knowledge, development and understanding o f tourism marketing model development. Special thanks are also extended to Dr. Edward M. Mahoney for his thoughtful and technical counseling, particularly during the final phases o f the study. I would also like to thank the other members o f my committee, Dr. Joseph D. Fridgen and Dr. Bonnie J. Knutson, they were most helpful and supportive o f this dissertation. Special thanks are given to my friend Prof. Gaylan A. Rasmussen who has supported and encouraged me in a very personal way. I would also like to express my gratitude to Dr. Daniel M. Spotts for his willingness to share his research experiences with me. Special recognition is extended to my loving parents, Chen-Ho and Shu-Yuan W ang for their love and unwavering faith in my entire student life; without their understanding, patience, and gentle pushes, I could not have done any o f this. Finally, I would like to express my deepest gratitude and thanks to my best friend and wife, Hui-Chuan. Thank you for your support and many sacrifices during the crazy times and for your encouragement always. T A B L E O F C O N T EN T S CHAPTER I. Page IN T R O D U C T IO N .......................................................................................... 1 Objectives o f Inquiry Conversion Research ........................................ Conversion Study M ethods and Limitations ........................................ Statement o f Problem ............................................................................ Study Site ................................................................................................ Study Objectives and Hypotheses .......................................................... Objectives ................................................................................................ Hypotheses ................................................................................................. Definition o f Terms ................................................................................... 3 5 8 9 11 12 13 13 II. L IT E R A T U R E R E V IE W .......................................................................... 15 Tourism Definitions .................................................................................. M arketing Segmentation in Tourism ..................................................... Importance o f Brochure Use in Tourism .............................................. Consumer Behavior in Tourism ............................................................. Information Search in Tourism .............................................................. Importance o f Repeat Visitors to the Tourism Industry .................... Influence o f Socioeconomic Factors on Tourism and Travel ..................................................................... Logistic Regression: Overview ................................................................ Linear Regression Model: Function and Violations ............................ Logistic Regression: Function and Modeling ....................................... Comparison between Discriminant Analysis and Logistic Regression ............................................................. 42 III. M E T H O D S ...................................................................................................... Sampling Design ......................................................................................... Review o f the Questionnaire D esig n ........................................................ Data Collection ........................................................................................... Data Preparation ........................................................................................ 45 46 48 49 53 v 18 20 23 25 29 30 32 34 34 37 CHAPTER Page ...................................................................................... 56 Data Analysis ............................................................................................... 57 Descriptive Statistics ................................................................... Statistical Significance Tests ....................................................... Developing a Prediction Function with Logistic Regression Analysis ........................................ 59 59 Variable Selection 62 IV. FIN D IN G S AND R ESU L TS ..................................................................... 64 Characteristics o f Visitors vs. Non-Visitors ........................................ First-time Visitors and Repeat Visitors ................................................. 64 70 Test o f Hypotheses .................................................................................... 81 Hypothesis 1 Hypothesis 2 .................................................................................. 81 86 Results o f Logistic Regression in Prediction ....................................... 97 Forward Stepwise Procedure .................................................... Goodness-of-Fit in Logistic Regression ................................. Testing the Significance o f Parameters in Logistic Regression .......................................................... Interpreting the Estimated Parameters ..................................... 98 100 105 108 Model 1: Prediction o f the Propensity to Visit Frankenmuth ................................................................... 110 Model 2: Prediction o f a Repeat Visit to Frankenmuth .............................................................................. 115 Sub-model 2.1: Prediction o f Re-visit to Frankenmuth Using Sociodemographic Variables ................................................. Sub-model 2.2: Prediction o f Re-visit to Frankenmuth Using Brochure Variables ................................................................ Sub-model 2 .3: Prediction o f Re-visit to Frankenmuth Using Satisfaction Variables .............................................................. vi 1 17 119 122 CHAPTER Page Sub-model 2.4: Prediction o f Re-visit to Frankenmuth Using Travel Behavior ................................ Model 2: Full Model Derived from Sub-model 2.1 - 2.4 Using Sociodemograpic, Brochure Evaluation, Trip Satisfaction and Travel Behavior Variables ..................... 126 V. SU M M A R Y , L IM IT A T IO N S AND IM P L IC A T IO N S ............................................................................... 131 Study Limitations .......................................................................... Implications ................................................................................... Future Research ............................................................................ 134 136 138 L IS T O F R E F E R E N C E S A PP E N D IC E S 124 .................................................................................. 140 ....................................................................................................... 153 Appendix A. Sample o f Frankenmuth Inquirer's Questionnaire ........................ B. Results o f the Survey o f Frankenmuth Inquirers .......................... vii 153 161 L IST O F TA BLES ige 1994-1995 Projected State Advertising Budget by Rank 2 Elapsed Time between Inquiry and Receipt o f Information by Visitors and Nonvisitors .................. 66 Distance between Hom e and Travel Destination by Visitors and Nonvisitors ......................................... 67 Demographic and Socioeconomic Characteristics o f Sample ........................................... 68 State/Province o f Residence o f Respondents: Visitors vs. Nonvisitors ............................................... 70 State/Province o f Residence o f Respondents: Repeat vs. First-time Visitors .................................... 72 Sources o f Information Used by Repeat and First-time Visitors ......................................................... 73 Accommodations Used Repeat and First-time Visitors 75 Party Size, Length o f Stay, and Makeup o f Party by Repeat and First-time Visitors ............... 76 Four M easurements o f Satisfaction with Frankenmuth: Repeat vs. First-time Visitors ..................................... 78 Demographic and Socioeconomic Characteristics o f Repeat and First-time Visitors .............................. 80 Significance Test Results for Selected Variables: Visitors vs. Non-visitor ................................................ 82 vui Table Page 13. Significance Test Results for Selected Variables: Repeat vs. First-time Visitors ....................................................................... 14 15 Logistic Regression Procedure with Dependent Variable Visit or not (Full Model) ............................................................... 103 Logistic Regression Procedure with Dependent Variable Visit or not (Reduced M odel) ........................................................ 106 16. Logistic Regression Procedure with Dependent Variable Visit or not (Goodness-of-fit Reduce M odel) .......................................... 17 87 107 Logistic Regression Procedure with Dependent Variable "Repeat Visit or not" for Sociodemographic Variables (Goodness-of-fit Reduced Model) ............................................. 119 Logistic Regression Procedure with Dependent Variable "Repeat Visit or not" for Brochure Evaluation Variables (Goodness-of-fit Reduced Model) ........................................... 122 19. Logistic Regression Procedure with Dependent Variable "Repeat Visit or not" for Satisfaction Variables (Goodness-of-fit Reduced Model) ............................................................ 124 20. Logistic Regression Procedure with Dependent Variable "Repeat Visit or not" Using Travel Behavior Variables (Goodness-of-fit Reduced M odel) ........................................................... 126 18 21 Logistic Regression Procedure with Dependent Variable "Repeat Visit or not" — Full Model (Goodness-of-fit Reduced M odel) ............................................................. IX 129 LIST O F FIG U R ES Figure 1. Location o f Frankenmuth, Michigan Page ..................................................................... 10 2. Three-dimensional Opportunity Set Matrix in Holiday (Destination) Choice ............................................................................................................. 16 3. The Logistic Function: Sigmoid or S-Shaped Model ....................................... 41 ...................................................................... 50 4. Flowchart o f Questionnaire Design x CHAPTER I IN T R O D U C T IO N For both domestic and international tourism, the consumption o f tourism exceeds US $2 trillion which represents about 12% o f the world's economy (W aters, 1988). According to Holecek (1993), in 1989 the tourism industry provided more than 100 million jobs, generated $166 billion in tax revenues and accounted for 7% o f the world's exports. Both suppliers and promotional organizations in the tourism industry spend significant amounts o f money on different types o f promotion to increase destination travel and to capture its economic benefits. These expenditures are directed at "advertising, personal selling, public relations, publicity, and sales promotions such as trade shows, point o f purchase, and store displays" (M cIntosh and Goeldner, 1990). Because o f the grow ing economic importance o f tourism, state governments have increased tourism prom otion spending (Kreisman, 1982). For example, the U. S.. Travel Data Center reported that the 1994-95 advertising budgets for 44 states totaled $114,134,154, an average o f $2,593,981 per state (See Table 1). Their advertising budgets accounted for 29% o f their total budgets ($399,152,053) for fiscal year 1994-95 (U. S. Travel Data Center, 1995) M uch tourism related advertising research has been aimed at determining the effectiveness o f promotional campaigns in converting inquirers into visitors (Burke & 1 2 Tabic 1. 1994-1995 Projected State Advertising Budget by Rank Rank State Money Rank State Money 1 2 3 4 5 HI TX IL AR VA $16,384,000 7,600,000 6,200,000 5,414,215 5,406,984 26 27 28 29 30 NM RI NV UT VT 1,350,000 1,330,000 1,284,700 1,260,000 1,188,477 6 7 8 9 10 LA SC WI MO FL 5,205,000 4,376,498 4,100,000 3,754,400 3,608,700 31 32 33 34 35 MN AL OH ND MD 1,099,014 990,000 977,575 837,739 828,420 11 12 13 14 15 MA MI CT CA MS 3,449,720 2,976,532 2,737,250 2,705,000 2,689,850 36 37 38 39 40 IN NH ID IA ME 779,937 772,525 750,000 622,500 573,750 16 17 18 19 20 AZ OK MT TN WV 2,650,000 2,563,097 2,517,777 2,513,900 2,260,000 41 42 43 44 WA OR NE KS 461,855 450,000 430,000 230,000 21 22 23 24 25 PA AK KY SD WY 2,080,000 2,033,300 1,777,000 1,462,661 1,452,778 Grand Total Average Source: U. S. Travel Data Cnctcr $114,135,154 $2,593,981 3 Gitelson, 1990; M anfredo et al., 1992) Inquiry conversion research measures the impact o f advertising in the tourism industry According to the U.S. Travel Data Center, approximately "22 million inquiries were received by all o f the states during 1993, or 464,550 inquiries per state" (U. S. Travel Data Center, 1995). Inquiry conversion research is used to evaluate the perform ance o f promotional programs. This research is based on information gathered in response to inquiries generated as a result o f advertisements placed in various media (W oodside & Reed, 1974; W oodside & Ronkainen, 1982; Perdue, 1985). Inquiry conversion research relies upon the direct action o f inquirers who must return reader service cards, call a toll-free telephone number, or take similar action to obtain tourism information. Several months after their inquires are received, a survey o f inquirers is typically conducted to determine whether they visited the destination from which they received information. A sampling o f the total number o f inquiries is used to determine the pattern o f visitation The overall goal o f conversion studies is to provide information that can be used to guide the tourism agencies or managers in planning and designing strategies for future advertising and promotion efforts that will yield maximum returns on investments in them. Objectives of Inquiry Conversion Research The primary objective o f inquiry conversion research is to evaluate the performance o f advertising campaigns and to determine the return on investment from advertising. Inquiry conversion research data are usually used to: (1) estimate the number 4 o f inquirers who have visited the destination as well as those who plan to visit in the near future; (2) understand the socioeconomic, demographic and psychographic characteristics o f the inquirers and those who visit after inquiry; (3) evaluate the effectiveness o f different media ads and advertisements; (4) assess the influence o f media information on decision making by inquirers; (5) calculate the amount o f revenue and spending generated from visitors to a destination area; and (6) determine the return on investment (ROI) for the advertising campaign. Advertising continues to be an important factor in vacationers' awareness o f destination, image o f destination, visitation to destination, and informing visitors o f changes in the tourism product at the destination (Ellerbrock, 1981; W oodside 1981; Hunt and Dalton, 1983; Ballmen et al, 1984; W oodside and Ronkainen 1984; M ok, 1990; W oodside, 1990; Burke and Gitelson 1990; Perdue and Pitegoff, 1990; Siegel and Ziff-Levine, 1990). Conversion research is often used by state and local tourism associations to determine the percentage o f destination inquirers that actually make a visit/trip. The data are used to determine the effectiveness in converting "inquiries" into "visitors". Conversion studies are conducted to answer the question o f how many inquirers, generated from travel ads, were converted to visitors and the converters' demographic and travel-behavior characteristics. This includes length o f stay, place o f lodging, party size, and destination expenditures. 5 C onversion Study M eth o d s and L im itatio n s It has been over two decades since the first publication o f an inquiry conversion study in the Journal o f Travel Research by W oodside and Reid in 1974 The authors used revenue per inquiry (RPI) to compare with cost per inquiry (CPI). For example, if the total number o f inquiries was 1000 and the proportion o f respondents who visited was 50%, the estimated total number o f parties would be 500 (e.g., 1000 x 50% = 500). Then, if the estimated total number o f parties is multiplied by an $100 average expenditure per party, the estimated total expenditures for inquiries would equal $50,000. The estimated total expenditures is then divided by the total number o f inquiries to produce the RPI (e.g., $50,000/1,000 = $50). With a CPI o f $10, the ratio o f return on investment is 4 (e.g., RPI-CPI/CPI = [$50 - $10] / $10 = 4). This means that a $1 advertising expenditure will produce $4 in visitor expenditures Many inquiry conversion studies have been published in advertising, marketing, and tourism journals (W oodside and Motes, 1981; Silberman and Klock, 1986; Ronkainen and W oodside, 1987; Davidson and W iethaupt, 1989; Burke and Lindblom, 1989; Perdue and Pitegoff, 1990; W oodside and Soni, 1990; Burke and Gitelson, 1990; Perdue and Gustke, 1992). While a few authors have expressed concern about the traditional methods o f conducting conversion research and have offered suggestions for their improvement, little has been done to comparatively test methods or validate suggested improvements (Ellerbock, 1981; Woodside, 1981; Hunt and Dalton, 1983, W oodside and Ronkainen, 1984; and Ballman et al, 1984). 6 M ost advertising conversion research studies are inadequately designed and implemented to provide valid answers as to whether an advertising and marketing campaign generates "new" visitors and produces new income for the destination The above-mentioned studies identify common difficulties with conversion studies, such as, nonresponse bias caused by improper sampling techniques, sampling imprecision, and failure to account for individuals who decided to visit the destination prior to being exposed to the advertising (W oodside and Ronkainen, 1984; Silberman and Klock 1986). R ather than identifying a specific causal relationship, conversion studies may be limited to making descriptive statements about the relationship between the advertisement and the search for information. According to Burke et al. (1984), " w hat caused decisions to be made and how much can be credited to the advertising campaign are separate and highly complex issues". To measure the effectiveness o f advertising, a well designed true experiment method is necessary (Woodside, 1990). The results o f a typical conversion study may be misleading and produce inflated estimates o f the return on an advertising investment. Thirty-one states performed advertising effectiveness/ conversion research studies during 1993-94 (U. S. Travel D ata Center, 1995). The difficulty in measuring the effectiveness o f tourism advertising campaigns is the tenuous causal link between prom otions and tourist behavior. Many persons requesting information from state and local tourism associations have already decided to visit the area before receiving the promotional materials (W oodside & Ronkainen, 1994) They are looking for help in planning "what to do", not "where to go". 7 Furthermore, many inquirers who request information are return visitors Therefore, the conversion rate for initial visitation may be overestimated (Ballman et al., 1984). For example, Gitelson's study (1986) showed that two-thirds o f the inquirers had visited a travel destination before requesting information. By not taking into account the full costs associated with advertising campaigns such as postage, phone, printing, and material handling /distribution, the majority o f conversion studies do not reflect actual costs and result in inflated estimates o f return on investment. In sum, conversion studies have attained wide use because they obtain managerially useful information at a reasonable cost. Perhaps their greatest advantage is in providing relative measures useful for examining trends or comparing different methods o f advertising (M anfredo et al., 1992). However, when applying a conversion study, the researcher needs to take the following into consideration: how to select a sample?, how to determine the precise sample size?, how to deal with nonresponse bias?, how to deal with recall bias in reported expenditures?, how to identify new visitors?, and how to measure the costs and return on investment? A more appropriate method and analysis is needed to determine the extent to which the information received by first-time visitors influence their choices o f travel destination, recreation participation, accommodation choices, length o f stay, and spending patterns. In addition, factors which influence destination choices such as motivations, preferences, image o f the destination, and the decision-making process should be investigated to extend understanding and knowledge o f tourists' behavior 8 Statement o f the Problem Does advertising increase visitation? Although research methods are available to test advertising's impact on sales, there are no scientific studies dealing with advertising's role in stimulating visitation (W oodside, 1990). A purpose o f tourism advertising is to attract potential visitors' attention and to generate additional destination visits. A brochure/booklet is one o f many advertising tools. The basic objective o f a brochure/ booklet is "to communicate a favorable impression o f the product advertised and the benefits it can offer" (Coltman, 1989). Evaluation o f the effectiveness and efficiency o f brochure/booklet advertising is essential in understanding the performance o f a brochure. Although conversion studies have been conducted by a number o f authors to evaluate the effectiveness o f promotion/advertising, the relationship between brochure/booklet advertising and sales is still difficult to assess precisely. There is no "one best" evaluation method that can be successfully applied to all products or promotions. The lack o f effective advertising evaluation increases the cost o f promotion/advertising. Furthermore, there is no research available which links the impact o f media and visitor vs. non-visitor characteristics and/or first-time visitors vs. repeat visitors. This information would assist managers or agencies in understanding the process o f destination choice by inquirers, allocating advertising budgets more efficiently, targeting the highest potential markets, and designing more effective and appropriate advertising campaigns. In summary, there is a need to identify interactions among different travel stimuli, such as prior experience at a destination, "word o f mouth" recommendations by friends and relatives, and other factors that may influence destination decisions. In addition, 9 further study is also needed to better understand relationships between independent variables such as the impact o f media, brochure design, characteristics o f inquirers, choice o f travel destination, and spending patterns in a destination area and the dependent variable travel decision. Study Site Frankenmuth, located off 1-75 between Flint and Saginaw, Michigan, was selected as the study site. Its geographic setting is depicted in Figure 1. Frankenmuth was founded in 1845 by a group o f fifteen German Lutheran missionaries who came to this area for the purpose o f teaching Christianity to the Chippewa Indians. "Franken" depicts the province from which the settlers came and "Muth" means courage in German. The name Frankenmuth means "Courage o f the Franconians." Today, Frankenmuth is a thriving community o f 4,400 residents who take great pride in preserving the German heritage. Area homes, businesses, and surrounding farms remain neat and clean reflecting thier German ethic. W ell-tended flowers and lush greenery abound in what many visitors describe as the most authentic Bavarian architecture to be found anywhere in the United States. Frankenmuth is famous for good food. Two o f the country's largest family restaurants, dating from 1856 and 1888, combine to serve nearly two million dinners each year. Bronner's Christmas Wonderland, the world's largest Christmas store, covering five acres under one roof, is also located in Frankenmuth. In addition, pretzel, cheese and sausage factories, woolen mills, leather shoes, fudge and candy kitchens, an art gallery and 10 Frankenmuth, A//' D etroit),^/ /^ >5-C lcvclan (l [A u tom ap R oad A tla s C op y r ig ht €> 1 9 8 8-1 995 M i c r o s o f t C orporation! Figure 1. Location o f Frankenmuth, Michigan 11 museums all combine with 100 quaint village gift shops and attractions to make Frankenmuth the most visited spot in Michigan with over three million visitors each year. The primary economic base o f Frankenmuth is in tourism and recreation activities. As early as 1906, community leaders were organizing to promote cooperation and tow n improvement. The local Board o f Trade grew into a Chamber o f Commerce that is acknowledged as "the largest and most active Chamber o f Commerce in the United States among communities with under 5,000 residents." (Frankenmuth Chamber o f Commerce and Convention & Visitors Bureau, 1994). In 1994, the Frankenmuth Chamber o f Commerce and Visitor Bureau (FCCVB) provided service to over 30,772 visitors inside the visitor information center, answered 29,590 telephone inquiries, and mailed over 15,810 information packets. For the 1994 promotion campaigns, the FCCVB spent over $33,000 on print advertising, $88,000 on radio promotion, and $6,000 on television advertising. The mission o f the FCCVB is "to promote Frankenmuth and the prosperity o f all businesses in the community through a unified organization." Therefore, effective and efficient promotion campaigns are crucial to achieving the FCCVB's mission. Study Objectives and Hypotheses The purpose o f this research is to enhance the understanding o f travel behavior with a special focus on identifying variables that influence the travel decision. One objective o f this study is to gain insight into who requests travel information and how this information is utilized in travel decision making. Another objective o f this study is to identify relationships among travel decisions, elapsed time between inquiry and receiving 12 information, elapsed time between receiving information and making a trip, brochure quality, and individual traveler characteristics including socioeconomic and demographic information. Finally, the Logistic Regression Analysis (LRA) model will be employed to identify variables which predict the propensity to travel to a destination In summary, this study not only provides the traditional conversion study results for a better understanding o f conversion based on the promotional literature, but also identifies the key variables influencing destination choices o f tourists in different market segments (e.g., visitors vs. non-visitors and first-time vs. repeat visitors). It provides better knowledge for the utilizing o f promotional tools, allocating budget resources, and predicting the probability o f future trips to the destination. Objectives In order to derive information which may eventually be used in the development o f marketing strategies and the prediction o f future visits to Frankenmuth, this study was designed around the following four objectives: Objective 1: To identify factors that influence the decision ( visitors vs. non-visitors) to visit Frankenmuth after requesting travel information from the Frankenmuth Chamber o f Commerce and Visitor Bureau (FCCVB). Objective 2: To provide a descriptive profile o f the characteristics o f first-time and repeat visitors to include: information sources used, travel behavior, and satisfaction with the travel experience. Objective 3 : Develop a model to predict the propensity to visit Frankenmuth among inquirers. Objective 4: Develop a model to predict the propensity for repeat visits to Frankenmuth among current visitors. n Hypotheses Null Hypothesis 1. There are no significant differences in destination-decision making between visitors and nonvisitors with respect to: familiarity with the destination, state/province o f residence, distance to the destination, readership o f advertising literature, elapsed time between inquiry and receipt o f information, interest in the advertising literature, quality o f the literature, and socioeconomic and demographic characteristics o f inquirers. Null Hypothesis 2: There are no significant differences in destination decision making between parties who are familiar with the travel destination and those who are first-time visitors in terms of: residence status, distance to the destination, medium/media used, travel behavior, travel satisfaction, brochure readership, on site brochure consultation, elapsed time, spending patterns, brochure quality; and socioeconomic and demographic characteristics o f inquirers. D efinition o f T erm s The terms used throughout this study are defined below. An inquirer: An individual who is not a resident o f Frankenmuth and who requested information from the Frankenmuth Chamber o f Commerce A visitor is an individual who made a trip to Frankenmuth after requesting information. A non-visitor is an individual who did not make a trip to Frankenmuth after requesting information A first-time visitor is an individual who had not visited Frankenmuth before requesting information and actually did visit Frankenmuth. 14 A repeat visitor is an individual who had visited Frankenmuth before requesting information. Quality o f brochure as measured by how inquirers rated the overall quality o f the brochure mailed to them by the Frankenmuth Chamber o f Commerce. Stimulation o f the brochure as measured by responses to the quiry "to what extent did the brochure influence your interest in visiting the destination." C H A P T E R II L IT E R A T U R E R E V IE W The decision to visit a particular destination can be seen as the individual's solution to the problem, "where should I go for my holiday?" For the tourist, the decision entails a series o f choices, including the budget for the holiday, the time available, who to travel with, and forecasts o f the benefits they are likely to experience at each possible destination. In general, destination purchases can be distinguished from other purchases by: (1) the interval o f time which elapses between purchase and the consumption o f a destination, (2) the high cost o f travel compared with most other purchases, and (3) the difficulty o f knowing what to expect in a distant, unfamiliar place (Laws, 1995) According to Dann (1981), the choice o f a destination is viewed as a process in which the various "pull" factors (or attributes) o f a destination are analyzed and compared with similar destinations on a competitive basis. In the same vein, a number o f authors have proposed the concept o f opportunity set that is presented in Figure 2. An opportunity set is defined as "destinations available at a particular time." The opportunity set includes: 1) a perceived opportunity set which includes all destinations known to the tourists, 2) an attainable opportunity set which depends on what tourists can afford and 3) a realizable opportunity set which combines both perceived and attainable opportunity sets together 15 16 PERCEIVED OPPORTUNITY SET ATTAINABLE OPPORTUNITY SET INCREASING PREFERENCE REALIZABLE OPPORTUNITY SET . Goal A CONSIDERATION SET Goal B CHOICE SET Goal C DECISION SET Goal D HOLIDAY CHOICE Figure 2. Three-dimensional opportunity set matrix in holiday (destination) choice (adapted from Kent, 1990) 17 and provides the knowledge to recognize the supply opportunities and constraints on access to destinations in the opportunity set The extent o f the realizable set differs among tourists. It serves to bring the number o f possible destinations down to a manageable level from which tourists can make an informed selection. This set is then further reduced through a screening process in which the traveler's goals are matched to her/his expectations with respect to each destination ultimately leading to a final decision set from which the ultimate destination is selected. Decision sets are likely to contain no more than seven choices, often fewer (M outinho, 1987; W oodside & Sherrell, 1977). The study o f destination choice as a cognitive process has important implications for the tourism industry in marketing its products. Destination choices derive from bundles o f attributes (e.g., destination information, destination accommodations, travel m ode..etc.) combining different benefits and costs to tourists. A tourist's awareness o f the destination, attitude toward the particular destination, and expectations o f the destination will influence her/his individual image o f the destination. To better understand the destination choice process, this literature review focuses upon some o f the factors which other researchers have found to influence the destination choice process. Literature relating to definitions o f tourism , market segmentation in tourism, the importance o f brochure use in tourism, information-seeking behavior in tourism, consumer behavior in tourism, the importance o f repeat visitors to the tourism industry, and influence o f socioeconomic factors on tourism and travel are first reviewed. Then, the concept o f logistic regression (LR), the linear regression model, function o f 18 logistic regression, and a comparison o f LR with other analytical tools (e.g., linear regression analysis and discriminant regression analysis) will be discussed and presented. Tourism Definitions Tourism is the largest business in the world. It is also a complex process comprised o f many parts and interconnections. Tourism involves not only the tourists and the process o f their spatial mobility but also the host community/destination. It also includes economic, psychological, social, cultural and other attributes. Because o f its complexity, tourism can not be precisely defined. For example, Przenclawski (1986) places tourism, based on the behavior o f the tourist, into one o f the following categories: cognitive tourism, recreational tourism, health tourism, creative tourism, educational tourism, professional tourism, pilgrimage tourism, family-bound tourism, sex tourism and profit-making tourism. Because tourism means many different things to different people, Wall (1992) suggested that the discussion o f tourism must be moved from a mono-phenomenon to an examination o f types o f tourism. He used tourism typology to explain tourism based on attraction type, location, spatial characteristics, and development status. Many disciplines have developed an interest in tourism; these include: psychology, sociology, anthropology, economics, marketing, ecology, political science, and planning, but no common definition has yet emerged. 19 Ryan (1991 p 6 ) defined tourism based on an economic activity point o f view as: "a study o f the dem and f o r and supply o f accomm odations and supportive services f o r those staying away fro m home, and the resultant patterns o f expenditure, incom e creation and employment This definition identifies two approaches to tourism research: 1). tourism as a scientific process to investigate the hypothetical relationships between causal and determined variables; and 2 ) tourism as a subset o f a business problem. McIntosh and Goeldner (1990, P 4)) summarize tourism as: "the sum o f the phenom ena and relationships arising fro m the interaction o f tourists, business suppliers, host governments, and host com m unities in the process o f attracting and hosting these tourists and other visitors . " The authors point out that tourism is a composite o f activities, services, and industries that deliver a travel experience and other hospitality services to those people, individuals or groups, traveling away from their residence. Fridgen (1990) indicated that tourism is behavior influenced by internal forces (e.g., attitudes, motives, perceptions, personality, learning, social and family role) and external forces (e.g., social class, culture, subculture, reference group, environmental conditions). In other words, tourism is a process o f decision-making influenced by different dimensions. These dimensions may be psychological, social and cultural, economic, and environmental. Mill and Morrison (1992) stated that tourism includes activities and impacts occurring when a tourist travels. Planning o f the trip, traveling to the place, staying in the 20 place, returning from the place, and post-trip memory are part o f the experience In addition, tourism includes all the interactions between hosts and guests in the destination community Mathieson and Wall (1982) suggested that the tourism industry is composed of three basic elements: ( 1) a dynamic element that involves travel to a selected destination, (2) a static element that involves a stay at the destination, and (3) a consequential element that is concerned with effects on the economic, physical, and social subsystems in which the tourist makes contact either directly or indirectly. Due to its heterogeneous nature, tourism has been defined in numerous ways. It is a complex, global activity and a significant global socioeconomic phenomenon supported by changes in lifestyle, higher incomes, higher levels o f education, and greatly enhanced mobility (Mill &.Morrison, 1992; U.S. Travel Data Center, 1991). No single definition can completely capture the tourism phenomenon. Thus, different tourism definitions are needed to enhance understanding o f tourism's multiple dimensions. Knowledge o f a broad range o f tourism definitions will help agencies deliver tourism opportunities, benefits, and experiences to the tourist and provide "the means o f transport, goods, services, accommodations and other facilities for travel out o f the home community for any purpose not related to local day-to-day activity" (U. S. Travel Data Center, 1987). Market Segmentation in Tourism Market segmentation is a technique used to divide a heterogeneous market into homogeneous sub-groups or market segments. It is based on the idea that a market is 21 composed o f subgroups o f people and that each o f these subgroups has different, specific needs and wants (Barnett, 1969) A number o f authors have suggested that market segmentation is a tool for enhancing understanding o f consumer behavior Through market segmentation, one can adjust a product or service and its price, promotion and distribution to meet the needs and wants o f distinct target segments (Wind, 1978; Stynes, 1985; Morrison, 1989). According to Hsieh and O'Leary (1993), market segmentation is a management strategy based on assumptions that among a population subgroup a homogeneous behavior exists within the larger heterogeneous behavior. Market segments have most frequently been based on demographics, geographic, and behavioral characteristics (Lewis and Chambers, 1989). Both demographics and geographic segmentation modes have a common limitation: they are largely nonproductive because they are "post hoc" The needs and wants motivating the behavior o f these tourists, however, are still largely unknown Thus, understanding tourists' attitudes and their behavioral characteristics can assist in developing marketing strategies focusing on a particular market segment. Market segmentation is an important element o f any tourism marketing strategy Much literature has been published in the tourism field based on the different approaches to market segmentation. For example, Woodside and Jacobs (1985) compared three nationality groups with respect to trip behavior, demographics, trip characteristics, accommodations, attractions visited, and benefits realized. Woodside et al. (1986) segmented the timeshare resort market into owners and non-owners. Perdue (1985) segmented travel information inquirers based on whether information was acquired before 22 or after the vacation destination choice was made Calantone and Johar (1984) found that pleasure travelers sought different benefits from their trips throughout different seasons. Spotts and Mahoney (1991) segmented visitors based on the amount o f money they spent in a destination region. Snepenger (1987) used the degree o f novelty sought by vacationers as a segmentation base. Similarly, Davis and Sternquist (1987) used travelers' judgments o f the availability and importance o f vacation destination attributes in a cluster-based segmentation analysis. In order to achieve an efficient use o f marketing resources, Kotler (1984) and Morrison (1989) suggested the following criteria for effective segmentation: 1. Measurable: It is inadvisable to pick target markets that can not be measured with a reasonable degree o f accuracy. 2. Accessible: The essence o f market segmentation is being able to select and reach specific customer groups 3 Substantial: A target market must be big enough to be worth considering for market segments 4. Defensible: The marketer must be sure that each group receives individual attention 5. Durable: Some market segments are short term or medium term, meaning they exist for less than five years. The prudent marketer should be convinced that each target market has long-term potential. 6. Competitive: The competitiveness o f this service is relative to a particular market segment The more precisely the service fits the needs o f a 23 particular segment, the more likely it is to succeed On the other hand, if a service does not match the needs well, there is little point in pursuing the segment Importance o f Brochure Use in Tourism Tourism marketing is made up o f a variety of activities designed to meet the needs o f travelers Promotion is recognized by many as a critical element o f tourism marketing The travel brochure is a most important and widely utilized element in destination promotion programs (Holloway & Plant, 1988). For years, travel brochures have been used by national, state, and local convention and visitors' bureaus as an inexpensive and adaptable communication medium Moss (1977) defined promotion through the use o f brochures as "a booklet or pamphlet used in sales solicitations or promotion activities." The author suggested that sales promotion materials must do more than just remind consumers o f the existence or value o f a service. Consumers are known to use brochures in making travel destination decisions Coltman (1989) commented that potential customers "will compare the brochure o f one destination or supplier with the brochure o f its competitors, and it is likely that the one with the best brochure will receive the business." In other words, brochures can convey the quality o f destinations to the potential customers. However, this is not always true. According to Wicks & Schuett (1991), an individual tourist conducts an internal and external information search to help fulfill his/her need to learn about the tourism product under consideration for purchase. An internal search is the recall o f information, such as memories o f a previous visit to the site or a similar destination An external search 24 is the process o f seeking new information through communication with others including the media or travel brochures from the destination. Gitelson and Crompton (1983) indicated that external searches are important in the tourism industry because a trip is a high risk purchase and involves the use o f discretionary money during a individual's free time. In addition to money and time, the tourist is unable to actually observe the potential purchase, especially if it is a new destination Therefore, brochures or secondary sources must be used. Etzel and Wahlers (1985) reported that destination-specific travel literature (DSTL) including booklets, brochures and pamphlets " is an attractive promotion alternative because it presumably affords the marketer a more accurate reach and a potentially greater impact than mass circulation broadcast media alternatives " They also note that "DSTL represents an opportunity to provide the prospect with a significantly greater number o f strong selling messages as compared to a magazine ad or a short radio or television commercial." Although external information is not used in all purchasing situations, the relative ease and low cost o f obtaining travel brochures makes them an attractive information source for many vacation decision makers. In addition, the travel brochure may serve as a long-term reminder or reference for those information-seekers who keep it. Furthermore, the brochure may be passed on to others, thus multiplying its effectiveness 25 Consumer Behavior In Tourism Tourism destination choice can be influenced by a number o f different factors Among the key factors believed to be involved are: a combination o f needs and desires, availability o f time and money, supply o f opportunities, image o f the destination, perceptions and expectations built on experiences and information gathered A number of authors have produced publications that stress actions involving a balance o f biological, social and psychological needs (Woodside & Sherrell, 1977; Parker, 1983; Kripperdorf, 1987; McIntosh & Goeldner, 1990; Um & Crompton, 1990; ). Furthermore, McIntosh and Goeldner (1990) explain that tourism can be ethnic, cultural, historical, environmental, recreational and business tourism. These approaches focus on outcome or activities rather than process. Relatively little is known about the motivation for the travel, the length o f stay, the on site activities or post-consumption benefits sought from travel (Middleton, 1988; Jefferson & Lickorish, 1988). Although Murphy (1985) provides a historical review o f the major trends in tourism by using motivation, ability and mobility to account for the growth o f tourism, his approach focuses much more on sociological changes than on motivational changes. A distinction should be made between the role o f motivations (push a tourist to make decision) and the attraction exercised by destination images (pull the tourist towards a particular destination). Breaking from work, escaping from a routine, or respite from everyday worries are examples o f motivation. The stimulus of new places and attractions available at destinations are examples o f destination images 26 To explain travel behavior, one must first understand travel motivations Understanding what motivates people to travel allows is to better define the values sought from travel. The term motive has been used to refer to internal forces and external goals and incentives that guide, direct, and integrate a person's behavior for future and potential satisfaction (Murray, 1964; Iso-Ahola, 1982; Pyo and Uysal, 1990). Therefore, motivation is an interpersonal phenomenon. This fact has positioned leisure travel as a psychological experience (Dann, 1977, 1981; Chon, 1989). Mill & Morrison (1985) indicate that motivation comes from people's needs and desires. Therefore, tourists must determine their preferences and set their goals to identify the most satisfactory option from among the current available travel destinations. On the one hand, needs are intrinsic, an innate condition arising from a lack o f something necessary to the individual's well-being. On the other hand, wants are extrinsic, a feeling that the individual would get pleasure or satisfaction from doing something. Together, needs and wants determine motivations. Kotler (1982) identifies three stages o f "need arousal" (p. 236). The first stage includes internal or external stimuli which trigger a leaning toward a particular product class. The second stage is activating existing needs that might be served by the product class The third stage is the specific wants that become activated by the recognized needs Iso-Ahola identified motivational forces for optimal stimulation and arousal (equilibrium) as being approach (seeking) or avoidance (escaping) in nature. He indicated that people pursue leisure activities for feelings o f mastery and competence or to leave the routine environment. He, therefore, suggested two motivational forces that become 27 determinants o f tourism behavior: (1) The desire to leave the everyday environment behind - escaping personal and /or interpersonal environments, and (2) The desire to obtain psychological or intrinsic rewards through travel to a contrasting environment - seeking personal and/or interpersonal intrinsic rewards. In addition to Iso-Ahola's social/psychological motivational forces, the concept o f push and pull factors has been examined by a number o f authors (Dann, 1977; Crompton, 1979; Pearce and Caltabiano, 1983; Yuan and McDonald, 1990). Push factors that have been identified include: socioeconomic variables, demographic variables, and attitudes, interests, and opinions that the traveler possesses, along with knowledge about the market (Smith, 1983). Specific push factors could include: age, gender , income, education, family structure and size, occupation, and other personal variables that influence the traveler's decision to travel. Pull factors are destination attributes that respond to and reinforce push factors or motivations. According to Smith (1983), destination attributes can either be tangible resources or the perceptions and expectations o f the traveler Plog (1974, 1990) used the Psychocentrism/Allocentrism approach to examine travel behavior and motivation. He classified the U.S. population along a psychographic continuum delineating personality types, ranging from the psychocentric at one extreme to the allocentric at the other. He found that psychocentrics tend to prefer familiar destinations, including areas that promote relaxation and low-level activities. On the other hand, allocentrics tend to be self-confident and enjoy discovery, new experiences, and new areas, and they like to travel to different or even exotic destinations 28 Moutinho (1987) used the concept o f social influence to explain travel decisions. He suggests that travel decisions are very much affected by forces outside the individual, including the influences o f other people. He divided social influence into four major areas: (1) role and family influence, (2) reference groups, (3) social classes, and (4) culture and subculture. Social influence has a tremendous impact on motivating an individual to travel. Therefore, this influence on travel decision behavior and travel motivation is extremely important in understanding the overall tourism industry Hill et al. (1990) studied the relationship between motivation and family life cycle. The life cycle involves grouping people based on their stage in life, rather than simply classifying them by their chronological age. Hill et al. examined how the different motivations toward the resort vacation cross four different life cycle stages: ( 1 ) single without children, (2) married without children, (3) single with children, and (4) married with children. They found no significant difference between life cycles for the relaxation and escape motivation and the novelty, education, and prestige motivations. However, the motivation o f enhancement o f kinship relationships is more important to those who are married than those who are single. In summary, motivation is just one o f the many variables that contribute to explaining and predicting tourist behavior. Travel decisions are the result o f several motives. Although several different types o f needs and motives have been mentioned to help explain the push or psychological motivations for travel, there is no single theory o f travel motivation that can completely explain tourist behavior. Understanding the tourist's motivations is important to destination areas Such understanding would enable providers 29 to create the activities, attractions, and services which fulfill the needs o f each individual traveler Information search in Tourism For many tourists, gathering, processing, and evaluating information is an integral part o f the travel experience Information search activities help to fulfill vacation motives such as achievement, social affiliation, culture experience, escape from the everyday environment, relaxation and novelty seeking (Plog, 1974; Crompton, 1979). Furthermore, an information search is essential before deciding among alternative destinations, attractions, activities, and lodging choices The marketing literature generally distinguishes between internal and external information searching behaviors. Gitelson and Crompton (1983) found that tourists like to search external sources in order to learn about alternative destinations which may meet their needs, the characteristics and attributes o f those destinations, and their relative desirability Etzel and Wahlers (1985) assert that the extent o f consumers' information-seeking depends on factors such as: the perceived utility o f the information, the amount of uncertainty involved, the perceived importance o f the decision, and the cost o f acquiring information. Similarly, Fesenmaier and Johnson (1989) indicate that tourists use more technical sources when making more involved decisions. Chon (1991) indicates that significant differences exist among tourists with regard to their socioeconomic circumstances and their travel-information-seeking behavior Travel-information-seeking 30 behavior has been found to be influenced by the following variables: ( 1 ) gender; ( 2 ) age; (3) previous visitation to the destination area; (4) type o f lodging accommodations used; (5) frequency o f vacation trips taken per year; and ( 6 ) the likelihood o f repeat visits to the same destination. Since identifying and understanding information seeking behavior can assist in planning promotional campaigns, it is highly important to identify who requests travel information and how it contributes to tourism decision-making. As Gunn (1987, pi 71) notes: "Communications o f all types are becoming more and more im portant to link the consum er to the product, Simply, i f tourists do not know about travelways, attractions, sen'ices, and facilities, and do not know how to get to them, tourism can be less than satisfactory f o r both consum ers and suppliers. Certainly, the planning f o r tourism m ust include understanding o f the essential com ponent o f prom otion/inform ation". Importance o f Repent Visitors to the Tourism Industry Competition in the tourism industry is intense. Unlike most other retail purchases, the tourist can neither directly observe what is being bought, nor try it out inexpensively. Tourists have high quality expectations o f their forthcoming vacation and also demand value for their money This trend is increasingly evident as value-conscious consumers seek destinations that offer the best value (Reid and Reid, 1993). Frequent-Flyer programs offered by airlines and Frequent-Stayer programs offered by hotels were introduced to attract repeat patronage. According to Gitelson and Crompton (1984), building repeat visitation is a means by which tourism suppliers can increase revenues and decrease costs. It can also reduce 31 reliance on the difficult task o f attracting new visitors. In other words, focusing on repeat business permits tourism industries to target a particular segment and solicit direct responses to promotions. Furthermore, suppliers can more effectively measure promotional success and accurately forecast revenue from promotion investments. Haywood (1989) reported that it is five times more expensive to obtain a new customer than it is to retain a current customer Although the figure is speculative, it is generally agreed that it is cheaper to promote to current visitors than to promote to those not familiar with a destination. Fakeye and Crompton (1991) examined destination image differences held by prospective, first-time, and repeat visitors to the Rio Grande Valley in Texas. They report that repeat visitors appear to have greater awareness o f social opportunities, attractions and may have enhanced social networks, leading to a more complex image o f the destination. Gyte and Phelps (1989) reported that repeat visitors are more likely to return to the same destination the following year They also indicate that tour operators wishing to cultivate repeat business need to ensure that clients have a good holiday experience and visit on each and every visit to a given destination. In sum, repeat visitors to tourism destinations constitute an unique market segment and one which can be effectively exploited at relatively low cost. Destinations should stress external and internal marketing communications directed at the repeat visitor segment to capture this profitable market (Reid & Reid, 1993). 32 Influence o f Socioeconomic Factors on Tourism and Travel Income is probably the most significant determinant o f a household's probability of traveling and upon its level o f travel expenditures (Hagemann, 1981). The top 20% household income bracket accounts for almost one-half o f total expenditures for pleasure travel (Linden, 1980). Mak, Moncur and Yonamine (1977) reported that higher income visitors to Hawaii spent more per day and stayed longer than did their lower income counterparts. Similarly, Ethridge (1982) suggested that international travel originates in areas o f greater economic development. Education also influences the propensity to travel. Jorgenson (1976) reported that 45% o f travelers in the 1972 National Travel Survey had completed some college, whereas only 22% o f the total population had a similar level o f education. Educational attainment o f the "head o f household" is likely to significantly influence travel behavior. A higher education increases one's awareness o f the world and appears to be a powerful predictor o f a household's propensity to travel (Hagemann, 1981). Just as education influences the propensity to travel, it also influences vacation expenditures. Mak, Moncur, and Yonamine (1976) reported that visitors to Hawaii who were more highly educated spend less on average per day than did less educated visitors. Linden (1980) and Hagemann (1981) approached the issue o f education as an influencing factor on vacation expenditures from a macro perspective. Both researchers found that persons with higher educational attainment spent more on travel than did less educated persons. When examining the influence o f a graduate education on travel expenditures, the research findings differed. Hagemann (1981) found that households in which the head had 33 attended graduate school spent less on travel than at the next lower level. Conversely, Linden (1980) reported that: " Dollar expenditures for vacations by householders whose head has more than four years o f college runs two-and-a-half times higher than the all-country average " Children also influence travel behavior. The presence o f small children tends to act as a physical constraint on the family Hagemann (1981) noted that family size is negatively correlated with travel. In other words, larger families have a decreased propensity to travel and take shorter pleasure trips. The presence o f children younger than six had a more negative effect on travel than did children in the intervening years. Also, the greater the number of children under 18 years, the lower were household expenditures for travel (Hagemann, 1981). In fact, spending on vacation trips for families with children under six years o f age is approximately 30% below the national averages. Once the youngest child reaches 18 years o f age, however, vacation expenditures exceed the national average by almost 50% (Linden, 1980). Age is another influencing factor in one's propensity to travel. The National Travel Survey revealed that only one-third o f those persons under 25 years reported any travel during the survey period. Sixty-percent o f all travelers were in the 25-64 age category. Age was also found to be a significant variable in a study o f domestic travelers to Hawaii. Age was found to influence length o f stay and the amount o f expenditures while in the test area Persons in the young and the retired age groups stayed significantly longer than those in the middle years although they spend significantly less per day (Mak et al., 1977). While it is important to be aware o f the influence o f demographics on the travel industry. 34 it should be noted that demographic and socioeconomic characteristics, while enabling certain kinds o f consumer behavior, are not sufficient to guarantee a particular behavior. Logistic Regression: Overview Logistic regression (LR) is a robust statistical procedure for modeling the relationship between one dichotomous dependent variable and several independent variables. Since the Framingham study by Truett, Cornfield, and Kannel (1967), logistic regression has become the standard method for regression analysis o f dichotomous data. The logistic regression model has been applied in many fields, especially by economists and epidemiologists, but it has been infrequently used by tourism and travel researchers. Since it is often the most appropriate approach for analyzing dichotomous variables, the technique should be familiar to tourism researchers so that their work can keep pace with improvements in the field o f statistical analysis. In this study, logistic regression was used to examine the relationship among visitors and to predict the propensity to visit Frankenmuth in the near future. This section begins with a brief but necessary description of linear regression, discriminant analysis, and logistic regression. Comparisons are made among these types o f statistical analysis. Linear Regression M o d e l: Functions and Violations Regression is a set o f statistical procedures that can be used to make predictions about one variable (called either predicted, criterion, or dependent variable) based on the knowledge o f another variable (called the predictor or independent variable). Linear 35 regression tests whether two or multiple variables are linearly related and calculates the strength o f the linear relationship, if the relationship between the variables can be described as: r=oc + fk Where Y: dependent variable c l : intercept, presents the value o f Y when X is zero P : slope o f the line, presents the change in Y associated with a one-unit increase in X. X: independent variable Normally, the regression procedure involves three steps: (1) identify two or multiple variables that are correlated to establish the regression equation, ( 2 ) estimate the goodness-of- fit o f the regression equation, and (3) apply the regression equation to data from subjects not included in the original sample to predict the outcome or dependent variable. For more than one independent variable, regression analysis involves investigating the dependence o f Y on the independent variables (X, X, X, ....X ). The ordinary multiple linear regression model may be written as Y = oc + p iYi + p 2 + ■•■+ P />■+■£ where a represents the intercept, the value o f Y when X is zero; Pi, p 2 , P/> represent the regression coefficients or partial slopes that characterize the relationship between the independent variables, X, , X 2 ,...X , and the dependent variable Y ; and e represents the error term, a random variable representing the error in predicting Yfrom X. 36 For an individual case i, Y, = a , + pi,Yi, + $ 21X 2 , +... + $P,XP, + e, , the subscript indicates that the equation is predicting values for specific cases, indexed by i (i = 1 for the first case, i = 2 for second case, etc.). This equation is used to calculate the value o f Y for a particular case, i, rather than describing the relationship among the variables for all o f the cases in the sample or the population. Estimates o f the intercept, a , and the regression coefficients, p or ( P i , P 2 , . , P/,), are obtained mathematically using the method o f ordinary least squares (OLS) estimation. A A These estimates produces the equation Y = a + hX, or in the case o f several predictors, A A A A Y = ci + b ]X] + />2-T2 + A + bPX p , when Y is the value o f Y predicted by the linear A A A regression equation, a is the OLS estimate o f the intercept a , and b (or b 1. />2. the OLS estimate for the slope P( or the partial slopes Pi, P 2 , A A b,) is , P/0 Residuals for each A case, e, , are equal to ( Y —T,), where Y, is the estimated value o f Y for case i. The basic assumptions o f the OLS regression model are: ( 1) (error term) is a random variable with mean zero and variance c r, that is E(e, ) = 0, V(e,) = c r ; and (2) e, and e} are not correlated, i ^ j so that Cov( e,. e , ) = 0; thus the variance o f Y, = cr; and Y, and Yj where i ^ j are not correlated. A further assumption which is not necessary for estimation, but is required in order to apply statistical tests such as the T-test or F-test, is t h a t , is a normally distributed random variable with zero mean and variance :, that is e ~ N(0, a 2 ) (Fraper and Smith, 1966). However, a number o f authors have noted several inadequacies and limitations o f the linear regression model in cases where the dependent variable is dichotomous (Cox, 1970; Anderson, 1980; Bull and Donner, 1987). The main disadvantages o f the linear 37 regression model are violations of assumptions that Y, 's are normally distributed with mean 0 , and variance c r , and 0 , is linearly dependent on X, 's. The limitations and the disadvantages o f the linear model when the dependent variable is dichotomous are summarized below: 1. It is quite possible that the predicted values of, will exceed one or take on a negative value. 2. Since Y, takes only the values 0 and 1, then Y ,2 = Y, and variance o f Y, = 0, (1- 0, ). This violates the assumption o f the least squares estimate that variance (Y, ) = a 2 (i.e., the assumption o f homoscedasticity). Using the OLS estimate could give us an unbiased estimate of, , but it is not an efficient estimator. This problem has led to the development o f the logistic regression model which addresses the above problems by transforming the probability o f success into a continuous variable that can take on any value along the real line (-«>,«>). Loeistic Regression: Function and M odeling Predicting whether an event will or will not occur, as well as identifying the variables useful in making the prediction, is important in most academic disciplines and in the "real" world. A variety o f multivariate statistical techniques can be used to predict a dependent variable from a set o f independent variables. Multiple regression analysis and discriminant analysis are two related techniques used to develop such predictions. However, these techniques are problematic when the dependent variable can have only tw o values: an event occurring or not occurring. 38 Logistic regression represents an alternative method o f classification to consider when the multivariate normal model is not justified. The logistic function describes the mathematical form for the base o f the logistic model. The function is given by 1 over 1 plus e to the minus z. as follows: Probability(event) = P(Y=1) = |^ ^ . P|A|^ AV ^ where e is the natural logarithm z varies from -OO to +00 . Thus, the value o f logit P(Y=1) is in the range between 0 and 1, regardless o f the value o f z. Therefore, it is not possible to obtain a risk estimate either above 1 (e.g., absolutely certain) or below 0 (e.g., totally impossible) which explains why the logistic model is often the first choice when a probability is to be estimated. The logistic regression model is also the solution to a problem involving the distribution o f normal errors with binary outcomes. In fact, this model was designed for analyzing binary data (Cox, 1970). One can directly estimate the probability o f an event occurring for a single independent variable where (Y,) takes the value "0" and "1" . The expected value o f Y, is E(Y, ) =P(Yt ) and Y ^ l . where P(Y, ) represents the probability o f Y, equal to 1 (Probability o f event occurring) 1-P(Y1) represents the probability o f Y, equal to 0 (Probability o f event not occurring) 39 W e could try to model the probability that Y I as P(Y I) a + pA”, but we would run into the problem that although observed values o f P(Y= I) must lie between 0 and 1, predicted values may be less than 0 or greater than 1 . A step toward solving this problem would be to replace the probability Y with OddsfY I) Here OddsfY I) is the ratio o f the probability that Y probability that Y I I to the 0 OddsfY 1) is equal to P(Y 1)/ 1-PfY I ) . For example, if the probability o f event occurring equals 0 .2 , then the probability o f event not occurring is 0 .8 and the resulting odds calculation is 0 2/0 8 or one-fourth. The meaning o f odds in the context is that the probability o f the event occurring is one-fourth the probability o f the event not occurring. An odds ratio greater than 1 indicates an increased likelihood o f the event occurring, while an odds ratio less than 1 indicates a decreased likelihood o f the event occurring. A further transformation o f the odds term produces a variable that varies from negative infinity to positive infinity which is called the /ogil o f Y. It is the natural logarithm o f the odds term written as ln{P(Y 1) 1-PfY 1)}. The transformation o f logit PfY 1) yields the following equation: /■Y Y I lA V PfY 1 (l /) = O Oi .■“ ’M _ ) l+exp(u+P|.V ) where a and Pi are coefficients estimated from the data, X is the independent variable, and e is the base o f the natural logarithms which is approximately 2 .718. level, the logit transformation (0^ ) , in terms o f 0 ,, is defined as follows: 0' = logit ( 0 ,)= In*[ 73^ ] = ln„ = lnt-|c(u+,i‘n ) = a + p,.V For a single 40 For more than one independent variable, the model can be written as ''P 'P | ,1 r\ 1 ° r etlulvalentlV 0 - ^ The logistic regression model is represented as follows: Q' —logit (0 ,) = l n ^ - p p J = a + PuVi + $ 2 X 2 +... + $PX P Logistic regression is a type o f log-linear analysis used with a binary dependent variable Logistic modeling is based 011 the assumption that the underlying relationship between the dependent and independent variables is an S-shaped function called a sigmoid curve (See Figure 3). The values for the dependent variable are presented as probabilities that range between 0 and 1 with the maximum slope o f the curve in the mid-range. This implies that the independent variable has its greatest impact at some midpoint, where the slope o f the curve is the greatest, and less impact at the ends o f the range where the slope o f the sigmoid curve is smaller. 41 P ro b ab ility o f Y=1 D ependent V ariab le In d e p e n d e n t V ariable Figure 3: The Logistic Function: Sigmoid or S-Shaped Model 42 As the value o f an independent variable changes, the value o f the exponent changes, and the change in the dependent variable is exponential Thus, the logistic function is nonlinear, and a unit change in an independent variable has a different impact on the dependent variable at different values o f the independent variable. The logistic regression model is a sensible method for regression analysis o f dichotomous data for two primary reasons: " 1 ) it is an extremely flexible and easily-used function and 2) it lends itself to a substantively meaningful interpretation" (Hosmer & Lemeshow, 1989). However, it is important to understand that the probability, the odds, and the logit are three different ways o f expressing exactly the same thing. The logifform o f the probability is the best one to analyze dichotomous dependent variables, although the probability or the odds is more easily understood. Comparison between Discriminant Analysis and Logistic Regression Discriminant analysis (DA) is a statistical procedure for identifying characteristics that are important for distinguishing among groups It is similar to multiple regression in investigating the relationship between dependent and independent variables except the criterion is dichotomous rather than continuous (Ghiselli, 1981) To minimize the rate o f misclassification errors, the discriminant function distinguishes between pre-defined groups by maximizing between-group variance relative to within-group variance. Discriminant analysis is appropriate when the discriminating variables have. (1) a multivariate normal distribution, and ( 2 ) equal within-group covariance matrices. Without an underlying normal distribution, the statistic can be very misleading such that 43 individual group classification rates will be distorted. A number o f authors have concluded that when within-group covariance matrices are not equal, discriminant functions can include meaningless variables, produce inconsistent coefficient estimates, have a poor fit to the data, and generate substantial bias (Fienberg, 1991; Gilbert, 1969). Therefore, another multivariate technique for estimating the probability o f an event occurring, the logistic regression model, was applied in this study. This model requires fewer assumptions than does discriminant analysis, and, even when all the assumptions related to discriminant analysis are satisfied, logistic regression is as efficient as discriminant analysis.. In linear regression, the least squares method is usually applied to estimate the parameters o f the regression model. This means that regression coefficients that result in the smallest sums o f squared distances between the observed and the predicted values of the dependent variable are selected. Unfortunately, when the method o f least squares is applied to a model with a dichotomous outcomes, the estimators no longer have these smallest sums o f squared distance between the observed and the predicted values o f the dependent variable. In logistic regression, the parameters o f the model are estimated using the maximum - likelihood estimation (MLE) method. That is, the coefficients that make our observed results most "likely" are selected. However, logistic regression is a mathematical modeling approach that can be used to describe the relationship o f one or several independent variable(s) to a dichotomous dependent variable In general, with the goal being the assignment o f observations to correct categories, DA is used to identify a set o f predictors which best discriminates two groups 44 o f observations. Logistic Regression (LR) is used to make inferences about the relationship o f an independent and dependent variable, and the interpretation o f the effect o f the independent variable is straightforward When choosing between DA and LR, the researcher should consider: (1) whether an investigation's primary purpose is classification, description, or prediction, ( 2 ) the characteristics o f the sample, and (3) the assumptions o f methods. In sum, the advantages o f logistic regression are: (1) there are no necessary assumptions about independent variables and ( 2 ) logistic regression can accommodate an extremely skewed distribution o f the dependent variable. In building a model, Klecka (1980) recommended that " unless there are strong theoretical reasons for keeping them, it is advisable to eliminate weak and redundant variables. Their presence only complicates the analysis and they may even increase the number o f misclassifications." One objective o f this study is to predict the propensity to take a trip to Frankenmuth. Therefore, LR is better than DA because LR requires fewer assumptions, and it is more parsimonious to build a simple and plausible model for the data presented and to estimate the effect o f each explanatory variable on the dependent variable (Hanushek and Jackson, 1977). C H A P T E R III M ETHODS In this study, the cross-sectional mail survey method was employed. A survey design provides a quantitative description o f some fraction o f the population -the samplethrough the data collection process o f asking questions o f people (Fowler, 1988). This data collection enables a researcher to generalize the findings from the sample to the full population. According to Babbie (1990), the purpose o f survey research is to generalize from a sample to a population so that inferences can be made about some characteristic, attitude, or behavior o f this population. Therefore, through the careful design o f data collection, a representation sample can be selected from the population. Instruments are used to collect data. In addition to data collection, independent and dependent variables are selected to correspond with study objectives and hypotheses to be tested. Furthermore, appropriate data analyses are employed and findings are presented. Finally, models are developed to predict the dependent variable, in this case, for example, the propensity to visit Frankenmuth. The procedures used in this study include: sample design, data collection, variable selection, statistics used in data analyses, and the research models used to predict travel decisions. 45 46 Sampling Design The main objective o f a sample design is to insure that the sample selected is representative o f the population from which it is drawn. Sampling designs may be divided into two basic classifications: namely probability and non-probability sampling. The essence o f probability sampling is that each member o f the sample population has a known probability o f being selected. Probability sampling permits one to generalize from sample results to the population. Basic types o f probability sampling include simple random sampling, stratified random sampling, systematic sampling, cluster sampling, and multi-stage sampling. The most frequently used methods o f non-probability sampling are purposive or judgmental sampling, and quota sampling. However, these non-probability sampling methods are not commonly used in tourism research. In this study, probability sampling was used. As was noted in Chapter I, the population o f interest in this study is people who requested travel information from the Frankenmuth, Michigan Chamber o f Commerce. Specifically, the sampling frame consisted o f the 5,967 Americans or Canadians who requested information via a toll free call or by mail from the Frankenmuth Chamber o f Commerce between September 1, 1993 and March 15, 1994. According to Woodside and Soni (1988), the highest quality o f inquiries come from toll-free call campaigns that may attract more upscale respondents in terms o f income, education, and occupational status. For this study, inquirers from businesses and institutions (e.g., travel agencies, tour brokers, libraries), and inquirers from Mexico and overseas countries were excluded from the sample frame because the focus was on English speaking pleasure travelers. The 47 questionnaire was sent via certified mail to a systematic sample o f 1263 individuals randomly selected from the above frame. O f the 1263 surveys mailed, 53 were returned due to insufficient addresses, and 595 completed and usable questionnaires were returned. The response rate based upon the 1210 questionnaires delivered was thus 49.2% This response rate is above average for a mail survey o f this kind and is especially high considering the length o f the questionnaire and the absence o f follow-up mailings Follow-up mailings were not employed because their impact on response rate was assumed to be minimal given that the initial mailing was sent via certified mail. However, there is the potential for non-response bias in this study's results. Non-response bias is a systematic error that occurs when a sample is not fully representative o f the population from which it was drawn It results from respondents differing from non-respondents on key variables measured in the survey. For example, respondents may have been more likely than non-respondents to have visited Frankenmuth and/or to have spent relatively large sums o f money there, possibly resulting in exaggerated estimates o f the percentage o f inquirers who visited Frankenmuth and/or their average expenditures. In the tourism research field, it has been observed that non-visitors are sometimes less inclined to respond to a survey o f this type due to a feeling that, since they did not visit, their participation in the survey is unimportant. (To minimize this potential error the cover letter explicitly encouraged non-visitors to respond). 48 Review' of the Questionnaire P csi 2 n The questionnaire was designed to obtain the following general information: ♦ familiarity with the destination, especially, had respondent visited Frankenmuth either before or after requesting information ♦ kinds o f media used to obtain the phone number or address used in requesting information and other information sources used ♦ travel behavior o f visitors (e g , number in party, purpose o f trip, length o f stay, and accommodations used) ♦ level o f visitor satisfaction (e.g., satisfaction with the visit to Frankenmuth, intention to visit again, recommendation o f Frankenmuth as a travel destination to others, and overall ranking o f Frankenmuth as a travel destination) ♦ elapsed time between inquiry, receipt o f information, and visit to Frankenmuth ♦ expenditures during the visit to Frankenmuth ♦ brochure quality (e.g., usefulness o f brochure, interesting to read, attractiveness o f brochure, accuracy o f brochure, and overall quality o f brochure) ♦ brochure effects (e g , increased interest in visiting Frankenmuth, influenced decision to visit Frankenmuth, and increased spending in Frankenmuth) ♦ demographic and socioeconomic characteristics o f inquirers (e.g., gender, residence, employment status, age, and income level) The full set o f questions included in the questionnaire is contained in Appendix A. Note that responses are tabulated and included with each question in Appendix B. As can 49 be seen in Figure 4, a rather complex survey design was required to obtain the information required for the study from respondents The nature o f the data obtained (nominal, ordinal, interval, and ratio) is noted with each question in Figure 4 Data Collection As been mentioned earlier, a mail survey was applied to collect data for this study. The mail survey is an important and extensively used data collection technique o f both industry and academic researchers The advantages o f mail surveys include geographic flexibility, relatively low cost, lack o f interviewer bias, respondent convenience, respondent anonymity, and relative speed o f completion. Although there are several disadvantages o f the mail survey, low response rate is probably mentioned most frequently One o f the unfortunate side effects o f a low response rate is the potential for non-response bias. Many techniques have been designed and utilized to increase response rate. Linsky (1975) divided these techniques into three categories: (1). Mechanical and perceptual techniques that include pre-contact, postcard enclosure, follow-up, types o f mailing for outgoing and return envelopes, length o f the questionnaire, printed versus mimeographed questionnaires, pre-coded versus open-ended questionnaires, and color o f the questionnaires; (2). Broad motivational techniques that include anonymity, cover letters, letterhead, sponsoring organizations and titles, and use o f deadlines; (3). Direct motivational techniques/rewards that include cash rewards, enclosure o f prizes premiums and other non-cash rewards. 50 1. H o w w an th e p h o n e # or ad d re ss o f th e F ra n k e m m u th C h a m b e r of C o m m e rc e a n d V is ito rs B u rea u (F 'C C V D ) o b ta in e d 7 P r i o r to r e q u e s t in g in fo ., h ad >thj ev e r v i s i te d F r a n k e m m u th '1 VIS NO 2a. W h e n w a s th e m o st rece n t o f th e s e v isits? I I ▼ 2b. D id \o u re q u e s t info, from th e F C C Y H to h e lp p la n th e v is it? N ^ .V W h a t w a s the p u rp o s e o f re q u e s tin g info? N 4. I la d you d e c id e d to c o m e to F ran k en m u th b e fo re re q u e s tin g in fo ? N1 5. D id you v is it F ran k en m u th a lte r re q u e s tin g in fo .? N VI S ▼ 6 5 a A r c y o u c o n s i d e r i n g a v i s i t t o F r k m u t h ’’ N W h e n did y ou m a k e th e trip to F ra n k e n m u th 5 b W h e n w i l l t h i s v i s it m o s t l i k e l y o c c u r ' ’ < > alte r r e q u e s tin g the in to fro m F r a n k e n m u t h ’ K 5c I h d you re c e iv e in fo a l t e r your r e q u e s t '' N 7 P rio r t o trip, d id you h av e a n y V FS o t h e r i n f o r m a t i o n ’’ N 5 J W a s th e into y ou re ceive d th e info y ou r e q u e s te d ’ N W h e n d i d y o u r e c e i v e t h e i n f o '* K NO 5 f | > j d y o u r e a d t h e h n v h u r e s t h a l w e r e s e n t t o s o u ’ N' 7a W h a l other so u rc c(s)'’ N VIS Y | 5 h D i d t h e h n v h u r e s d e c r e a s e or i n c r e a s e ^ ^ S W a s F ia n k e m n iu lh the P rim a ry D e stin a tio n ’ N y«*ur i n t e r e s t m M M t i n g F ' r a n k t a i m u l h ’ ( ) * * h>u w o u l d y o u r a l e t h e o v e r a l l q u a l i t y ▼ . .. . . . . .,., 5^. \ M u t d d yiHi ik» wit h m l o ’ N u| t h e b r o c h u te 1( ) Sa W h a l w a s t h e p r i m a r y I W hat d e s t i n a t i o n '1( o p e n j i ▼ 9 ▼ w a s th e P n m a r y P u rp o se o f this trip * N j ▼ ▼ 1*r q iu Q .2 H N = Nominal data I = Interval data NO O - Ordinal data R = Ratio data Figure 4: Flowchart of Q uestionnaire Design 51 from Q9 V 10. I>id you s p e n d a n y n ights a w a y from ho m e on th is trip ? N y i :f ' 1()a. D id you s p en d any n ig h ts in Ir k m u th on th is trip ? N vi s ^ i i ▼ 10b. W hat w as th e p rim a ry d e s tin a tio n 9 (o p e n ) 10c. H o w m an y n ig h ts d id you sp en d in F rkm uth? R l()d. W h ere w e re th e se n ig h ts in F rkm uth sp en t? N yks ^ 1 1. W h en you v isite d I rk m u th on this trip , w ere you ^ a mcml>er of an o rg an i7 cd to u r gro u p ? N i ^ 11a. l>id an y o n e acco m p an y you on th is trip ? N j V Yf - . S 11b. H ow m an y p erso n (not in d u c in g v o u rs c ll) -• ac co m p an ied you? R 1 Ic. W lio w e re th sc s p ers o n s ? N ! ▼ 1 i i ▼ 1 2 . W h ile you w ere in F rk m u th on this trip , did you or any o th e r m e m b e rs of'yvuir S P K N D IN G I 'N I T spend any m oney? N Y FS I NO I ▼ 12a. H ow m a n y p erso n s, in c lu d in g y o u rself, w e ic in your S PF N D IN C i U N H o n th is trip ? 1 12b. H ow m u c h m o n e y d id y o u r .SPKNDINCt D N M s p e n d th is tr ip in F tk m u th ? R 12b6 ( ta so lin c , o il, re p a irs, an d o th e r 1 2 b l. L odging? v e h ic le -re la te d ite m s ? 12b2. C a m p in g I:cc? I2 b 7 )2b.V ( h i t s , cra fts, so u v cn ic n cc, ^ () c to ss -u )u n (ry sk is, e tc .) c lo th in g , a n d /o r sp e c ia lty ? 12b4. (iro e e ry a n d c o n v e n ien ce sto re lb o d a n d b ev e ra g es? 12b5. R e s ta u ra n t a n d b ar m e a ls d rin k s? ▼ R en tal fees (fo r golf ca rts, 12bK. G u id e lo u rs ? 12bV. A ll o th e r item s? 1 2 b l0 . Total? ▼ I .V W as y o u r e x p e rie n c e in F ik m u lh on th is trip ? (> 14. H ow lik e ly arc you to v is it I rk m u th ag ain ? () 15. A lte r th is trip , d id ytni reco m m en d a v isit to I rk m u th to an y o n e ? N 16. ( )verall, how w o u ld y o u rale I rk m u th as a to u rist d e s tin a tio n ? O 1 7. D id you rece iv e any in lo . in resp o n se to your ic q u e s t? N ▼ N = Nominal data I = Interval data O = Ordinal data R = Ratio data Figure 4 (co n t'd ) 52 Q 17 vi* Q 5i Q Ss Q17(N O ) | 17a W as the in to vnn received the info you requested? N 17b. W hen did you receive the info ? I 17c. D id you receiv e info, front Ihc I C C’ V H ’ N A fter this trip He fore the trip 1 18. D id you read the brochure that w as sent to you? N VMS I I t NO 18a. W hat did you do w ith in lo ? N 1 19. D id w xt c o n s u lt th e b ro c h u re w h ile v is itin g F rk m u th on th is trip ? N 2 0 . I lo w u s e fu l w a s th e in fo , in th e b ro c h u re ? O 21. T o w lint e x te n t w a s th e b ro c h u re in te r e s tin g to re n d ? O 22. T o w h a t e x te n t w a s th e b ro c h u re a ttr a c tiv e in d e s ig n ? O 23. T o w h a t e x te n t d id th e b ro c h u re d e c r e a s e o r in c re a s e y o u r in te re s t in v is itin g F r a n k c n im ith ? O 2 4 . T o w h a t e x te n t d id th e b ro c h u re in llu e n e e s o u r d e c isio n to v is it F ra n k c n im ith ? O 2 5 . T o w h a t e x te n t d id th e b ro c h u re c a u s e y o u to s p e n d m o re m o n e y in I'rn n k e m m u th on th is tr i p ? ( ) 20. B a s e d o n y o u r e x p e r ie n c e on th is trip , h o w a c c u ra te w a s th e in fo in th e b ro c h u re ? O 27. I lo w w o u ld y o u r a le th e o v e ra ll q u a lity o f th e b ro c h u re ? C) 2 8. W hat is the Z IP C O D Ii o f your perm anent residence? (open) 29. W hat is your gender? N 30. W hat is your present em ploym ent situation? N 31 H ighest year o f form al schooling com pleted? I 32. In w hat year w ere you IK irn ? I 33. IIow m any people reside in your household (including yourself)? 1 34. I low m any e m p lo y e d persons age 18 o r o ld e r re sid e in your household? I 35 D o any c h ild r e n u n d e r age 18 r e s id e in your iKUtsehold? I yks | ▼ I n o 1 - 35a. W hat is the ACiK o f tlie oldest child living at Ivome? I 35b. W liat is the AG K o f the youngest child living at hom e? I 36. W hat w as your total 1993 household incom e before taxes? O N = Nominal data I = Interval data O = Ordinal data R = Ratio data F igure 4 (co n t'd ) 53 In this study, cover letters with letterhead were personalized and individually signed The questionnaires were mailed with a stamped return envelope via certified mail to the designated sample. C oupons and other incentives were enclosed with the questionnaire. In addition, three prizes were sent to respondents that w ere selected from returned questionnaires. As has been noted, follow-up mailings were not part o f the sample design, rather available project funds were allocated to response rate boosting strategies, such as using certified mail, which have proven more effective in enhancing response rates in mail survey. A total o f 595 usable questionnaires were returned The response rate was 49.2% after an adjustment for undeliverable mail was made. Data Preparation Data preparation is the procedure o f organizing data for use in statistical analysis All data to be entered into the computer must be given a structured form or format so that they can be processed by a computer. The document explaining how the data were coded and the form o f the file, is called a code book. In coding data, each unit for which data is collected is called a case. There are a number o f methods for assigning values to each piece o f information. The process o f assigning values to an item is known as scoring. The different kinds o f values that are assigned are known as "level o f measurement". The analysis technique used to answer a researchable question will depend upon the level o f measurement employed. The level o f measurement indicates the ordering o f the information or the facts, and the distance o f one value from another. 54 Data may be characterized as either discrete or continuous. In this study, for example, the variable relative rating o f brochure attractiveness was measured on a 7 point scale ranging from 1 (very unattractive) to 7 (very attractive) and thus is a discrete variable. The actual distance respondents travel away from home is an example o f a continuous variable used in this study. Four levels o f measurement (nominal, ordinal, interval, and ratio) are commonly used in survey research applications. Definitions o f these four levels o f measurement follow. Nominal data are those where values assigned do not relate to the characteristics o f a case nor the relationship o f one case to another The numbers are merely labels. An example o f a nominal variable used in this study is gender (e.g., coding male = 1 and female = 2 would be a nominal scale; males do not com e first, two males do not add up to a female). Ordinal data are those which can be ranked according to some hierarchical scheme but whose absolute position with respect to the other values can not be assigned. An example from this study o f an ordinal variable is "satisfaction after trip" (Question 13). Scores on this attitude scale are assigned 1 as "much worse than expected", 2 as "somewhat worse than expected", 3 as "about expected", 4 as "somewhat better than expected", and 5 as "much better than expected". The difference between the ranks need not be equal. Interval measurements also utilize numbers to describe conditions, but these numbers have more meaning than do ordinal measurements. Interval data are those where the exact distance between each value and the magnitude o f this distance are known. The most common example o f an interval measurement is the Fahrenheit temperature scale. The difference between 70 degrees and 80 degrees is the same as the difference between 50 degrees and 60 degrees Ratio data 55 have all the characteristics o f interval data, but they have the additional characteristic o f a true zero value. In comparison with the Fahrenheit tem perature scale, the Kelvin tem perature scale is a ratio measurement. Thus, while 40 degrees Fahrenheit is not twice as warm as 20 degrees Fahrenheit, 200 Kelvins is twice as warm as 100 Kelvins. In this study, age, length o f stay, travel party size, and year o f attaining school are examples o f ratio measurement. The data preparation procedures used for this study are described below. Questionnaires were returned to the Travel, Tourism and Recreation Resource Center (TTRRC) at Michigan State University. Returned questionnaires were dated and checked for completeness. SPSS for Windows Release 6.1 was employed to conduct data entry and data analysis. SPSS for Windows Release 6 .1 brings the full power o f the mainframe version o f SPSS to the personal computer environment. It enables one to perform many analyses on a PC that were once possible only on much larger machines. A code book was prepared to guide coding. The total number o f questionnaires received was 600. Five useless questionnaires were deleted. Hence, the 595 questionnaires were available for analyses by the program. The frequency statistics method was applied to perform data cleaning. In addition, several crosstabulation tests were performed to further enhance data quality. In order to obtain useful data for analysis, data transformation procedures were employed. Data transformations are often used to collapse categories o f nominal or ordinal data to obtain a smaller, more useful number o f categories. For example, in this study, household income (Question 36) wass collapsed from 12 levels into 5 levels in comparisons o f 56 socioeconomic characteristics between visitors and non-visitors. Binary type data, such as male/female, yes/no, were coded as a dummy variable (e.g., male = 1 and female = 0 , yes = 1 and no = 0) to perform data analyses. Several other data transformations and recoding processes were employed as needed to perform desired statistical analyses Variable Selection 1. Medium/media sources: These include: newspaper article, newspaper advertisement, magazine article, magazine advertisement, radio advertisement, television advertisement, friend/relative/co-worker, Frankenmuth area business, travel show, travel agent, Michigan Travel Bureau, regional tourist association, telephone directory, brochure, and others. These data were assigned as dummy variables (e.g., 0 = did not choose this medium as an information source and 1 = did choose this medium as an information source). 2. Travel behavior variables: These variables include party size, length o f stay in Frankenmuth, accommodations used in Frankenmuth, and purpose o f trip. Party size and nights stayed in Frankenmuth are ratio data. Accommodations used and purpose o f trip are nominal data 3. Travel satisfaction variables: These variables include trip experience rating, intention to visit again, recommendation to others, destination ranking. Experience, intention, and destination ranking were measured using a Likert scale and thus are ordinal data; and recommendation is a nominal variable 57 4. Brochure readership variables: These variables include "read" brochure and "consult" brochure. Both are nominal variables. 5 Elapsed time variables: These variables include time between inquiry and receipt o f information for all inquirers and time between information receipt and the visit for visitors. These variables are continuous (ratio) variables. 6. Spending pattern variables: There involve the actual dollars spent by parties during their stay in Frankenmuth. They consist o f the following categories: money spent on lodging, camping fees, gifts, grocery and convenience store purchases, restaurant and bar spending, gasoline and auto-related purchases, rental fees, guided tours, and all other items. All these variables are continuous (ratio) variables. 7. Brochure variables: These variables include brochure: usefulness, interest, attractiveness, influence, accuracy, and quality. All o f these variables are ordinal variables. 8. Socioeconomic variables: These variables include: sex, age, employment status, educational background, household size, and household income. Age, educational background and household size are continuous variables. Household income is an ordinal variable. Sex and employment status are categorical variables (nominal variables). Data Analysis In general, analytic techniques are used to gain an understanding o f phenomena by discovering relationships between variables which are thought to affect the phenomena. The proper approach for examining data is to draw on prior empirical studies and theoretical knowledge and insight about the social processes that might be involved and 58 then to test whether the data in question exhibit the characteristics expected from prior research and accepted related theory In other words, to guide the search for patterns in the data, we can use theoretical models o f what might be happening. A model is a theory or set o f hypotheses which attempts to explain the connections and interrelationships between concepts (Gilbert, 1969). What particular statistical test should be used to decide whether a result is statistically significant or not depends on: (1) Are the variables related or from the different groups9, (2) Are the variables categorical or non-categorical?; (3) Are the variables normally distributed or distribution free?; (4) Do the tw o or more groups being compared consist o f different individuals (unrelated) or ones which are the same or have been matched (related)?; and (5) How many variables groups are compared? In order to study the individual characteristics o f inquirers and test the factors that influence their travel decisions, inquirers were divided into: ( 1 ) those who made a visit after inquiry and those who did not make a visit after inquiry , and ( 2 ) first-time visitors (those visitors who had not visited Frankenmuth before) and repeat visitors (those visitors who had been to Frankenmuth before) The research was designed to further understanding o f : ( 1 ) behavior differences between visitors and non-visitors; ( 2 ) behavior differences between first-time and repeat visitors; and (3) factors that influence travel decision making in general. Data analyses included the following procedures: (1) descriptive statistical analysis for the overall data; ( 2 ) statistical significance tests were applied to test for differences between visitors and non-visitors, and first-time and repeat visitors, (3) relationship tests were conducted to investigate association between factors 59 and travel decision making; and finally (4) Logistic Regression Analysis (LRA ) was used to predict future visits Descriptive Statistics Descriptive statistical analyses involve procedures used to summarize, organize, and describe quantitative information. Univariate analysis is one o f the descriptive statistics procedures used for gaining understanding o f the nature o f and characteristics o f a single variable. The most common way to present data in tabular form is as a frequency distribution. In addition to the real numerical values for items in the distribution, their percentage o f the total is often also included. Another type o f univariate analysis is the examination o f the distribution o f the observation to include providing measures o f central tendency exhibited by the data. There are three measures o f central tendency that may be used for this purpose: mode, median, and mean. The mode and median generally can be read directly or easily estimated from the frequency table. Descriptive statistics used in this study include: frequency counts, percentages, modes, medians, and means. In addition, cross-tabulations, also called contingency tables, were used to explore individual characteristics across different segments Statistical Significance Tests A bivariate analysis is conducted to examine the significance o f the relationship or association between two variables A bivariate analysis is used to test hypotheses developed to guide analyses In general, bivariate analysis is based on tw o concepts: 60 correlation and cross-tabulation. Correlation coefficients indicate the degree to which variation in one variable is related to variation in another variable (e.g., pearson product moment correlation and Spearman rank-order correlation) A cross-tabulation is a joint frequency distribution o f two variables that are at the nominal or ordinal level o f measurement Correlation coefficients can be either positive or negative A positive relationship between tw o variables indicates that a high score on one variable is associated with a high score on the variable correlated with it. A negative relationship or correlation indicates that a low score on one variable is associated with a high score on the other variable. Correlation may be high or low between variables. Low correlation coefficients may be either due to the fact that the variables are not associated or that the variables are related but in a non-linear relationship. To determine whether a statistically significant relationship exists between variables, an analysis o f nominal data with a two-variable cross-tabulation could be conducted using the chi-square (A'*) statistic as the decision criterion. The chi-square test for independence provides a standard for deciding whether tw o variables are statistically independent. In other words, chi-square provides a measure o f how much the observed and expected frequencies differ for two variables. But, how much difference between observed frequencies and expected frequencies is needed to reject the null hypothesis? The choice depends on two additional considerations: the degrees o f freedom (Jf) involved in the data set and the desired level o f significance (/■*). According to Reynolds (1984), the chi-square test involves the same logic as advanced multivariate procedures. The 61 chi-square analysis consists o f four parts: (1) the null hypothesis (H o); (2) expected frequencies derived under the assumption that the null hypothesis is true, (3) a comparison o f these expected values with the corresponding observed frequencies; and (4) a judgment about whether or not the difference between expected and observed frequencies could have arisen by chance. However, when a statistically significant relationship is found it only indicates an association between two variables and does not imply that a causal relationship exists between the tw o variables. In this study, chi-square was applied to test for differences between tw o variables (e.g., first-time vs. repeat visitors) in terms o f media sources used, travel behavior, spending patterns, prior experience with Frankenmuth, intention to visit Frankenmuth, brochure quality, brochure effects, and socioeconomic and demographic characteristics o f respondents. The t-test is used to test the hypothesis that two groups have the same population mean. M ore generally, whether the difference between tw o means differs significantly is the basis for the t-test. The procedures for conducting the t-test are as follows: (1) calculate the mean for each group to be compared; ( 2 ) subtract one mean from the other to generate the difference between the two; (3) calculate the t statistic by dividing the difference o f the tw o sample means by its standard error, (4) calculate the observed significance level; (5) reject the hypothesis that two means are equal in the population if the observed significance level is small (P < 0 05) 62 Developing a Prediction Functions with Logistic Repression Analysis The validity o f inferences drawn from modern statistical modeling techniques requires that the underlying assumptions o f the models are met. A critical step in assessing the appropriateness o f such a model is to examine its fit, or how well the model describes the observed data. W ithout such an analysis, the inferences drawn from the model may be misleading or even totally incorrect. In the analyses o f dichot omous dependent variables, researchers often use an ordinary least squares (OSL) regression procedure and defend its use on several grounds, including its simplicity, robustness, and straightforward interpretation (Cleary & Angel, 1984). However, the use o f OLS with a binary dependent variable is technically incorrect because several assumptions o f the OLS model are violated. Binary dependent variables are not normally distributed, the dependent and independent variables do not have a continuous linear relationship, and the error terms are not independent nor homoscedastic Also, the predicted values o f the estimated dependent value will not be constrained between 0 and 1 , which in essence violates the definition o f probability. Logistic regression and discriminant analysis both could be used to analyze a dichotomous dependent variable. Discriminant analysis is used to assign an observation to two groups (1 or 0 for the dependent variable) based on the predicted value that resulted from applying the discriminant analysis equation. Logistic regression is a technique to analyze the effects o f a set o f independent variables on a dichotomous dependent variable which involves minimal statistical bias and loss o f information (Walsh, 1987). In other words, discriminat analysis assigns observations into two different groups without 63 prediction, and logistic regression predicts whether or not an event will occur in future The purpose o f logistic regression differs slightly from discriminant analysis. Logistic regression is used to make inferences about the relationship between an independent and a dependent variable. For this study, logistic regression is used as a model to investigate factors that influence destination decision making and to predict the propensity to visit Frankenmuth in the near future. In summary, this study involved a cross-sectional mail survey research. Data were collected by means o f a questionnaire containing 36 questions. The majority o f these were Likert-scale items based on a scale from "low degree" to "high degree" and dichotomous items based on a yes/no scale. Five hundred ninety-five questionnaires were used in data analysis. Chi-square analyses and t-tests were performed to test the difference between two groups. In addition, logistic analysis was employed to predict the propensity o f visits to Frankenmuth using independent variables identified from applications o f chi-square analyses and t-tests. C H A P T E R IV FINDINGS AND RESULTS In this chapter, survey findings are presented in the form o f descriptive statistics useful in comparing visitors to non-visitors and first time visitors to repeat visitors. Key differences found will be noted and discussed. Data needed to test the study hypotheses will be presented along with results relevant to the statistical tests performed. Finally, results from applying the logistic regression model to the data will be presented along with an analysis o f its effectiveness as a predictor o f visits to Frankenmuth. The chapter is divided into the following four major sections: 1) the characteristics o f visitors vs. non-visitors; 2 ) a comparison o f the characteristics o f first time vs. repeat visitors; 3) the results from testing o f the hypotheses; and 4) the results from the application o f the logistic regression model to predict visits to Frankenmuth. Characteristics o f Visitors vs. Non-Visitors There are many internal and external factors that influence an individual's travel behavior. A better understanding o f what these factors are is crucial to destination marketers interested in enhancing the effectiveness o f their marketing programs. According to M outinho (1987), acquiring and organizing information play an important role in travel decision-making. Thus, researchers should investigate factors that influence 64 65 the process o f travel decision-making including elapsed time between when information is requested and when it is received. Dholakia cl. al. (1993) asserted that the elapsed time between when information is received and decision-making is an important factor influencing buying behavior. In the case o f the tourism industry, a tourist does not want to take the risk and make a visit if s/he does not have enough destination information. An inquirer will choose alternative destinations if s/he has to wait too long to receive requested travel destination information. Therefore, the elapsed time between inquiry and receipt o f information is a factor which could influence the decision to visit a destination. Thus, the longer the elapsed time the less likely one is to take a visit . Results obtained for elapsed time between inquiry and receipt o f information for visitors and nonvisitors are presented in Table 2. Fifty-seven percent o f all visitors indicated that they received their requested information within 7 days o f requesting it whereas only about 33% o f nonvisitors received their requested information within one week. The mean elapsed time reported by visitors (9.44 days) was about half that reported by non-visitors (17.92 days). T-test results indicate that this difference is statistically significant. Thus, if one can totally discount factors which may have biased respondents perceptions o f elapsed time, results provide convincing evidence o f the importance o f expediting delivery o f information requested. 66 Table 2. Elapsed Time between Inquiry and Receipt o f Information by Visitors and Nonvisitors Elapsed time 1 2 3 4 5 6 7 8 9 10 11 12 13 * week weeks weeks weeks weeks weeks weeks weeks weeks weeks weeks weeks weeks and more Visitors % 57.1 26.6 9.8 4.3 Nonvisitors % 32.8 31.2 12.3 8.3 2.4 4.0 2.4 All Respondents % 43.0 29.3 11.2 6.6 0.8 0.8 0.8 1.4 2.3 1.4 0.9 0.5 0.7 0 0 0.4 0.2 1.6 0.9 0.5 2.4 1.6 0 0 0 1.1 0 0.5 m e a n e l a p s e d t i m e fo r v is it o r s (9 .4 4 d a y s ) a n d n o n - v i s it o r s ( 1 7 . 9 2 d a y s ) * T -tc s t p r o b a b i li ty is .0 0 0 i n d i c a t i n g a s i g n if i c a n t d i f f e r e n c e b e t w e e n v i s i t o r s a n d n o n - v i s it o r s Crompton (1966) introduced the gravity model to forecast trips between a single origin and a single destination within a specified time period. The gravity model is designed to reflect the relative strength o f distance as a deterrent to travel. W ith his model, Crampon demonstrated that the longer the distance between origin and destination, the fewer the number o f trips that will be taken. The gravity model appears to apply to the data collected in this study which is summarized in Table 3. Eighty-percent o f visitors 67 Table 3. Distance between Home and Travel Destination by Visitors and Nonvisitors Distance under 1 0 0 miles 101 - 2 0 0 miles 201 - 300 miles 301 - 400 miles 401 - 500 miles 501 - 600 miles 601 - 700 miles 701 - 800 miles 801 - 900 miles 901 - 1000 miles more than 1001 miles Visitors % 42.2 19.3 19.7 1 1.5 2.5 0.8 1.6 0.8 1.2 0 0.4 Nonvisitors % 21.3 18.9 17.7 20.7 4.3 4.0 3.4 4.6 All Respondents % 30.2 19.1 18.5 16.8 3.5 0.6 0.6 4.0 2.6 2.6 3.0 0.9 0.3 2.4 * m e a n d i s t a n c e for v is ito rs ( 1 9 6 . 9 8 m ile s ) a n d n o n - v i s it o r s (3 3 6 .7 1 m i le s ) * T -tc s t p ro b a b ility is .0 0 0 i n d i c a t i n g a s i g n if i c a n t d if f e r e n c e b e t w e e n v is it o r s a n d n o n - v i s it o r s reported that they lived within a 300 miles distance o f Frankenmuth, whereas only 58% o f nonvisitors indicated that they lived within 300 miles o f Frankenmuth. The tendency o f visitors to live closer and travel a shorter distance than non-visitors is confirmed by the reported mean distance for visitors (196.98 miles) and non-visitors (336.71 miles). T-test results confirm that the observed difference is statistically significant. The demographic and socioeconomic characteristics o f the total sample, visitors, and non-visitors are presented in Table 4. Results indicate that there are no significant differences across employment status, education, age, and income for visitors and non-visitors. Results also show that the income bracket mode o f inquirers was $25,000 to 68 Table 4. Demographic and Socioeconomic Characteristics of the Sample Variable Visitors (n=255) % Nonvisitors (n=330) All Respondents (n=595) % % 51 .4 12.5 1.2 15.7 14.9 2.4 1.9 5 4 .4 12.8 1.5 9.7 16.7 3.3 1.5 5 3 .4 12.6 1.3 12.1 16.0 2 9 1.7 G r a d e sc h o o l S o m e h ig h school H i g h sc h o o l g r a d u a t e S o m e c o ll e g e /u n i v e r s i ty C o lleg e/u n iv ersity g raduate P o st g r a d u a t e /a d v a n c e d d e g r e e 1.4 2 .4 34.4 3 2.0 17.8 11.9 0 .9 2.5 3 3 .7 3 1 .0 13.8 18.0 1.2 2.5 3 4 .0 3 1.2 15.8 15 4 U n d e r 25 2 5 - 34 35 - 49 50 - 65 A b o v e 65 4 .3 2 0 .0 41.1 2 4.5 10.2 4 .0 13.3 4 1 .5 3 0 .0 11.1 4.1 16.4 4 1 .2 2 7 .5 10.9 U n d e r $25,00 0 $ 2 5 ,0 0 0 - $49,999 $ 5 0 ,0 0 0 - $74,999 $7 5 ,0 0 0 -$ 1 0 4 ,9 9 9 14.8 3 8.6 29 .2 14.0 3 8 .8 30.1 14.7 3 8 .7 2 9 .6 12.1 11.6 11.6 $ 1 0 5 ,0 0 0 an d Above 5.3 5.5 5 .4 Emnloymcnt W o r k i n g full li m e W o r k i n g p a r t li m e T e m p o ra rily unem ployed H om em aker R e ti r e d S tudent O th er Education Income 69 $49,999 The most frequently indicated employment status for visitors and nonvisitors is working full time (51.4%, and 54 4% respectively). The majority o f the respondents were in the 34 to 49 year age bracket. Approximately, 33.7% o f respondents indicated that they graduated from high school, and 3 1 2 % indicated that they had some college/university education. In summary, demographic and socioeconomic variables are so similar for visitors and non-visitors that they contribute little to understanding o f trip decision making in this case and, thus, provide no basis for enhancing marketing program effectiveness. The distribution o f visitors, nonvisitors, and all respondents for the top seven state/province origins is present ed in Table 5 Michigan residents accounted for 44% o f all respondents. The adjacent states o f Ohio, Illinois, and Indiana and the province o f Ontario accounted for an additional 40% o f all respondents. This information corresponds to the results presented in Table 3 which show that the longer the travel distance the fewer the trips that will be made. It should be noted, however, that the data in Table 5 are not based totally on gravity distance concept because differential awareness stimulated by advertising is also involved. Thus, where Frankenmuth places advertising influences the geography from which inquiries are received The geographic sources o f inquirers in this case is generally consistent with the Frankenmuth Chamber o f Commerce/Convention & Visitors Bureau's geographic advertising/program. In 1994, Frankenmuth Chamber o f Commerce spent $33,000 in promoting Frankenmuth as a destination in the Chicago Sun Times, Cleveland Plain Dealer, Columbus Dispatch, Fort Wayne News, Indianapolis Star News, Michigan Living, Saginaw Valley News Magazine and USA Weekend (Frankenmuth Chamber o f 70 Table 5. State/Province o f Residence of Respondents: Visitors vs. Nonvisitors. State/Province Michigan Illinois Ohio Indiana Ontario (Canada) Wisconsin New York Visitor Rank % 1 56.9 3 10.6 12.2 8.6 2 4 5 7 6 4.3 .8 1.2 Nonvisitors Rank % 1 2 3 4 5 6 7 32.7 16.1 14.5 7.9 6.4 2.1 2.1 All Respondents Rank % 1 2 3 4 5 6 7 44.0 13.3 13.3 8.2 5.3 2.7 1.7 Commerce and Convention & Visitors Bureau, 1994). With the exception o f the latter, these are all media with readership concentrated in Michigan and adjacent states A detailed evaluation o f the geographic allocation o f Frankenmuth's advertising program budget and its effectiveness would be needed to assess the relative importance o f distance and advertising in generating inquirers, and this was not within the scope established for this study. First-time Visitors and Repeat Visitors In this section, visitors who were familiar with the travel destination from one or more prior visits and those who were first-time visitors are compared. Note that first-time and repeat visitors here only apply to this sample population (i.e. inquirers who responded to the survey) and not to the general population o f visitors to Frankenmuth. The two 71 groups are compared across the followings variables: residence status, travel information sources used, trip purpose, travel behavior, elapsed time between receipt o f information and making the trip, ratings o f trip experience and satisfaction with Frankenmuth, perceived quality o f the brochure and social economic and demographic characteristics During the study period, Frankenmuth received 246 individual/party visits from Michigan and other states or provinces from among inquirers who responded to the survey. Residence o f respondents is reported in Table 6 . Michigan is the major source o f tourists to Frankenmuth. In addition, 32% o f visitors come from Ohio, Illinois, and Indiana. Flowever, the distribution o f residences o f repeat vs. first-time visitors differs markedly with Michigan being the source o f almost 75% o f repeat visitors and only about 33% o f first-time visitors. Proximity would appear to be a major stimulus for repeat visitation. This result offers a marketing opportunity for the Frankenmuth Chamber o f Commerce/ Convention and Visitor Bureau. Since repeat and first-time visitors are likely to respond to different marketing stimuli, rather than designing one brochure for both groups, two separate brochures could be designed to service the specific information needs o f the tw o groups. Inquiries received from Michigan residents would be satisfied by sending a "repeat-visitor" brochure. A more refined strategy would be to screen inquiries when feasible regardless o f residence into first-time and repeat visitors and then to satisfy inquirers with the appropriate brochure for that group. 72 Table 6. State/Province of Residence of Respondents of Visitors: Repeat vs. First-time Visitors. State/Province Michigan Ohio Illinois Indiana Ontario Repeat Visitors First-time Visitors Total Visitors (%) 73.2 8.7 7.4 4.0 4.0 (%) 32.6 16.8 15 8 17.9 (%) 57.3 2.1 1 1.8 11 0 9.3 3.3 What are the main sources o f information, other than FCCVB's brochures, sought when planning a travel destination? Evidence suggests that the social environment, specifically the influence o f friends and family, is instrumental in selecting a travel destination (Table 7). Engel et al. (1990) have observed that "hundreds o f studies have found that consumers obtain information about products and services from other people , particularly family members, friends and neighbors, and other acquaintances" Results from this study further confirm the importantce o f other peoples' recommendations in the travel decision process. In Table 7, the most frequently used medium by all visitors was "friend, relative, and co-worker". The difference in use o f this source between repeat and first-time visitors is dramatic with only 20% o f repeat visitors using it vs 43 75% o f first-time visitors. Given the above noted striking difference in use o f friends and relatives as an information source, it is appropriate to explore its root cause in a behavioral sense and its marketing implications. Underlying this behavior is most likely differences in perceived 73 Table 7. Sources o f Information Used by Repeat and First-time Visitors Sources Newspaper article Newspaper advertisement Magazine article M agazine advertisement Radio advertisement Television advertisement Friend/re!ative/co-worker Frankenmuth area business Travel agent Michigan Travel Bureau Regional tourist association Brochure Other Repeat Visitors (%) 4.62 3.08 8.46 10.00 1.54 5.38 20.00 8.46 2.31 10.00 4.62 16.92 4.62 First-time Visitors Total Visitors (%) 3.13 1.56 4.69 1.56 1.56 1.56 43.75 4.69 6.25 10.94 4.69 7.81 7.81 (% ) 4.12 2.58 7.22 7.22 1.55 4.12 27.84 7.22 3.61 10.31 4.64 13.92 5.67 risks associated with choosing a travel destination with perceived risks being far greater for first-time visitors. Risk avoidance behavior would dictate that first-time visitors would seek out the most creditable information sources available to them. O f the sources listed in Table 7, friends and relatives and the Michigan Travel Bureau would clearly be perceived as the most creditable in part because the others are likely to be perceived as less objective. H alf o f first-time visitors sought information from friends and relatives or the Michigan Travel Bureau whereas only 30% o f repeat visitors used these sources. 74 The marketing implications o f these resists are clear; if a destination's focus is on attracting first-time visitors, its strategy should be to stimulate recommendations from prospective visitors friends and relatives and/or the Michigan Travel Bureau The former are best influenced through on-site marketing to visitors and doing everything possible to insure that visitors leave happy. Working with the Michigan Travel Bureau could include providing it with the most update information about Frankenmuth and providing valuable discount coupons for it to distribute to inquirers. Visitors, other than day trippers, mainly stayed in a "hotel or motel" during their trip in Frankenmuth (60.4%) (See Table 8 ). A small portion o f visitors stayed in campgrounds (2.2%). The data indicate that the use o f accommodations is very similar for repeat and first-time visitors; however, nearly half o f repeat visitors were on day trips using no accommodations in Frankenmuth whereas only about 20% o f first-time visitors were on day trips. This difference in proportion o f day trips is consistent with the earlier finding that repeat visitors are significantly more likely to be in state residents and to travel a shorter distance to reach Frankenmuth. The information needs o f first-time visitors are more focused on accommodations than are those o f repeat visitors who are more likely to have acquired such knowledge from previous visits and less likely to be seeking such information in any case. Thus, only minimal space in brochures targeting repeat visitors needs to be devoted to accommodations; a brief listing o f properties and their locations would probably be sufficient. Whereas brochures targeting first-time visitors should present much more information on accommodations since their interest in such information can be expected to very high. 75 Table 8. Accomm odations used by Repeat and First-time Visitors Accommodation Hotel or motel Bed & Breakfast Campground Friend's/relative's home Second home Other None (Day trips) Repeat Visitors First-time Visitors Total Visitors % 52.1 % 76.0 60.4 0.0 0.0 0.0 1.7 2.9 2.2 0.0 0.0 0.0 0.0 0.0 0.0 0.0 21. 1 0.0 46.2 0.0 37.4 As shown in Table 9, the reported mode party size was 2 persons. First-time visitors were far more likely to be members o f two person parties than are repeat visitors. The average party size was 3.4 persons; however, repeat visitors' party size was found to be larger (3.6 persons) than that o f first-time visitors (3.1 persons). As mentioned earlier, day trippers are a very important portion o f total visitors, and Table 9 shows that most o f visitors were day trippers. The composition o f the parties was mainly relatives. The repeat visitors segment included a higher percentage o f the "friends and relatives" category. In summary, these results confirm that Frankenmuth is a short stay destination which attracts primarily couples and families. First-time visitors are more likely to be made up o f related couples who spend at least one night in Frankenmuth. Overall, first-time visitors are more likely to stay longer than are repeat visitors. 76 Table 9. Party Size, Length of Stay, and Makeup o f Party by Repeat and First-time Visitors Sources Repeat Visitors First-time Visitors Total Visitors % % % 0.0 2.2 37.6 14.3 34.6 13.5 49.5 17.2 24.8 6.3 58.2 47.5 18.8 19.8 13.9 54.1 20.4 16.1 9.4 18.3 58.0 23.7 22. 1 19.9 63.1 17.1 3.6 .76 3.1 Partv size 1 2 3 4-5 6 and more Lenuth o f stav (day trip) 1 night 2 nights 3 nights or more 0 M akeup o f partv Friends Relatives Friends and relatives m e a n o f P a rty size m e a n o f le n g t h o f sta y ( n i g h t s ) 21.6 13.7 6.5 70.9 7.0 1.06 0.9 42.3 15.4 30.4 1 1.0 3.4 .84 In the following section, satisfaction with Frankenmuth as a travel destination among repeat and first-time visitors is investigated. Satisfaction is frequently used to refer to the fulfillment o f a motivating state or the meeting o f an expectation, through the purchase o f a product or service. In this study, satisfaction is the result o f interaction between tourists' experiences on the Frankenmuth trip and the expectations that they had prior to visiting Frankenmuth. In other words, satisfaction with the experience depends on how trips are viewed by tourists. Satisfaction is examined in the context of: (1) travel experience, (2) intention to visit again, (3) likelihood to recommend Frankenmuth to 77 others, and (4) overall ranking o f Frankenmuth as a travel destination. As can be seen from the results presented in Table 10, first-time visitors reported a higher level o f satisfaction with their trip than did repeat visitors. The Frankenmuth trip exceeded the expectations o f the majority o f all visitors, but it was somewhat better or much better than expected for almost 70% o f first-time visitors but exceeded expectations for only about half (52% ) o f repeat visitors. This is very good news since such a high performance in exceeding custom er satisfaction is indicative o f an exceptionally good product offering. However, it may suggest that the brochure used to promote Frankenmuth may be perceived as understating the product Frankenmuth has to offer its visitors. Reported intentions to visit again is another, but less direct, measure o f visitors satisfaction with Frankenmuth as a travel destination. Sixty-one percent o f repeat visitors indicated they were certain to visit again while only 33 .3% o f first-time visitors indicated that they were certain to visit again Ninety-five percent o f repeat visitors indicated that they were "likely to visit again" or "certain to visit again" whereas slightly fewer first-time visitors (81% ) indicated such strong intentions to visit again. Ninety-two percent o f all visitors would recommend Frankenmuth as a destination choice to others. First-time visitors are more likely to recommend Frankenmuth to others than are repeat visitors Approximately, 94 1% o f the visitors ranked Frankenmuth as a somewhat better than average destination to visit; 32.2% o f all the visitors identified Frankenmuth as an excellent travel destination. Repeat visitors ranked Frankenmuth more often as an excellent travel destination than did first-time visitors. 78 Table 10. Four Measures o f Satisfaction with Frankenmuth: Repeat vs. First-time Visitors Satisfaction Repeat Visitors First-time Visitors Total Visitors Experience (this trip) Much w orse than expected Somewhat worse than expected About what you expected Somewhat better than expected Much better than expected % 0.0 5.5 42.5 28.1 24.0 % 1.1 2.2 26.9 36.6 33.3 % 0.4 4.2 36.4 31.4 27.6 Intention to visit (again) Certain to not visit again Unlikely to visit again Uncertain whether visit again Likely to visit again Certain to visit again 0.7 1.4 3.4 33.3 61.2 1.1 1.1 17.0 47.9 33.0 0.8 1.2 8.7 39.0 50.2 Will recommend to others Yes No Don't remember 81.1 11.5 7.4 94.6 2.2 3.2 86.1 8.3 5.6 Overall ranking 1 - terrible 2 3 4 5 6 7 - excellent 0.0 0.7 0.7 1.4 26.8 33.6 37.7 0.0 1.1 1.1 8.6 35.5 30.1 23.7 0.0 0.8 0.8 4.2 29.7 32.2 32.2 79 In summary, the experience derived from the tourist product or service evaluated in this study was positive and better than expected by consumers. These results confirm that Frankenmuth is an attractive travel destination which generally exceeds consumer expectations. Repeat visitors appear to be somewhat more satisfied with Frankenmuth than first-time visitors who indicated a lower likelihood for a repeat visit and a somewhat low er overall ranking o f Frankenmuth as a desirable destination to visit. For many first-time visitors, Frankenmuth proved to be worth a visit but lacking in what is needed to insure a repeat visit. Thus, in future marketing research, it would be beneficial to assess what changes in the product offering are needed to encourage more repeat visits to Frankenmuth. Information obtained on the demographic and socioeconomic characteristics o f repeat and first-time visitors are presented in Table 11. The mode reported working status o f visitors is "working full time" and about twice as many first-time visitors are retired (19.8) as are repeat visitors (11.8%). Respondents are relatively well educated; most (96% ) have completed at least high school. First-time visitors are somewhat better educated than repeat visitors. The mode age group o f respondents is 35 to 49 years, followed by the 50 to 65 years age segment First-time visitors are more likely to be senior citizens. The modal income bracket o f visitors is $25,000 to $49,999 (39.2% ), followed by the $50,000 to $74,999 (30.7% ) category. First-time visitors are somewhat more likely to report higher incomes 80 Table 11. Demographic and Socioeconomic Characteristics o f Repeat and First-time Visitors Demographic & Socioeconomic Characteristics Repeat Visitors First-time Visitors Total Visitors % % % E m n lo v m c n l W o r k in g fu ll tim e W o r k in g p a r t lim e T e m p o r a r ily u n e m p lo y e d H om em aker R e tire d S tu d e n t O th e r 52 .3 12.4 1.3 17.6 11.8 2 .6 2 .0 5 0 .5 12.9 1.1 12.9 19.8 2 .0 10 5 1 .6 12.6 1.2 15.7 15.0 2 .4 16 E d u c a tio n G ra d e school S o m e h ig h sc h o o l H ig h s c h o o l g ra d u a te S o m e c o lle g c /u n iv e rs ily C o l le g c /u n iv c rsily g r a d u a tc P o st g r a d u a te /a d v a n c e d d e g re e 2 .0 2 .0 .34.0 3 4 .7 17.7 9 .5 0 .0 2.1 .34.7 2 9 .5 18.9 14.7 1.2 2 1 .34.3 3 2 .6 17.8 11.9 A ce U n d e r 25 2 5 - 54 3 5 -4 9 50 - 65 A bove 65 4 .0 2 1 .5 4 1 .6 2 4 .8 8.1 5.3 18.1 4 0 .4 2 3 .4 12.8 4 .5 2 0 .2 4 1 .2 2 4 .3 9 .9 In c o m e U n d e r $ 2 5 ,0 0 0 $ 2 5 ,0 0 0 - $ 4 9 ,9 9 9 $ 5 0 ,0 0 0 - $ 7 4 ,9 9 9 $ 7 5 ,0 0 0 - $ 1 0 4 ,9 9 9 $ 1 0 5 .0 0 0 a n d A b o v e 10.0 3 8 .5 3 4 .5 1.3.1 3 .5 18.3 4 0 .2 2 4 .4 9 .8 7 .3 13.2 .39.2 3 0 .7 11.8 5 .2 81 Tests o f H ypotheses As noted in Chapter I, tw o broad hypotheses were developed to guide pursuit o f the objectives o f this study. Hypothesis 1, reproduced below, focuses on visitors vs. nonvisitors; Hypothesis 2, presented later in this Chapter, focuses on repeat vs. first-time visitors For purpose o f statistical testing o f the two broad hypotheses, each was disaggregated into a series o f subhypotheses. Each o f these is introduced below followed by the results obtained from testing it. Numerical results for the set o f subhypotheses associated with Hypothesis 1 are presented in Table 12 and those related to Hypothesis 2 are provided in Table 13. Hypothesis 1. There are no significant differences in destination-decision making between visitors and nonvisitors with respect to: familiarity with the destination, state/province o f residence, distance to the destination, readership o f advertising literature, elapsed time between inquiry and receipt o f information, interest in the advertising literature, quality o f the literature, and socioeconomic and demographic characteristics o f inquirers. Hypothesis 1-1 Familiarity with Frankenmuth does not influence the decision to visit Frankenmuth To operationalize this hypothesis, familiarity was defined by whether or not respondents had made a prior visit to Frankenmuth. The chi-square statistic for visitors and nonvisitors with respect to the familiarity variable is significant. Thus, the null hypothesis is rejected at p < .05. This result indicates that familiarity with the destination 82 Table 12. Significance Test Results for Selected Variables - Visitor vs. Non-visitors I Iy]X)thcsis N um ber 1 -1 1 -2 1 -3 1 -4 1 -5 1 -6 1 -7 1 -8 1 -8.1 1 -8 2 1 -8 .3 1 -8.4 1 -8 .5 * V a ria b le s Tested F a m ilia rity w ith d e s tin a tio n 1 R e sid e n c e ( % M ic h ig a n )2 M ile s a w ay from h o m e 2 R e a d b ro c h u re K lap scd tim e b e tw e e n in q u iry &. re c e ip t o f b ro c h u re In te re st in b ro c h u re ’ Q u a lity o f b ro c h u re '1 S o c io ec o n o m ic & d e m o g ra p h ic G e n d e r ( % M a le) K m p lo y m en t (% F u ll tim e ) E d u ca tio n A ge In c o m e ’ V isito rs N o n -v isito rs 6 1 .2 % 57% 196.98 m ile s 9 9 .6 % 5 0 .6 % 33% 336.71 m ile s 9 6 .2 % 9.44 days 5 67 6.05 17.92 days 5 .39 5.74 2 7 .5 % 514% 14.02 years 4 4 .6 4 years 5.24 2 9 .1 % 5 4 .4 % 14 2 6 y ears 4 7 .0 5 years 5.20 c h i-sq u a re S ig n ific a n c e I - tc s t P ro b a b ility 0.01* 0.00* 0.00* 0.01* 0.00* 0.01* 0.00* 0 .6 9 0 .4 4 0.03* 0.04* 0 .8 3 in d ic a te s sig n ific a n c e a t CX= .05 1. Q u e s tio n 2, "I la v e y o u e v e r v isite d F runk en in u lh ?" P rccen t r e sp o n d in g "Yes." 2. " R esid en ce" an d " m ile a g e a w a y from h om e" w e r e o b ta in e d v ia a p p lic a tio n o f A u to m a p PC s o ftw a r e to z ip c o d e s u s e d to m a il q u e stio n n a ir e s. 3. T h e r a tin g s w e r e b u sed on a se v e n p o in t sc a le , w ith 1 r e p r e s e n tin g "greatly d e c r e a s e d m y in terest" an d 7 r e p r e s e n tin g " greatly in crea sed m y in terest." 4 . T h e r a tin g s w e r e b a se d o n a se v e n p o in t s c a le , w ith 1 r e p r e se n tin g "terrible" an d 7 r e p r e s e n tin g " ex cellen t." 5. T h e ra tin g w a s b a se d on ord in al s c a le from 1 to 12, w ith 1 - u n d er $ 1 5 ,0 0 0 ; 2 - $ 1 5 ,0 0 0 to $ 1 9 ,9 9 9 ; 3 - $ 2 0 ,0 0 0 to $ 2 4 ,9 9 9 ; 4 - $ 2 5 ,0 0 0 to $ 3 4 ,9 9 9 ; 5 - $ 3 5 ,0 0 0 to $ 4 9 ,9 9 9 ; 6 - $ 5 0 ,0 0 0 to $ 7 4 ,9 9 9 ; 7 - $ 7 5 ,0 0 0 to $ 1 0 4 ,9 9 9 . 8 - $ 1 0 5 ,0 0 0 to $1 1 9 ,9 9 9 ; 9 - $ 1 2 0 ,0 0 0 to $ 1 3 4 ,9 9 9 ; 10 - $ 1 3 5 ,0 0 0 to $ 1 4 9 ,9 9 9 ; 1 1 - $ 1 5 0 ,0 0 0 to $ 2 9 9 ,9 9 9 ; 12 - $ 3 0 0 ,0 0 0 or m ore. 83 has a significant influence on choosing it. In other words, people have a tendency to select destinations that they have visited before. Hypothesis 1-2 State o f residence does not influence the decision to visit Frankenmuth. State o f residence was defined as the percentage o f Michigan residents in the visitor and nonvisitor responding populations. Michigan residents accounted for 57% o f visitors and 33% o f nonvisitors, and, as indicated in Table 12, the related chi-square statistic is significant. Thus, the null hypothesis is rejected at p<.05 indicating that there is a significant difference in decision making by residence status between visitors and nonvisitors. Hypothesis 1- 3 Distance does not influence the decision to visit Frankenmuth. This hypothesis was operationalized by applying the t test to the mean distance between residence and Frankenmuth for visitors vs. nonvisitors. The mean o f mileage away from home is 196.98 miles for visitors and 336.71 miles for nonvisitors. The average trip mileage away from home is 139.73 miles shorter for visitors compared with nonvisitors' trip mileage. This null hypothesis is rejected at p < 0 5 indicating that there is a significant difference in mileage away from home between visitors and nonvisitors 84 Hypothesis 1-4 Reading the brochure mailed to respondents did not influence respondents' decision to visit Frankenmuth. Ninety-nine percent o f visitors reported that they read the brochure after receiving it while 96% o f nonvisitors indicated that they read the brochure after receiving it. Readership o f the brochure had a slight positive influence on the decision to visit Frankenmuth and this difference is statistically significant as indicated in Table 12. Thus, the null hypothesis is rejected at p < .05. Reading the brochure had a positive influence on respondents' decision to visit Frankenmuth. Hypothesis 1- 5 Elapsed time between inquiry and receipt o f the brochure did not influence the decision to visit Frankenmuth. The average elapsed time for visitors is 9.44 days; 17.92 days for nonvisitors. The difference is 8.47 days. While the survey results in this case are strong, there is reason for caution in their interpretation. Since the Frankenmuth CVB seeks to respond promptly to all requests for information, the elapsed time reported by non-visitors appears excessive. Quite possibly non-visitors' perceptions o f elapsed time were unrealistically high as a result o f a rationalization process for their decision not to visit Frankenmuth. As indicated in Table 12, the mean elapsed time is statistically different for visitors and non-visitors. This null hypothesis is rejected at p< 05. Its rejection indicates that elapsed time between request for and receipt o f travel information has a significant role in trip decision-making 85 Hypothesis 1- 6 Respondents interest in the brochure did not influence their decision to visit Frankenmuth A 7 point Likert scale was used to measure interest in the brochure. The reported means for visitors and nonvisitors were 5.67 and 5.39 respectively. As indicated in Table 12, these means were found to be statistically different. Thus, the null hypothesis is rejected at p < 0 5 indicating that interest in the brochure did influence the decision to visit Frankenmuth. Hypothesis 1- 7 Perceived quality o f the brochure received by respondents did not influence their decision to visit Frankenmuth. A Likert scale question with 1 being terrible and 7 excellent was used to measure respondents' perception o f the quality o f brochure they received. The reported means for visitors and nonvisitors were 6.05 and 5.74 respectively. T test results reported in Table 12 indicate that these means are statistically different. Thus, the null hypothesis is rejected at p < .05 indicating that quality o f the brochure did influence the decision to visit Frankenmuth. Hypothesis 1 - 8 Socioeconomic and demographic variables did not influence the decision to visit Frankenmuth. 86 Hypothesis 1-8.1. 1-8.2. 1-8.3. 1-8 4. 1-8 5 The variables: (1) gender, (2) employment status, (3) educational background, (4) age, and (5) household income did not influence the decision to visit Frankenmuth None o f these null hypotheses, except 1-8.4 (age), are rejected at p < .05. The results show that gender, employment status, educational background, and household income are irrelevant to the decision to visit Frankenmuth. The null hypothesis related to age is rejected at p < .05 indicating age o f respondents did influence the trip decision. Hypothesis 2: There are no significant differences in destination decision making between parties who are familiar with a travel destination and those who are first-time visitors in terms of: residence status, distance to the destination, medium/media used, travel behavior, travel satisfaction, brochure readership, on site brochure consultation, elapsed time, spending patterns, brochure quality; and socioeconomic and demographic characteristics o f inquirers. This broad hypothesis was disaggregated into a series o f sub-hypotheses to facilitate statistical analyses. The related data and test results are summarized in Table 13. Each hypothesis is presented below with an accompanying brief discussion o f findings. Hypothesis 2-1. 2-2 There is no significant difference in destination decision making between repeat visitors and first-time visitors in terms o f ( 1) state/province o f residence or ( 2 ) distance from Frankenmuth. 87 Table 13. Significance Test Results for Selected Variables - Repeat vs. First-time Visitors. H y p o th e s is N um ber 2 2 2 2 2 -1 -2 -3 -3.1 -3 .2 2 -3 .3 2 2 2 2 2 2 2 2 -3 .4 -3 .5 -3 .6 -3 .7 -3 .8 -3 .9 - 3 .1 0 -3 .1 1 2 - 3 .1 2 2 -4 2 -4.1 2 -4 .2 2 -4 .3 2 - 4 .4 2 -4 .5 2 -5 2 -5.1 2 -5 .2 2 -5 .3 2 -5 .4 2 -6 2 -6.1 2 -6 .2 2 2 2 2 -7 -7.1 -7 .2 -7 .3 V a r ia b le s T ested R e s id e n c e (% M ic h ig a n ) M ile s a w a y from h o m e M e d ia so u r c e s u se d N e w s p a p e r a r ticle N e w s p a p e r a d v e r tise m e n t M a g a z in e a rticle M a g a z in e a d v e r tise m e n t R a d io a d v ertisem en t T e le v is io n a d v e r tise m e n t l-'r ie n il/rela tiv c/co -w o rk er l-'rankenm uth area b u s in e s s Travel agen t M ic h ig a n Travel B u rea u R e g io n a l tou rist a s s o c ia tio n B r o ch u r e Travel b e h a v io r P u r p o se o f trip P arty s iz e (P erso n ) L e n g th o f stay (n ig h ts ) D a y trip (% Y e s ) A c c o m m o d a tio n s u se d (H o te l or M o te l) S a tis fa c tio n Trip e x p e r ie n c e 1 In ten tio n to v is it a g a in 1 R e c o m m e n d to o th e rs Travel d e stin a tio n 2 R e a d e r sh ip R ea d C o n su lte d on trip K la p sed tim e (d a y s ) in q u ir y - receip t r e c e ip t - v isit in q u iry - visit (c o n tin u e d on n ex t p a g e ) R ep en t V isito r s 73% 1 5 0 .9 8 m ile s 4 .6 % 3.1% 8 .5% P ir st-tim e V isito r s 33% 2 6 6 .7 1 m ile s c h i-sq u a r e S ig n ific a n c e 0 .0 0 * 0 .0 0 ' 3 .1% 1.6% 0 .3 7 0 .3 5 4 .7 % 0.1 3 0 .0 0 ' 1.5% 5.4% 2 0 .0 % 8 .5% 2 .3 % 10,0% 4 .6 % 16.9% 1.6% 16% 1.6% 4 3 .8 % 4 .7 % 6 .3 % 10 .9 % 4 .7 % 7 .8 % 0.81 0 .1 0 0 .0 1 ' 0 .1 3 0 .3 3 0 .6 2 0 .6 8 8 9 .3 % 92.6% . 0 .4 9 3 .6 3.1 1.06 13 .9 % 10.0% 0 .7 6 35 9% 9 6 .8 % 3.71 4 .5 3 8 1 .1 % 6 .0 4 9 6 .1 % 3 .9 8 4 .1 0 94.6% . 0 .0 1 ' 0 .0 4 ' 0 .0 1 ' 0 .0 0 ' 0 .5 2 002' 0 .0 0 ' 0 .0 1 ' 5 .6 3 100% 9 9 .3 % 7 8 .4 % 9 4 .6 % 9 .7 3 2 8 .4 3 3 5 .9 0 8 .9 7 2 3 .6 0 3 2 .5 1 T -tc st P r o b a b ility 0 .0 2 ' 0 .4 2 0 .0 2 ' 0 .6 4 0 .3 2 0 .3 2 88 Table 13 (cont’d) H y p o th e s is N um ber 2 2 2 2 2 2 2 -8 -8.1 - 8 .2 - 8 .3 -8 .4 - 8 .5 - 8 .6 V a r ia b le s T e ste d E x p e n d itu r e s ( D o lla r s ) L o d g in g C a m p in g fe e G ilt s , c ra fts, so u v e n ir s G r o c e r ie s R esta u ra n ts G a s o lin e , v e h ic le rep air 2 - 8 .7 G u id e d tou rs 2 - 8 .8 O th e r ite m s Total 2 2 2 2 2 -9 -9.1 -9 .2 -9 .3 -9 .4 2 - 9 .5 2 - 9 .6 2 - 9 .7 2 - 9 .8 2 2 2 2 2 2 -1 0 -1 0 .1 - 1 0 .2 - 1 0 .3 - 1 0 .4 - 1 0 .5 B roch u re u s e f u ln e s s 2 in te r e stin g to read 2 a ttr a c tiv e n e s s 2 in c r e a se d in terest in d e s tin a tio n 2 in flu e n c e d d e c is io n 2 in c r e a s e d e x p e n d itu r e 2 accuracy2 q u a lity 2 S o c io e c o n o m ic & d em o g ra p h ic G e n d e r (% M a le ) E m p lo y m e n t (% L u ll tim e ) E d u ca tio n A ge In co m e R ep ea t V isito r s 7 2 .2 6 0 .7 7 1 3 8 .2 5 1 4 .2 4 7 8 .6 0 1 2 .1 3 F ir st-tim e V isito r s c h i-sq u a re S ig n ific a n c e 9 3 .7 6 0 18 1 6 0 .5 2 9 .6 7 T -tc st P ro b a b ility 0 .2 4 0 33 0 .5 3 0 .1 3 0 .2 4 0 .4 4 7 .2 0 3 2 9 .1 8 1 0 6 .9 7 1 4 .4 2 1.44 1 4 .7 0 4 0 1 .6 8 5 .7 9 5 .6 3 5 .8 3 5 .7 7 5 .6 6 5 .8 2 0 .9 1 0 .8 7 0 .9 0 5 .5 8 4 .1 8 3 .3 0 6 .1 7 6 .0 8 5 .8 6 4 90 3 .0 8 6 .0 3 6 .0 2 0 .0 6 0 .0 1 * 0 .3 6 3 1 .7 % 2 3 .8 % 5 2 .3 % 1 3 .7 4 4 4 .4 7 5 .5 7 5 0 .5 % 14.45 4 5 .8 2 5 .0 3 3 .7 4 0 .1 9 0 .1 9 0 .2 3 0 .3 5 0 .6 0 0 .2 3 0 67 0 .0 3 * 0 .4 6 0 .1 8 . * in d ic a te s sig n ific a n c e a t < .05 1 The ra tin g s w e re b a se d on a liv e p o in t L ik e rt sc a le w ith a n e u tra l m id p o in t, "5" b e in g p o sitiv e a n d " 1" n e g a tiv e 2 The ra tin g s w e re b a se d on a se v e n p o in t scale. In th e o rd e r th at th e v a ria b le s a re liste d , p o les o n th e L ik e rt sc a le s a re ; 1 n o t a t all - 7 very u se fu l; 1 very u n in te re s tin g - 7 very in te re stin g , etc. 3 H ie ra tin g w a s b a se d o n a 12 p o in t o rd in a l scale, w ith 1 - u n d e r $ 1 5 ,0 0 0 ; 2 - $ 1 5 ,0 0 0 to $ 1 9 ,0 9 9 ; 3 - $ 2 0 ,0 0 0 to $ 2 4 ,9 9 9 ; 4 - $ 2 5 ,0 0 0 to $ 3 4 ,9 9 9 ; 5 - $ 3 5 ,0 0 0 to $ 4 9 ,9 9 9 ; 6 - $ 5 0 ,0 0 0 to $ 7 4 ,9 9 9 ; 7 - $ 7 5 ,0 0 0 to $ 1 0 4 ,9 9 9 ; 8 - $ 1 0 5 ,0 0 0 to $ 1 1 9 ,9 9 9 , 9 - $ 1 2 0 ,0 0 0 to $ 1 3 4 ,9 9 9 ; 10 - $ 1 3 5 ,0 0 0 to $ 1 4 9 ,9 9 9 ; 11 - $ 1 5 0 ,0 0 0 to $ 2 9 9 ,9 9 9 ; 12 - $ 3 0 0 ,0 0 0 o r m ore. 89 The null hypothesis regarding state/province o f residence is rejected at p < .05 indicating that there is a significant difference between repeat visitors and first-time visitors with respect to where they reside. The data presented in Table 13 show that 73% o f repeat visitors were from Michigan; only 33% o f the first-time visitors w ere from Michigan. Data for the mileage variable indicates that the null hypothesis is rejected at p < .05 indicating that distance is a significant variable distinguishing repeat from first-time visitors The average distance traveled to Frankenmuth is 150.98 miles for repeat visitors and 266.71 miles for first-time visitors Hypotheses 2-3.1 - 2 .- 3 .1 2 Repeat and first-time visitors exhibit no significant differences in use o f the following 12 sources o f travel information: ( 1) newspaper article, ( 2 ) newspaper advertisement, (3) magazine article, (4) magazine advertisement, (5) radio advertisement, ( 6 ) television advertisement, (7) friend/relative/co-worker, ( 8 ) Frankenmuth area business, (9) travel agent, (10) Michigan Travel Bureau, ( 1 1 ) regional tourist association, and ( 1 2 ) the brochure. Three o f these null hypotheses (2-3 .4, 2-3 .7, and 2-3 .12) are rejected at p < .05 indicating that magazine advertisement, friend/relative/co-worker, and the brochure were used differently by repeat and first-time visitors The data show that magazine articles and brochures have a significant positive influence on repeat visitors. First-time visitors are significantly more likely to rely on information from the friend/relative/co-worker source Frequency o f reported use was similar for the other nine sources listed across repeat and first-time visitors. 90 Hypothesis 2-4 There are no significant differences in travel behavior between repeat and first-time visitors with respect to: (1) purpose o f trip, (2) party size, (3) length o f stay, and (4) accommodations used Hypothesis 2-4.1 There is no significant difference in "purpose-of-trip" between repeat and first-time visitors. Eighty-nine percent o f repeat visitors reported that recreation/pleasure was their trip purpose; 93% o f first-time visitors indicated that purpose o f trip was recreation or/and pleasure. Results presented in Table 13 indicated that this observed difference is not statistically significant. Thus, this null hypothesis can not be rejected at p < .05 which indicates that there is no significant difference with respect to "purpose-of-trip" between repeat and first-time visitors. Hypothesis 2-4.2 There is no significant difference in mean party size between repeat and first-time visitors The mean travel party size is 3.6 persons for repeat visitors and 3.1 persons for first-time visitors. Repeat visitors have slightly more persons accompanying them to Frankenmuth, and T test results indicate that this difference is statistically significant. Thus, this null hypothesis is rejected at p < 05 indicating that there is a significant difference in mean travel party size between repeat and first-time visitors. 91 Hypothesis 2-4 3 There is no significant difference in length o f stay in Frankenmuth between repeat and first-time visitors Average length o f stay in the destination area is 1.7 days for repeat visitors, 2 days for first-time visitors While first-time visitors mean stay is slightly longer than that o f repeat visitors, the difference is too slight to be statistically significant. Thus, this null hypothesis can not be rejected at p < .05 indicating that length o f stay in the destination area can not be used to distinguish between repeat and first-time visitors. Hypothesis 2-4.4 There is no significant difference in choosing Frankenmuth as day-trip destination betw een repeat and first-time visitors. R epeat visitors are much more likely to choose Frankenmuth as a day-trip destination than are first-time visitors. The percentages choosing Frankenmuth as a day-trip destination is 36% for repeat visitors and 14% for first-time visitors. As indicated in Table 13, this difference is statistically significant. Thus, this null hypothesis is rejected at p < .05 indicating that there is significant difference in choosing Frankenmuth as a day trip destination between repeat and first-time visitors. Hypothesis 2-4.5 There is no significant difference in accommodations chosen between repeat and first-time visitors. 92 Ninety-seven percent o f repeat visitors and 96% o f first-time visitors who stayed overnight stayed in a hotel or motel The calculated chi-square statistic was not found to be significant, thus this null hypothesis can not be rejected at p < .05 indicating that there is no significant difference in choice o f accommodations between first-time and repeat visitors Hypothesis 2-5 There are no significant differences between repeat and first-time visitors with respect to how they rated their trip experience, intention to visit Frankenmuth again, inclination to recommend the destination, and ranking o f Frankenmuth as a attractive travel destination. Hypothesis 2-5.1 There is no significant difference in trip experience rating between repeat and first-time visitors. Satisfaction with the trip was measured via a five-point Likert scale question with 1 representing "much worse than expected" and 5 representing "much better than expected". The mean trip experience rating was 3.98 for first-time and 3.71 for repeat visitors, respectively. Data in Table 13 indicate that these means are statistically different Thus, this null hypothesis is rejected at p < .05 indicating that there is a significant difference in trip satisfaction between repeat and first-time visitors with respect to their trip experience. 93 Hypothesis 2-5.2 There is no significant difference between repeat and first-time visitors' intentions to visit Frankenmuth again. A five-point Likert scale question was used to measure intention to visit again with 1 representing "certain to not visit again" and 5 representing "certain to visit again." Although first-time visitors rated their trip experience higher than repeat visitors, their mean intention to visit again score was less (4.10) than that reported for repeat visitors (4.53). Since this difference is a statistically significant, this null hypothesis is rejected at p < .05 indicating that there is a significant difference in intention to visit Frankenmuth again between repeat and first-time visitors. Hypothesis 2-5.3 There is no significant difference in intention to recommend Frankenmuth as a travel destination between repeat and first-time visitors. This null hypothesis is rejected at p < .05 indicating that there is a significant difference in intention to recommend Frankenmuth as a travel destination between repeat and first-time visitors. First-time visitors are more likely to recommend Frankenmuth than are repeat visitors. Hypothesis 2-5.4 There is no significant difference in ranking o f Frankenmuth as an excellent destination between repeat and first-time visitors. 94 A seven-point Likert scale was used to rank Frankenmuth's appeal as a travel destination, with 1 representing "terrible" and 7 representing "excellent". The reported mean was 6.04 for repeat visitors and 5.63 for first-time visitors. As indicated in Table 13, these means are statistically significant. Thus, this null hypothesis is rejected at p < .05 indicating that repeat visitors rank Frankenmuth higher as a travel destination than do first-time visitors. Hypotheses 2-6.1. 2-6.2 There is no significant difference between repeat and first-time visitors with respect to ( 1 ) reading the brochure they received or ( 2 ) consulting it during their stay in Frankenmuth. The null hypothesis with respect to reading the brochure can not be rejected at p < 05 This indicates that there is no significant difference between repeat and first-time visitors with respect to reading the brochure before their visit. Ninety-nine percent o f repeat visitors read the brochure after it was received; 1 0 0 % o f first-time visitors reported that they read the brochure. Thus, the null hypothesis regarding brochure consulting is rejected at p < .05. First-time visitors rely more on the brochure than do repeat visitors as can be seen in Table 13. Almost 95% o f them consulted the brochure on-site as opposed to 78.4% for repeat visitors, and, as indicated in Table 13, this difference is statistically significant. Hypotheses 2-7.2. 2-7.2. 2-7.3 Between repeat and first-time visitors to Frankenmuth, there is no significant difference with respect to elapsed time ( 1 ) between inquiry and receipt o f information, (2) between receipt o f information and visit, and (3) between inquiry and visit 95 Based upon the data provided in Table 13, none o f these hypotheses can be rejected at p < 05. Thus, significant differences were found between repeat and first-time visitors with respect to the above measures o f elapsed time Hypotheses 2-8.1. 2-8.2. 2-8.3. 2-8.4. 2-8.5. 2-8.6. 2-8.7. and 2-8.8 There is no significant difference in the amount o f money spent on: (1) lodging, (2) camping fees, (3) gifts, (4) groceries, (5) restaurant meals, ( 6 ) gasoline and vehicle-related items, (7) guided tours, and ( 8 ) all other items, between repeat and first-time visitors during their stay in Frankenmuth. For repeat visitors, the average amount o f money spent on lodging, camping, gifts, groceries, restaurant meals, gasoline and vehicle-related items, guided tours, and all other items was $72.26, $0.77, $138.25, $14.24, $78.60, $12.33, $3.74, and $7.20, respectively For first-time visitors, the average amount o f money spent in these same categories was $93.76, $0.18, $160.52, $9.67, $106.97, $14.42, $1.44, and $14.70, respectively. The average total expenditure for repeat visitors is $329.10 and $401.68 for first-time visitors. Statistically speaking, there is no significant difference in total expenditure between repeat and first-time visitors, although first-time visitors spent $80 dollars more per visit than did repeat visitors. Thus, none o f these hypotheses can be rejected at the p < .05 indicating that there are no significant differences across expenditures by individual category or in total expenditures between repeat and first-time visitors during their stay in Frankenmuth 96 H ypotheses 2-9.1. 2-9.2. 2-9.3. 2-9.4. 2-9.5. 2-9.6. 2-9.7. and 2-9,8 Between repeat and first-time visitors, there are no significantly differences in rating o f the following brochure related variables: ( 1 ) usefulness, ( 2 ) interesting to read, (3) attractive in design, (4) increased or decreased interest in visiting, (5) influenced decision to visit, ( 6 ) influenced expenditures, (7) accuracy, and ( 8 ) overall quality. For all o f these variables, seven-point Likert scales were used to obtain rankings. On each scale; 1 represents the lowest ranking and 7 represents the highest ranking (See Table 13, note 2 for more detail or refer to question 20 -27 in the appendix). For repeat visitors, the average score obtained for usefulness, interesting, attractiveness, interest, influenced expenditures, accuracy, and quality are 5.79, 5.63, 5.83, 5.58, 4.18, 3.30, 6.17 and 6.08, respectively. Similarly, the average scores assigned these variables by first-time visitors are 5.77, 5.66, 5.82, 5.86, 4.90, 3.08, 6.03 and 6.02, respectively. Statistical test results presented in Table 13 show that there is only one null hypothesis (2-9.5) which can be rejected at p < .05. Thus, it appears that the brochure did have more influence on first-time visitors' decision to visit Frankenmuth but no significance differences in rankings were found between the two groups with respect to the other six brochure related variables examined. Hypothesis 2-10 I. 2-10 2. 2-10 3. 2-10 4. and 2-10.5 There are no significant differences between repeat and first-time visitors with respect to (1) gender, (2) employment status, (3) education level, (4) age, and (5) income level. 97 None o f these hypotheses, except hypothesis 2-10.3 (education level), can be rejected at p < .05 indicating that there are no significant differences between repeat and first-time visitors with respect to gender, employment status, age, and income level The null hypothesis with respect to education level is rejected at p < 05 indicating that there is a significant difference between repeat and first-time visitors with respect to education level achieved. The average education level o f repeat visitors is 13 .74 years; the mean for first-time visitors is 14.45 years. First-time visitors have achieved a higher educational level than repeat visitors. Results o f Logistic Regression in Prediction Results from the logistic regression analysis (LRA) will be presented in the following sections. One o f the objectives o f developing a logistic regression model was to predict the probability o f a visit to Frankenmuth. Another objective o f this study was to use LRA to predict the propensity o f repeat visits to Frankenmuth from among current visitors who requested information from the FCCVB. This was accomplished by the application o f two different models. The first model was used to predict the probability o f visits to Frankenmuth by inquirers in accordance with the third objective in this study. The first model includes the following predictors: individual sociodemographic characteristics, elapsed time between inquiry and receipt o f information from the FCCVB, familiarity with the destination area, brochure (quality and interest in reading it), and actual readership o f the brochure. This 98 first model is titled "Model 1: Prediction o f the propensity to visit Frankenmuth", and is explained in detail on page 110 (Equation 1). In order to evaluate the performance o f the different variables in predicting the propensity o f a re-visit to Frankenmuth, the following four subgroups o f variables were considered: ( 1) sociodemographic characteristic variables, ( 2 ) brochure related variables, (3) satisfaction variables, and (4) travel behavior variables. The application o f these four subgroups o f variables derived from the fourth objective o f this study. This second model is titled "Model 2: Prediction o f repeat visits to Frankenmuth." It is presented in detail on page 127 (Equation 2). As discussed in Chapter II, LRA is the best tool to predict a dichotom ous dependent variable from a combination o f independent variables. The objective o f LRA is to find the most parsimonious model to predict the probability o f an event occurring. The approach used to develop this model entailed the following four procedures: ( 1 ) a forward stepwise process, (2) an examination o f the goodness o f fit o f the model, (3) testing o f the estimated significance o f parameters, and (4) interpreting the estimated parameters. Upon completing a best model, these procedures were used for predicting visits to Frankenmuth (M odel 1) and for predicting repeat visits (Model 2). Forward Stepwise Procedure According to Schroeder (1983), there are two general approaches to building a regression model: 1) preselection o f variables based on theory and results from previous research in which these variables were explored; or 2 ) in situations where prior related 99 research is not available and where related theory is not definitive, stepwise procedures are used to select significant variable(s) for inclusion in the model. The stepwise method is incremental in that independent variables are explored one at a time for their contribution to explaining variation in the dependent variable. After a variable is entered into the model, the stepwise procedure checks to see if previously entered variables can be removed without loss in the model's explanatory power. The procedure continues until no more variables can be entered or removed to improve the model's explanatory performance. The stepwise procedure uses the score statistic (p= .05) for inclusion and the likelihood-ratio (p=. 1 0 ) as the test for removal o f a variable from the model. In other words, the forward stepwise procedure entails a systematic evaluation o f the interactions among independent variables and a dependent variable If the addition o f a multiplicative interaction term to the logistic regression model is associated with coefficient changes in other predictor variables, an interaction between the new variable and one or more o f the other predictors is suspected. Interactions are evaluated by comparing models with and without selected variables included. If a statistically significant improvement occurs (p < .05) with the interaction term, the variable o f focus remains in the model. If statistically significant changes in predictor coefficients are observed, the model does not contain interaction terms. The first variable included in this model is the one with the largest acceptable value for the selection criterion After each variable is entered, all variables in the model are evaluated against a deletion criterion Alternative methods o f entry available in SPSS 100 are ( 1 ) the backward elimination method in which all variables that satisfy the selection criterion are entered simultaneously into the model, and then those meeting the removal criterion are deleted and ( 2 ) the simultaneous or direct method in which all variables that satisfy the selection criterion are entered simultaneously into the model. G oodness-of -Fit in Logistic Rearession In evaluating a linear regression model, researchers need to answer questions such as: How well does the overall model work?; If the overall model works well, how important is each o f the independent variables?; Is the relationship between any o f the variables attributable to random sample variation? If not, how much does each independent variable contribute to the prediction o f the dependent variable?; and finally. Does the form o f the model appear to be correct? In linear regression analysis, the assessment o f the significance o f coefficients is approached by two sums o f squares: total sum o f squares (SSI) and error sum o f squares (SSIs). A third sum o f squares is simply the difference between SSI and SSF that is the regression sum o f squares (SSR SS'/'-SSF) The multivariate /'’ratio, used to test hypotheses H „: R2 = 0 and Hn: Pi = p 2 = ••• = P* = 0, can be calculated as: SSR F k l F / k with d*>fr e,,g . = k, dft = N -k-I and J res ’ where; R~ k squared multiple correlation coefficient number o f independent variables 101 /V sample size The statistical significance (j?) associated with the F ratio indicates the probability o f obtaining an R2 as large as the observed R 2 , or P coefficients as large as the observed P coefficients, if the null hypothesis is true. Usually, if the calculated p is small ( < .05) then the null hypothesis is rejected indicating that there is a relationship between the independent variables and the dependent variable that can not be attributed to chance. If the calculated p is large (> .05) then the null hypothesis can not be rejected, and it is concluded that there is insufficient evidence to be sure that the variance explained by the model is not attributable to random sample variation. The coefficient o f determination, R2, is an indicator to judge whether the relationship is strong enough to be captured in the regression equation. For example, if R2 was 0.14, this means that 14% o f the variability in the dependent variable can be explained by the independent variable. R 2 is used to measure the proportion by which use o f the regression equation reduces the error o f prediction. R2 ranges from 0 (indicating that independent variables do not help at all to predict the dependent variable) to 1 (indicating that independent variables can be used to predict the dependent variable perfectly). R2 can be calculated as SSR SSI] (SST-SSE)SS7] or I-(SSEZSST). The /-'ratio and R 2 can be expressed as function o f one another: J2 (R2/k)//(I-R 2)/N- k-l],an d R2 k i' (kE i N-k-1). In a logistic regression model, the log-likelihood is the criterion used for selecting parameters for the model; as the sum o f squared errors is the criterion for selecting parameters in the linear regression model. Likelihood is the probability o f the observed result given the estimates o f the population parameters. 102 The statistical package usually presents the log-likelihood multiplied by -2 (e.g., -2LL) instead o f log-iikelihood itself. The reason for using minus twice its log is to create an approximate X2 (chi-square) distribution to test the hypothesis in order to investigate w hether or not a particular independent variable should be included in the LRA model The value o f -2LL for a logistic regression with only the constant included is called "Initial Log Likelihood Function -2 Log Likelihood" (See Table 14). This initial logistic regression is commonly designated as D. For a dichotomous dependent variable (coded as 0 or 1), if nv , is the number o f cases for which Y P(Y I) 7, N is the total number o f cases, and nv , N is the probability that y is equal to 1, then I) -2{(nv , ) / n / P ( Y 1)1 ‘ In/1-P(Y 1)/ -2 f(ny ,) h ilP ( Y I)J t (ny J In /P (Y ())/ The value o f -2LL for the logistic regression model that includes independent variables and the constant is designated as "-2 Log Likelihood" The statistic I) is called "deviance" chi-square by Hosmer and Lemeshow (1989) or "deviation" chi-square by M enard (1994), and it is as an indicator o f the model's goodness o f fit with all o f the independent variables in the equation. The D statistic from logistic regression plays the same role as the errorsum o f squares (SSE) plays in linear regression. To estimate the significance o f an independent variable, we can compare the value o f D without and with the independent variable in the equation. The change in D ,due to the inclusion o f the independent variable in the model, called "Model Chi-Square" under the chi-square column in SPSS output (See Table 14), is denoted as G. The larger the 103 Table 14 Logistic Regression Procedure with Dependent Variable Visit or not (Full Model) T o ta l n u m b e r o f c a se s: N u m b e r re je c te d b e c a u s e o f m is s in g d a ta : N u m b e r o f c a s e s in c lu d e d in th e a n a ly s is : 595 176 419 D e p e n d e n t V a ria b le E n c o d in g : O r ig in a l In te rn a l V a lu e V a lu e 0 1 0 1 D e p e n d e n t V a ria b le : V is ito r (V is it F ra n k e n m u th a fte r re q u e s tin g in f o rm a tio n ) B e g in n in g B lo c k N u m b e r 0 . In itia l L o g L ik e lih o o d F u n c tio n -2 L o g L ik e lih o o d 5 7 1 .3 4 8 7 7 ♦ C o n s ta n t is in c lu d e d in th e m o d e l. B e g in n in g B lo c k N u m b e r 1. M e th o d : E n te r V a r ia b lc ( s ) E n te re d o n S te p N u m b e r E s tim a tio n te r m in a te d a t ite ra tio n n u m b e r 4 b e c a u se L o g L ik e lih o o d d e c re a s e d b y le s s th a n .01 p e rc e n t 4 7 8 .7 6 2 -2 L o g L ik e lih o o d G o o d n e s s o f F it 4 3 5 .8 6 4 M o d e l c h i- s q u a r e Im p ro v e m e n t c h i- s q u a re 9 2 .5 8 7 9 2 .5 8 7 df 11 11 S ig n ific a n c e .0 0 0 0 .0 0 0 0 C la s s ific a tio n T a b le fo r V is ito r P re d ic te d O b ser v ed n o n -v is ito r s v isito r s n o n -v isito r v isito r s P ercen t C orrect 188 53 78.01% 69 109 61 .2 4 % O v e ra ll 7 0 .8 8 % 104 Table 14 (cont'd). Variables in the Equation B V a ria b le D 1S T RESI QUAL ELA PRI W ORK EDUC IN C O M E AGE FA M LAR READ GENDER C o n s ta n t -.0 0 1 8 .7 1 6 6 .2961 -.0 6 0 0 -.3 1 3 6 - .0 7 7 5 -. 1363 -.0 0 7 1 - .3 7 2 7 5 .5 2 3 3 .0 1 8 0 -4 .9 1 1 5 S .E . W a ld df S ig .0 0 0 0 .3031 . 1159 .0131 .2 4 2 0 .0 4 2 7 .2651 .0 0 9 4 .2 5 5 6 13.5051 .2 6 4 5 1 3 .5 5 1 7 7 .5 2 3 0 5 .5 9 1 5 6 .5 2 3 4 2 0 .9 3 4 8 1 .6 7 9 2 3 .2 9 2 6 .2 6 4 4 .5 6 0 8 2 .1 2 7 4 .1 6 7 3 .0 0 4 6 .1 3 1 4 1 1 1 1 1 1 1 1 1 1 1 1 .0061 .0 1 8 0 .0 1 0 6 .0 0 0 0 .1 9 5 0 .0 6 9 6 .6071 .4 5 3 9 .1 4 4 7 .6 8 2 6 .9 4 5 9 .7 1 7 0 R -.0 9 8 3 .0 7 9 3 .0 8 9 0 -.1 8 2 0 .0 0 0 0 - .0 4 7 6 .0 0 0 0 .0 0 0 0 -.0 1 4 9 .0 0 0 0 .0 0 0 0 E x p (B ) .9 9 8 2 2 .0 4 7 5 1 .3 4 4 6 .9 4 1 7 .7 3 0 8 .9 2 5 4 .8 7 2 6 .9 9 3 0 .6 8 8 8 2 5 0 .4 6 7 7 1.0181 D 1ST : D is la n c c b e tw e e n F r a n k e n m u th a n d re s id e n c e , c o d e d a s re a l m ile a g e . R E S I: R e s id e n c e s ta tu s , c o d e d a s 1= M ic h ig a n a n d 0 = n o n -M ic h ig a n . Q U A L : O v e ra ll q u a lity o f b ro c h u re . 7 -p o in t L ik e rt sc a le c o d e d a s 1 (te rrib le ) to 7 (e x c e lle n t). E L A P R I: E la p s e d tim e b e tw e e n in q u iry a n d re c e ip t o f in f o rm a tio n , c o d e d a s a c tu a l d a y s .. W O R K : E m p lo y m e n t s ta tu s , c o d e d a s 1 = full tim e a n d 0 = o th e r. E D U C : E d u c a tio n a l b a c k g ro u n d , c o d e d a s a c tu a l y e a rs a tte n d e d sc h o o l IN C O M E : H o u s e h o ld in c o m e , c o d e d a s 1= e q u a l to o r m o re th a n $ 3 5 ,0 0 0 a n d 0 = le s s th a n $ 3 5 ,0 0 0 . A G E : Y e a rs o ld . F A M L A R : V is ite d F ra n k e n m u th b e fo re , c o d e d as 1 = y es a n d 0 = no. R E A D : R e a d th e m a te ria l a fte r re c e iv in g . c o d e d a s 1 = y e s a n d 0 = no. G E N D E R : 1 = m a le a n d 0 = fe m a le . 105 calculated G the greater the improvement in goodness-of-fit contributed by the associated independent variables. For the logistic model, G is analogous to the multivariate / 'te s t in linear regression. G is used to test the null hypothesis that . If G is (3i = P 2 = statistically significant at the P = P* = 0 .05 level, the null hypothesis is rejected indicating that information about the independent variables allows us to make better predictions with the particular independent variables than we could make without the same independent variables. Testing the Significance o f Parameters in Logistic Regression In logistic regression, there are two ways to test for statistical significance o f the estimated parameters: the coefficient to standard error ratio (CoefT/S.E.) and the maximum likelihood estimation (MLE) chi-square statistic. The Wald statistic, which has a chi-square distribution and is used to test the null hypothesis that the coefficient equals zero, is calculated by dividing the coefficient estimate by its standard error and squaring the result (W .=[B /S.E.]‘ ). For example, the variable RESI's Wald statistic equals (.7166/.3031 )2, or 5.59 (See Table 14). When the ratio (CoefES.E.) approaches 2, which would lead to an approximate level o f significance o f 0.05, there is a case for statistical significance (Hosmer & Lemeshow, 1989). Because the Wald statistic is unreliable with large coefficients, another approach to testing the significance o f an independent variable commonly used is to run tw o models with likelihood-ratio: one with the full model containing all the variables and one with a 106 reduced model without the variables to be tested The difference between the chi-squares for the tw o models represents the change due to the effect o f the individual variable being tested. For example, the difference between the two models as can be seen in Table 14 and Table 15 is the exclusion o f the variable DIST from the full model. The difference is denoted as (/, using the definition that was mentioned earlier. The value o f the G statistic test in this case is G - -2[(-247.448) - (-239.381) = 16.134 which, with 1 degree o f freedom, has a p-value o f p[X: (1) > 16.134] = .0000. Since the p-value is less than .05, there is a significant difference between the full model and the reduced model The full model including the DIST variable yields better predictions than the reduced model. Thus, it is advantageous to include DIST in the model. Table 15. Logistic Regression Procedure with Dependent Variable Visit or not (Reduced M odel) Variable RESI QUAL ELAPRI WORK EDUC INCOME AGE FAMLAR READ GENDER Constant B 1.1096 .3015 -.0595 -.2221 -.0626 -.1247 -.0107 -.2261 5.5164 .0108 -5.7625 SE Wald df Sig .2595 .1141 .0131 .2362 0411 .2583 .0092 .2469 13.5049 .2576 13.5458 18.2862 6.9750 20.5754 .8836 2.3215 .2333 1.3649 .8387 .1669 .0018 1810 1 1 1 1 1 1 1 1 1 1 1 .0000 .0083 .0000 .3472 1276 .6291 .2427 .3598 .6829 .9666 .6705 R .1682 .0929 -.1796 .0000 -.0236 .0000 .0000 -.0000 .0000 .0000 Exp(B) 3.0330 1.3518 .9423 .8009 .9394 .8827 .9894 .7976 248.7433 1.0109 E s tim a te d C o e ffic ie n ts fo r a M u ltip le L o g istic R e g re ss io n M o d e l e x c lu d in g D IS T (D is ta n c e ) V a ria b le * L o g L ik e lih o o d = - 2 4 7 .7 7 8 107 The overall goal o f logistic regression is to obtain the best fitting model with the minimum number o f parameters. According to Hosmer and Lemeshow (1989), the purpose o f logistic regression is to find the model containing only those variables thought to be significant. Then this reduced model is compared with the full model containing all the variables The results o f fitting the reduced model are given in Table 16. The difference between the tw o models as can be seen in comparing Table 14 and Table 16 is the exclusion o f the variables WORK, INCOME, AGE, FAMLAR, READ, and GEND ER from the full model. The difference is G=-2[(-243.637)-(-239.381))=8.512, with 6 degree o f freedom, and a p-value o f p[X : ( 6 ) > 8.512] = .21. Since the p-value is larger than ,05, there is no significant difference between the full model and the reduced model. The reduced model is as good as the full model as a prediction tool Thus, there is no advantage to including WORK, INCOME, AGE, FAMLAR, READ and GEN D ER in the model. Table 16. Logistic Regression Procedure with Dependent Variable Visit or not (Goodness-of-fit Reduced Model) Variable DIST RESI QUAL ELAPRI EDUC Constant B -.0017 .6581 .2885 -.0577 -.0829 -.0413 S.E. Wald df Sig .0006 .2750 .1117 .0129 .0415 1.0358 6.8368 5.7269 6.6661 19.8641 3 9814 .0016 1 1 1 1 1 1 .0088 .0167 .0098 .0000 0460 .9682 * L o g L ik e lih o o d = -2 4 3 .6 .1 7 R -.0919 .0805 .0900 -.1761 -.0587 Exp(B) .9983 1.9311 1.3344 .9440 .9205 108 Interpreting the Estimated Parameters In logistic regression., the coefficient (B) represents the change in the log o f the odds o f an event that is associated with a one unit change in an independent variable. To facilitate their interpretation, coefficients are expressed as Exp(/J), which is obtained by computing the natural antilogarithm o f the coefficient. In Table 16, E xp(/i) represents the factor by which the odds - the probability o f visiting Frankenmuth to the probability o f not visiting Frankenmuth - change as a function o f a one unit change (e.g., one more year o f school-EDUC) in a particular independent variable. If the coefficient is positive, Exp(/i) will be greater than one . For instance, the coefficient for "residence status - RESI" (1 = Michigan resident , 0 = non-Michigan resident) is 6582 (See Table 16), and its Exp(/I) or odds value is 1.9311. This means that respondents who are Michigan residents are almost 2 times more likely to visit Frankenmuth than those respondents who are not Michigan residents On the contrary, if the coefficient is negative, Exp(/J) will be less than one. For example, the coefficient (/i) for "years attending school - EDUC" is -.0829 and its Exp(/i) is .9205. This odds value means that respondents who attended 12 years o f school are almost 10% less likely to visit Frankenmuth than who attended 11 years o f school. When a coefficient equals zero, Exp(/i) is one, the odds are not affected by a one unit change in an independent variable Interpretation in logistic regression involves estimating the effects o f the various independent variables; the direction o f the relationship as well as the magnitude o f the relationship are o f interest as in OLS regression In logistic regression, the direction o f the 109 relationship is indicated by the sign o f the parameter estimate. From the data presented in Table 16, we can see that the reduced form model can be expressed as: logitfl'lsit) -.0413 0017'D IST ■ .6581 'R E S T • ,2885'QUAL -. 0577'ELAPRI -. OS29*EDUC Thus, when REST (residence status) and QUAL (overall brochure quality) increase, the probability o f visiting Frankenmuth will increase. On the contrary, when /9AS7'(mileage between home and Frankenmuth), ELAPRI (elapsed time between inquiry and receipt o f information), and EDUC (years attending school) increases, the probability o f visiting Frankenmuth will decrease. By using the changes in the log odds associated with an independent variable, interpretations about the effects o f that variable can be made. For Equation 1 presented below, each one-unit increase in D /ST is associated with a decrease o f .0017 in Iogil(Vi.sit). Being a resident o f Michigan increases logit(Visit) by .6581. Each one-unit increase in OUAL is associated with an increase o f .2885 in logit(lrisif). Each one-unit increase in ELAPRI is associated with a decrease o f .0577 in logitflrisit). Each one-unit increase in EDUC is associated with a decrease o f .0829 in Iogit(\risit). 1 10 Model 1: Prediction of the Propensity to Visit Frankenmuth In order to select the most significant variables in the decision among inquirers to visit Frankenmuth, stepwise logistic regression was applied. The best prediction model was found to be o f the following form:: !ogit(Visit) -.0413 -.0 0I7*D IST -t .6851*RESI < .2885*QUAL - .0577*171APR1 - .0829*EDIIC -E q u a tio n 1 W here the five independent variables are: DIST. mileage between Frankenmuth and residence REST. Michigan resident = 1 , Non-Michigan resident = 0 OVAL. quality o f brochure, Likert-scale, 1 represents terrible and 7 represents excellent ELAPRI. elapsed time between inquiry and receipt o f information (days) EDUC. years attended school (years) This function shows that increasing trip mileage, years o f attending school, or elapsed time between inquiry and receipt o f information will result in a decrease in logit(visit) (the probability o f visiting Frankenmuth). Living in Michigan or an increased ranking o f the overall quality o f brochure will result in an increase in logit(visit). The overall probability o f correct predictions is 68.32%, which is 18.32 percentage points better than the 50% odds o f correct predictions using a pure random process such as tossing a coin. This improved probability to predict the propensity to visit Frankenmuth derives from knowledge o f the above five mentioned independent variables. The results indicate that inquirers who are Michigan residents, live near Frankenmuth, who have less educational background, who must wait only a short period o f time for information from Ill the FCCVB and who rank the information as being o f high quality are more likely to visit Frankenmuth The stepwise logistic regression model can be used for future predictions o f who will visit and not visit Frankenmuth. The marketer can predict the probability that an individual inquirer will take a trip to Frankenmuth by using the following equation: P(Y =1/X h, X2l, X , , X , , X 5i) = I -v 11+h 2X 2 i+h y v 3, +fc4 •' 4 1 + ftVv s, W here P(Y=1): the probability o f taking a trip to Frankenmuth e: the base o f natural logarithms, a: the intercept b, x (i: the distance coefficient b, multiplied by the distance o f individual /' s residence from Frankenmuth. b,x,,: the residence status coefficient b., multiplied by the individual i's resident status value(l for Michigan residents, 0 for non-Michigan residents) b, x,,: the quality o f brochure coefficient b 3 multiplied by the individual i's ranking value o f quality o f brochure (1 representing terrible, 7 representing excellent) bjX^: the elapsed time coefficient b4 multiplied by the value o f days between inquiry and receipt o f information by individual i. bsx5l: the education coefficient b 5 multiplied by the years o f school attended by individual i. 1 12 Predictions for individual cases may be obtained by replacing the variables in the equation with their values for specific cases. For example, for a non-Michigan resident with a bachelors degree (16 years o f school) living 200 miles from Frankenmuth, who waited 10 days for requested information, and who gave a ranking o f 6 for the overall quality o f the brochure, the probability o f a visit is calculated as follows: Logitfll.si!) -.0413 -.0017*(200) '■ .6581 *(0) -i .2885*(6) -.0577*(10) -.0829*. (16) -.5537. This corresponds to a probability o f a visit o f e 5537 (1 t e"s5i ) .365. Thus, this individual has a less than 50-50 probability o f making a visit to Frankenmuth. This information can be useful in understanding and interpreting differences in the Frankenmuth travel market and can help the FCCVB better predict the probability o f a visit to Frankenmuth by individual inquirers. But, how could FCCVB actually use this knowledge (Equation 1) to predict the probability o f visits to Frankenmuth?; Should it send out different information to different inquirers?; and/or Should it ignore certain inquiries where probability o f visit is low? Those questions are very important issues to FCC VB's promotion strategies and budget allocation. The Equation 1 demonstrated above could be utilized to guide what information to send to inquirers only if the FCCVB could design a system to obtain information from each inquirer for the five variables included in Model 1. Screening inquiries in this fashion would be costly and probably not acceptable to many inquirers. Thus, marketers probably can not use Model 1 to send different materials to different segments o f inquirers. However, the model might be exercised to identify market segments with a high propensity to visit Frankenmuth. Existing secondary data could be analyzed to identify concentrations o f best prospects, and the FCCVB could then seek to focus prom otion on these concentrations or the promotion itself could be designed to appeal to target segments. Thus, Equation 1 allows the FCCVB to zero in on those particular variables with the greatest prediction power. Furthermore, the model might be simplified to include only independent variables which are simple to access for the FCCVB such as: distance and residence. For example, the FCCVB already must ask inquirers for their address, therefore residence status (RESI) is ready available. Therefore, residence (RESI) is an available independent variable that can be used to divide inquiries into "Michigan" and "Non-Michigan" segments. The probability o f visit to Frankenmuth using RESI as an independent variable is .57 for Michigan residents and .33 for Non-M ichigan residents respectively. The overall probability o f correct predictions using RESI is 62.56% which is 12.56 percentage points better than the 50% odds o f correct predictions using a pure random process. In addition, the marketer can estimate the probability o f a visit to Frankenmuth using distance as an indicator. For example, the probability o f a visit to Frankenmuth is more than .50 if distance to Frankenmuth is under 152 miles. The overall probability o f correct predictions is 62.54% which is 12.54 percentage points better than 50% odds o f correct predictions using a pure random process. The probability o f correct predictions can be enhanced by using residence and distance in combination in an abbreviated version o f Model 1. Again, since inquirers' addresses are obtained in order to mail them the information they requested, the variable RESI is available and the variable DIST can be simply calculated. For example, the probability o f a visit to Frankenmuth is 114 more than .50 if distance to Frankenmuth is under 252 miles and the inquirer is a Michigan resident. On the other hand, if the inquirer is not a Michigan resident then no matter how close the inquirer lives to Frankenmuth, the probability o f a visit to Frankenmuth is still below .50. However, combining the RESI and DIST independent variables as an abbreviated version o f Model 1, the overall probability o f correct predictions is 62 .89%. This is 12.89 percentage points better than 50% odds o f correct predictions using a pure random process In addition to using it to target advertising and, in an abbreviated form, to custom ize information to inquirers. There is a third potential application o f Model 1. Even in cases where one does not have a prior information about elapsed time between inquiry and receipt o f information, its impact can be explored using Model 1. In essence, Model 1 can be used as a simulation model to explore how increasing or decreasing elapsed time will influence the probability that inquirers will visit Frankenmuth. Simulation results then can be used to evaluate the rate o f return from proposed alternative inquiry response strategies with varying impacts on length o f elapsed time between inquiry and receipt o f promotional materials. For example, results indicate that inquirers would more likely visit Frankenmuth if they received the requested information in 6 days (the probability is .5004). The overall probability o f correct predictions using ELAPRI as a predictor is 61.97% , which is 11.97 percentage points better than the 50% odds o f correct prediction using a pure random process. By examining values o f ELAPRI which are less than 6 , marketers can derive estimated o f cost-effectives o f enhanced inquiry response progams. 115 Model 2: Prediction o f a Repeat Visit to Frankenmuth A critical first step in the process o f building Model 2 was to develop an appropriate definition o f the dependent variable. Repeat visitors to Frankenmuth, the dependent variable, could not be directly identified since at the time respondents were surveyed a repeat visit was a future event. However, the questionnaire did include a question concerning intention to visit again. The question was a scaled question with "certain to visit again" and "certain not to visit again" on the opposite poles o f the scale. The midpoint on the scale was presented as a neutral point, and the opposite poles on the scale were "likely" and "unlikely to visit again " Since the repeat visit variable could not be measured directly, intention to visit again was selected as the best substitute. The next question to resolve was whether or not to aggregate "certain to visit again" and "likely to visit again" as the measure o f the dependent variable. Aggregating the two may have been appropriate if one was interested in using the model to predict total repeat visitors from among this study population. However, one would expect the "likely to visit again" group o f respondents would contain a considerably higher percentage o f non-repeat visitors than the "certain to visit again" group which would create a considerable error o f estimate (i.e., inflated estimates o f who is likely to be a repeat visitor). Using only the "certain to visit again" group as a measure o f the dependent variable would, on the other hand, yield a conservative estimate o f who is likely to be a repeat visitor. Considering how the results o f Model 2 would most likely to be applied, it was concluded that a conservative approach to defining the dependent variable was preferable. 1 16 The model's practical marketing application would most likely involve identifying prospective repeat visitors and then targeting them with special promotion efforts These will be costly, m ore costly than a general promotion program, hence it is desirable to direct repeat visitor promotion only at individuals who are the most likely to actually be repeat visitors This will insure maximum return on investment from a repeat visitor program designed around results o f applying Model 2. Thus, in developing Model 2, only respondents who indicated that they were "certain to visit again" were considered to be repeat visitors. All other respondents, including those who indicated that they were "likely to visit again" w ere considered to be non-repeat visitors. It was necessary to develop an alternative strategy for building Model 2 than was used for building Model 1 because the list o f possible independent variables was larger and the number o f observations was smaller in the case o f Model 2 vs. Model 1. As can be seen in Table 13, univariate hypotheses tests for examination yielded 15 prospective independent variables for examination in the multivariate logistic model 2 which exceeded the degrees o f freedom constraint imposed by the available sample size In accordance with suggestions found in Hosm er and Lemeshow (1989), logistic regression analyses were performed on the following four subgroups o f variables to reduce the number o f candidate variables for full model: ( ’ ) sociodemographic variables, ( 2 ) brochure evaluation variables, (3) satisfaction with the trip, and (4) travel characteristic variables. All variables which entered the four submodels were accepted as the set o f candidate variables for evaluation for inclusion in the full Model 2 117 Sub-model 2.1. Prediction o f Re-visit to Frankenmuth Using Sociodemographic Variables As discussed above, the dependent variable was derived from responses to Question 14 "How likely are you to visit Frankenmuth again?". Respondents who answered "certain to visit again" were assumed to be repeat visitors and were assigned the number 1 as part o f the data coding process. Zero was assigned to the other four possible choices that could have been selected on the five point scale. Thus, the dependent variable was "repeat visit" (yes = I, no = 0 ) to Frankenmuth and the independent variables were seven sociodemographic variables. The relationship can be expressed as: , F P r o b a b ility o f R e p e a l v is it ~] , t • \ lnL .-Pnb.biii.yofRcpc.vfa.iJ = rtSociodem ograpli.cs) In order to interpret independent variables accurately and to build a useful model with significant variables, a stepwise procedure was applied The empirical model which was examined using stepwise procedures is presented below: Where. P, : the probability that current visitors will visit Frankenmuth in the future x h . mileage between Frankenmuth and respondent's residence x,, gender x,,.- employment status; 1 = employed, 0 = unemployed x4l: years attended school x v age x6]: household income; more than or equal to $50,000 = 1, less than $50,000 = 0 x7l: residence status; Michigan = 1, non-Michigan = 0 When the stepwise procedure was applied, tw o variables were selected and placed into the model, The resulting equation is: Logit (Repeat Visit) 2.9367 - .0Q5*(DIST) - .H67*(EDUC) This function indicates that increasing trip mileage and years o f attending school causes a decrease in logit(repeat visitors). The overall probability o f a correct prediction is 61.04%. Thus, applying the goodness-of-fit logistic regression equation increases the probability o f correctly classifying respondents significantly above the 50% expected w ere a purely random assignment procedure to be used. The results indicate that current visitors who live near Frankenmuth with less educational background are more likely to visit Frankenmuth again. The results o f fitting the logistic regression model to these data are given in Table 17. This best stepwise logistic regression model for predicting the probability that an individual visitor will visit Frankenmuth again using sociodemographic variables then is: P(Y=1/X.IP X,,-1 )' = ■1 + t’ " ! • ' li ’ ^2^ W here P(Y=1): the probability o f taking a trip to Frankenmuth again e: the base o f natural logarithms, a: the intercept b, x n: the distance coefficient b, multiplied by the distance o f individual /'s residence from Frankenmuth. b, x,,: the education coefficient b, multiplied by the years o f school attended by individual /. 1 19 Table 17. Logistic Regression Procedure with Dependent Variable "Repeat Visit or not" for Sociodemographic Variables (Goodness-of-fit Reduced M odel) Variable DIST (X„) EDUC (X„) Constant B -.0050 -.1467 2.9367 S.E. .0 012 .0576 .8498 Wald df Sig 18.2728 6.4904 11.9411 1 1 1 .0 0 0 0 .0108 .0005 R -.2172 -.1141 Exp(B) .9950 .8636 E stim ated Coefficients for a M ultiple Logistic Regression Model including X„ (D istance between residence an d Frankcnm utli) and X 4| (actual years attending school) * Log Likelihood = -155.55 Suppose that we wanted to estimate the probability o f a current visitor returning to Frankenmuth if s/he has attended school for 15 years and lives 300 miles from Frankenmuth. By inserting this individual's distance from Frankenmuth and years o f education into the above model, her/his probability o f a repeat visit would be found to equal 0.3 1. Sub-model 2.2: Prediction o f a Re-visit to Frankenmuth Using Brochure Evaluation Variables In this model, the independent variables are brochure evaluation variables based on Question 20 through Question 27 in the questionnaire (See Appendix A). Although the measurements for these questions were taken from 7-point Likert-scales, most respondents rarely selected a ranking o f less than point 5 Thus, data were aggregated into tw o groups based on the median point, responses o f 1 to 5 were aggregated into 0 in the analysis, and number 6 , 7 were transformed into 1. The relationship between the dependent variable and independent variables thus can be expressed as: 120 P ro b ab ility o f R epeat Ti?it 1—P ro b a b ility o f R epeat visit J = /(Brochure evaluation variables) The empirical model including the eight predictors used in the stepwise regression procedure can be expressed as follows: Where P, : the probability that current visitors will visit Frankenmuth in the future x ,,: usefulness o f brochure; 1 = yes, 0 = no (Question 20) x;i interesting to read; 1 = yes, 0 = no (Question 21) x3l: attractiveness o f brochure; 1 = yes, 0 = no (Question 21) x ^ : interest in visiting Frankenmuth; 1 = yes, 0 = no (Question 23) xSl: influenced decision to visit Frankenmuth; 1 = yes, 0 = no (question 24) x6l: more expenditure stimulated by brochure; 1 = yes, 0 = no (Question 25) x7l: accuracy o f brochure; 1 = yes, 0 = no (Question 26) xSl: overall quality o f brochure; 1 = yes, 0 = no (Question 27) After the stepwise procedure was applied, only tw o variables (x4i\ increased interest in visiting Frankenmuth and xSl: overall quality o f the brochure) remained in the model. The equation follows: l.ogit (Repeat Visit) -.9905 i .64H3*(x4t) t .7620*(x-l ) This function shows that increasing "increased interest in visiting Frankenmuth by the brochure" and "overall quality o f the brochure" result in an increase in logit(repeat 121 visitors) The overall probability o f correct prediction is 61.40%. The results indicate that the current visitors who answered yes to "increase interest in visiting Frankenmuth by brochure" and "overall quality o f brochure" are more likely to visit Frankenmuth again. The result o f fitting the logistic regression model to these data are given in Table 18. This model for predicting the probability that a current individual visitor will take a trip to Frankenmuth again using brochure evaluation variables as predictors can be expressed as: o* /»1 A' 11♦ P(Y=1/X, X,, ) - -e— .. th v 11 1 -K ? I 1* 2 2f Where P(Y =1): the probability o f taking a trip to Frankenmuth again e: the base o f natural logarithms, a: the intercept b, x h: the interest in the brochure coefficient b, multiplied by the value o f "increased interest in visiting Frankenmuth created by the brochure" ranked by individual /. (1 for yes, 0 for no). b ,x ,r the quality coefficient b, multiplied by the value o f "overall quality o f brochure" ranked by individual i. For example, if an individual visitor answered "yes" to both "increase interest in visiting Frankenmuth by the brochure" and "overall quality o f the brochure", her/his probability o f taking a trip to Frankenmuth is 0.60. 122 Table 18. Logistic Regression Procedure with Dependent Variable "Repeat Visit or not" for Brochure Evaluation Variables (Goodness-of-fit Reduced M odel) Variable Interest (X4|) Quality (X J Constant B SE Wald .6481 .7620 -.9905 .2936 .3445 .2966 4.8772 4.8932 11.1504 df Sig R Exp(B) 1 1 1 .0272 .0270 .0008 .0954 .0957 1.9123 2.1425 E stim ated Coefficients for a M ultiple Logistic Regression Model including X 4| (Interest in visiting F rankenm uth by brochure) and XS) (Overall quality o f brochure) * Log Likelihood = -1 5 0 .1 3 Sub-model 2.3: Prediction o f a Re-visit to Frankenmuth Using Satisfaction Variables For this model, the independent variables are satisfaction variables derived from Question 13, Question 15 and Question 16 (See Appendix A). The relationship can be expressed as: . r iY ubabilitv o l'R c p c a t v isit "I |_ i —P r o b a b i l i t y o i ' K c p o a t v i s i t J c ■ t t \ =./(Satisfaction variables) The empirical model used in the stepwise procedure can be expressed as: Where P, : the probability that current visitors will visit Frankenmuth in the future x h : travel experience ( from a 5-point Likert scale with 1 representing "much worse than expected" and 5 representing "much more than expected" (Question 13)) x,, recommendation; 1 = yes, 0 = no (Question 15) 123 x ,,: destination ranking from a 7-point Likcrt scale with 1 representing "terrible" and 7 representing "excellent" (Question 16)). After the stepwise procedure was applied in formulating the prediction equation, only tw o variables remained in the equation: travel experience and destination ranking. The equation follows: Logit (Repeat Visit) - 7.8458 1 ,494*(xh) * 1.001 2 *(x 3i) This function indicates that increasing "travel experience" and "destination ranking" will result in an increase in logit(Repeat visit). The overall probability o f correct prediction is 72.29%. The results indicate that visitors whose expectations are exceeded and who rank Frankenmuth highly as a travel destination are more likely to visit Frankenmuth again. The results o f fitting the logistic regression model are given in Table 19. The equation for predicting repeat visitors from trip satisfaction variables is. P(Y=1/X ll, Xa ) Where P(Y=1): the probability o f taking a trip to Frankenmuth again e. the base o f natural logarithms, a: the intercept b, x h: the experience coefficient b, multiplied by the value o f "travel experience" reported by individual i. b,x,,: the destination coefficient b, multiplied by the value o f "destination ranking" reported by the individual / 124 Table 19 Logistic Regression Procedure with Dependent Variable "Repeat Visit or not" for Satisfaction Variables (Goodness-of-fit Reduced Model) V a ria b le B E x p e rie n c e ( X ,,) .4 9 4 0 R e c o m m e n d (X ,,) 1.0012 C o n s ta n t - 7 .8 4 5 8 S E W ald .1 8 5 0 .1885 1.1298 7.1271 2 8 .2 1 6 7 4 8 .2 2 5 3 df Sig 1 I 1 .0076 .0 0 0 0 .0 0 0 0 R E x p (B ) .1 2 1 9 .2 7 5 6 1.6338 2 .7 2 1 7 E s tim a te d C o e ffic ie n ts fo r a M u ltip le L o g istic R e g re s s io n M o d e l in c lu d in g X ,, (tra v e l e x p e r ie n c e ) a n d X ,, ( r e c o m m e n d a tio n ra n k in g ) * L o g L ik e lih o o d = - 1 3 8 .4 8 For example, if we wanted to estimate the probability o f a current visitor re-visiting Frankenmuth answered 5 to the "travel experience" question and 7 to the "destination ranking" question, applying the above model would yield a probability o f repeat visit estimate o f 0.84. Sub-model 2.4. Prediction o f a Re-visit to Frankenmuth Using Travel Behavior Variables In this model, the independent variables are travel behavior variables derived from Q uestion 2, Question 8 , Question 9, Question 10 and Question 1 lb. The relationship can be expressed as: , r lY ohabilitv o f R ep eat visit “I L T -^ o h ahniiy o n u ’peaTTisit J r* ^ , , • • x ^ T r a v e l charactent.cs) The empirical model used in stepwise regression analysis included five predictors and can be expressed as follows. ln[T^]=f V ,. 125 W here P, : the probability that current visitors will visit Frankenmuth in the future x ,,: previous experience; 1 = yes, 0 = no (Question 2) x2l primary destination; 1 = yes, 0 = no (Question 8 ) x ,,: primary purpose o f trip; 1 = recreation/leisure, 0 = other (Question 9) Xj,: day tripper; 1 = yes, 0 = no (Question 10) x5l: travel party size; actual travel party size After the stepwise procedure was applied, only one variable (xh: previous experience) remained in the model. The equation follows: Logit (Repeat Visit) -.5978 ■ I.()-i8*(xh ) This function indicates that a previous visit can result in an increase in logit(repeat visit). The overall probability o f correct prediction is 62.50%. The results indicate that current visitors who had been to Frankenmuth earlier are more likely to revisit Frankenmuth later The results o f fitting the logistic regression model are given in Table 20 The prediction model derived from travel behavior variables using the stepwise procedure can be expressed as: a . A , A ’ ,, P(Y=1/X„) v I,/ - 6 - W here P(Y=1): the probability o f taking a trip to Frankenmuth again e: the base o f natural logarithms, a: the intercept 126 Table 20. Logistic Regression Procedure with Dependent Variable "Repeat Visit or not" Using Travel Behavior Variables (Goodness-of-fit Reduced M odel) V a ria b le B p re v io u s v is it (X ,,) C o n s ta n t 1 .0 4 8 0 - .5 9 7 8 SE W ald df Sig R E x p (B ) .2 8 1 2 .2 1 6 7 13.8888 7 .6 0 8 6 1 1 .0 0 0 2 .0 0 5 8 .1 9 5 7 2 .8 5 1 9 * Log L ikelihood = -155.257 b, x h: the previous visit coefficient b, multiplied by the observed previous visit behavior by individual /. Suppose that we wanted to estimate the probability o f an individual visitor revisiting Frankenmuth. If this individual answered "yes" to the "previous visit" question, s/he would have a 0.61 probability o f making a future trip to Frankenmuth. Model 2: Full Model Derived from Sub-model 2.1 - 2.4 Using Sociodemouraphic. Brochure Evaluation. Trip Satisfaction and Travel Behavior Variables Having screened out the weakest independent variables via creating four Sub-models, the variables remaining were combined and analyzed using similar procedures to those employed in building the submodels. The significant predictors that remained in earlier mentioned Sub-models included: "distance away from Frankenmuth", "years attended school", "increased interest in visiting Frankenmuth stimulated by brochure", "overall quality o f the brochure", "satisfaction with travel experience", "destination ranking", and "previous visit". Before application o f the stepwise procedure, Model 2 contained seven predictors, and it can be expressed as follows: 127 W here P, : the probability that current visitors will visit Frankenmuth in the future x h : distance from Frankenmuth. x,, education. x ,,: interest in visiting Frankenmuth; 1 = yes, 0 = no x4l: overall quality o f brochure; 1 = yes, 0 = no x5l: satisfaction with travel experience; 5-point Likert scale with 1 representing "much worse than expected" and 5 representing "much better than expected". x6l: destination ranking; 7-point Likert scale with 1 representing "terrible" and 7 representing "excellent". x7l: previous visit; l=yes, 0 = no. When the stepwise procedure was applied, three variables (xh: distance to Frankenmuth, xS|: satisfaction with travel experience, and x(M: destination ranking) remained in the model. The resulting equation follows: Logit (Repeat Visit) -8.8774 - ,0094*(xh ) ■ .8 7 4 1 *(x 5 i ) » 1.2170* (x6) —Equation 2 This function indicates that an increase in "distance away from Frankenmuth" will result in a decrease in logitfrepeat visit). Increase in "travel experience" and "destination ranking" result in an increase in logit(repeat visitors). The overall probability o f a correct prediction was found to be 78.13%. The results indicate that current visitors who live near 128 Frankenmuth and register high scores for "travel experience" and "destination ranking" are most likely to revisit Frankenmuth. The results o f fitting the logistic regression model are given in Table 21. After variable screening via first building submodels and applying stepwise procedures to the remaining combined variables, the following best model for predicting repeat visitors resulted: P(Y=1/X. X, X J = L’ ... . .. „ .. v IP 21. .1 1/ W here P(Y =1): the probability o f taking a trip to Frankenmuth again e: the base o f natural logarithms, a: the intercept b, x h: the distance coefficient b, multiplied by distance from the individual /'s residence to Frankenmuth. b ,x ,r the satisfaction with travel experience coefficient b; multiplied by the ranking value o f "satisfaction with travel experience" by individual /. b,x,,: the destination ranking coefficient b, multiplied by the ranking value o f "destination ranking" by individual /. To illustrate application o f the Model 2, suppose that we wanted to estimate the probability o f a current visitor's desire to revisit Frankenmuth. If the individual lives 200 miles from Frankenmuth, scored "satisfaction with travel experience" as 3 and "destination ranking" as 7, the probability o f revisiting Frankenmuth for this individual is 0.79 129 Table 21. Logistic Regression Procedure with Dependent Variable "Repeat Visit or not" -- Full M odel (Goodness-of-fit Reduced Model) V a ria b le B distance Sig E x p (B ) W ald -.0094 .0017 29.3143 .0000 -.2967 .9906 .8741 .2267 14.8649 .0001 .2036 2.3967 ranking 1.217 .2336 27.1472 I Constant -8.8774 1.4106 39.6038 1 travel experience df R S .E . d e stin a tio n .0000 .2847 3.3770 .0000 * Log Likelihood = - 1 0 1 . 0 7 Model 2 results appear to have less potential for direct application by Frankenmuth m arketers than was the case for Model 1. A priori data for individual visitors that would be needed to exploit the demonstrated predictive power o f Model 2 are not available nor could they be readily obtained from visitors. The exception to this is the distance variable which would appear to offer a moderate basis for singling out repeat visitor prospects. Still, Model 2 does provide some useful general marketing information. Results confirm the importance o f insuring that visitors to Frankenmuth receive the best possible service to insure that they leave satisfied. Second, the significance o f a high ranking o f Frankenmuth as a travel destination points to the need for future research directed at uncovering what factors underly how visitors rank destinations. An improved understanding o f these factors would be useful in developing product development strategies with the goal o f enhancing 130 how Frankenmuth is rated as a travel destination. Implementing these strategies would, as the results suggest, stimulate repeat visits by current visitors. CHAPTER V SUM M ARY, LIM ITATIONS AND IMPLICATIONS The objectives o f this study were to (1) identify factors that influence different tourist market segment's (e.g., first-time and repeat visitors) decision to make trips to Frankenmuth after requesting information from the FCCVB and (2) develop models that estimate the probability that different market segments will visit Frankenmuth. Information useful for developing marketing and/or advertising strategies was obtained through: ( 1 ) descriptive analyses that produced profiles o f different market segments (visitors, nonvisitors, first-time visitors, and repeat visitors), ( 2 ) tests for significant differences between segments, and (3) models that predict the propensity o f different segments to visit Frankenmuth. A cross-sectional survey method was employed to collect data used in the study Random systematic sampling was applied to select the sample o f 1263 from a sampling frame consisting o f 5,967 Americans and Canadians who requested information via a toll free call to the Frankenmuth Chamber o f Commerce and Visitors Bureau (FCCVB) between September 1, 1993 and March 15, 1994. Questionnaires were sent certified mail to these 1263 individuals. This produced 595 usable questionnaires for data analysis Two hypotheses were established involving sets o f variables which were thought to influence travel decisions to Frankenmuth. Hypothesis 1 tested for differences between 132 inquirers who did and did not make visits to Frankenmuth after making an inquiry to the FCCVB Statistically significant differences were found between these two segments for the following variables: familiarity with Frankenmuth, state o f residency, distance from inquirers' residence to Frankenmuth, whether inquirers actually read the promotional materials (brochure) they were sent, elapsed time between inquiry and receipt o f requested information, interest in the brochure, the perceived quality o f the brochure, and the age o f those who did and did not visit after an inquiry Respondents were more likely to visit Frankenmuth if they had visited before, reside in Michigan, live a shorter distance away, received promotional materials sooner after requesting them, read the materials received, found the materials interesting, rate the materials high, and are relatively younger. Hypothesis 2 tested for differences between visitors who had and had not visited Frankenmuth previous by. Results indicated differences exist with respect to the following variables: state o f residence, distance from their residence to Frankenmuth, type o f media sources used for information about Frankenmuth, size o f travel party, day vs. overnight trips, satisfaction with their Frankenmuth experience, intention to make a repeat visit, intended to recommend Frankenmuth as a travel destination to others, ranking o f Frankenmuth as a travel destination, whether visitors utilized the brochure on site, whether or not the brochure influenced current visitors to visit Frankenmuth again, and age o f current visitors. Respondents were more likely to be first-time visitors if they were not residents o f Michigan, live longer distance away, relied more on word-of-m outh information, had a small travel party size, utilized accommodations in Frankenmuth, derived more satisfaction with their Frankenmuth experience, more likely to recommend 133 Frankenmuth as a travel destination to others, utilized the brochure on site, were more influenced by the brochure to make visits to Frankenmuth, and had more years o f schooling. A three step process w as employed to build the models (1 and 2) that predict the propensity to visit Frankenmuth. First, tests o f hypotheses were performed in order to identify (independent) variables that are significantly related to the propensity to visit Frankenmuth. Stepwise logistic regression analyses were then performed to delete variables that did not significantly affect the prediction o f propensity o f inquirers to visit Frankenmuth. Results from Model 1 indicate that travel distance to Frankenmuth, in-state vs. out-of-state, perceptions o f the quality o f the brochure received in response to inquiries, elapsed time between the inquiry and receipt o f the information, and level o f education most influenced whether an inquiry was followed by a visit. The predictive ability o f the Model 1 was evaluated using a random selection o f the 595 cases. The overall correct prediction o f the model is 68.32%. This is 18.32 percentage points better than the expected 50% from a pure random process. Model 2 was developed to predict the probability o f repeat visits to Frankenmuth. The results indicate that the likelihood o f a repeat visit is influenced by travel distance to Frankenmuth, satisfaction with previous Frankenmuth (travel) experiences, and ranking o f Frankenmuth as a travel destination. The overall correct prediction for this model is 78.13%, which is 28.13 greater than the 50% expected from a pure random process. 134 Study Limitations There a number o f sampling related limitations associated with this study First, the sample only consisted o f persons who requested information from the FCCVB. Persons who visited Frankenmuth but did not request information were not included in the sample. In addition, the sample frame was limited to persons that made inquiries during a six month period from September 1, 1993 to March 15, 1994. This raises questions regarding the extent to which these findings can be generalized to all visitors to Frankenmuth or even to all those who make inquiries to the FCCVB. This does not mean that the basic structures o f the prediction models are not valid. Some variation in coefficients for the variables included in the model and possibly some variation in the sets o f independent variables included could result if an expanded sample frame had been used to develop the models. However, there are no obvious intuitive or empirical reasons to suggest that the travel behavior or decision criteria o f the sample is significantly different from that o f all inquirers or visitors to Frankenmuth. Another limitation o f this study is the potential influence o f non-response bias. Non-response bias has been identified as a major limitation o f many previous inquiry conversion studies. This is also the single biggest impediment to any survey research based project. Questionnaires were sent to 1263 persons but only 595 usable returns were produced. In other words, slightly less than 50% o f the original sample was represented in the analyses involved in this study While the study design incorporated strategies to increase the response rate (e.g., certified mail), no effort was made (e.g., follow up questionnaires) to determine the extent or direction o f non-response bias. As a result, it is 135 impossible to determine whether the 595 questionnaires are representative o f persons that made inquiries. There are also certain limitations to using secondary data to estimate the logistic models. This original study was not designed to produce data to estimate or evaluate logistic models. Estimation o f the models was limited to variables that were included on the original questionnaire. In addition, many o f the response categories in the questionnaire were ordinal rather than precise interval scales Variable transformation represents another possible limitation in this study. Transform ations used in this study involved collapsing categories o f nominal or ordinal data to obtain a smaller, more usable number o f categories. There is no standard approach or rules to guide the transformation decisions that were necessary in this study. For nominal data, any categories could conceivably be combined, but for ordinal data only adjacent categories should be combined. The decision to collapse (combine) categories always involves a compromise between producing too many categories with some based upon very few respondents (e .g ., only three o f the 295 respondents were temporarily unemployed and five were students — Question 30) and collapsing so much that information is lost. For example, the categories relating to "brochure evaluation" were collapsed from the seven possible categories into tw o categories (yes and no). Although this did not reduce the predictive ability o f Model 2, information was lost. However, while all data transformations employed in this study were deemed desirable based upon sample size and other considerations, it is not possible to fully assess the degree to which they may have collectively impacted on overall results 136 Implications This study made progress toward developing a plausible model for examining the factors which influence destination decision-making. The significant positive relationships found between satisfaction and repeat visits to the same destination supports the findings o f other studies by Chon (1992, 1990, and 1989). It also verifies the importance o f continuous investment in service quality improvement. The significant negative relationship between distance and the decision to visit or not visit supports the findings o f a study by Smith (1983). The findings also suggest that first-time visitors have a greater propensity to use w ord-of-m outh advertising. This may be in part be attributable to the novelty o f the visit, and the resulting urge to "pass on" information to family and friends. This could provide a basis for a marketing campaign to provoke w ord-of-m outh by enlisting first-time visitors as marketing representatives. The results indicate that customer (not seller) perception o f the quality o f a promotional piece, in this case a brochure, is an important factor in determining whether potential first time visitors make a trip to Frankenmuth. Information on "how" potential visitors evaluate the quality o f promotions should be incorporated into promotional platforms, layout, graphics and copywriting. The relationship between elapsed time — between the time o f inquiry and receipt o f information — and decisions to visit confirms the importance o f timely follow-up to inquiries. The findings indicate that to maximize effectiveness, follow-up information should be sent within three days o f receipt o f the inquiry. The FCCVB could also enhance 137 the effectiveness o f their inquiry response program by focusing on factors/variables that comprise Model 1 Model 2 provides a more in-depth understanding o f the factors that influence repeat visits to Frankenmuth. Travel distance and satisfaction with their Frankenmuth experiences significantly influence likelihood o f repeat visit. These findings can be used to more precisely target marketing aimed at stimulating repeat visits. The results also indicate that the FVCCB should consider tailoring inquiry response packages for different segments o f inquirers. For example, packages sent to potential first-time visitors should include relatively more detailed and comprehensive information such as special events (i.e., winter art, snow and ice sculpting, summer music festival), facilities, and services available in Frankenmuth (e.g., accommodations, restaurant, touring). Packages designed for persons that have previously experienced Frankenmuth should focus on new features, programs and amenities. Model 2 provides a relatively efficient way to predict the probability that person who make inquiries and then visits Frankenmuth will make a repeat visit. It identifies factors that have the greatest influence on repeat visits. This information can also help the FCCVB decide on what types o f information to acquire about persons that make inquiries (e.g., w hether they have made a previous visit). Although application o f the model requires expertise in statistics and computers, this could be obtained from academic institutes or private research firms. 138 Future Research While there appears to be no obvious reasons to question the generalizability o f the results obtained for Model 1 and Model 2, there is no question that the sample used to estimate the models was limited and may be subject to non-response bias. It would be useful to verify results obtained using data collected from a more representative sample o f inquirers and also from visitors that make trips but not pre-trip inquiries. The focus would be on whether the independent variables and/or coefficients change significantly when the models are estimated with data collected from a different (expanded) sample. For example, the sample frame could be clustered into different geographic segments. Based on those geographic segments, a random household survey could be applied to collect similar information to that used in this study. While such a survey would be far more costly that was that used in this study, the costs could be spread by designing the survey to obtain information beyond that required for model verification purposes. Future research could also be directed to developing and testing m ethods that generate a representative sample o f persons that make inquiries. This could include different sampling schemes, alternative methods for increasing response rate, and methods to assess and correct for non-response bias. Future research should also be directed toward improvements in questionnaires to generate the data used in this study. For example, it would be useful to use more consistent measurement scales across questions. M ore consistent Likert scales (e.g., 5-point or 7-point) would simplify or eliminate the need for data transformations. Also, it 139 would be useful to experiment with ways to generate more interval, rather than ordinal data scales. The predictive ability o f models could also be evaluated by collecting similar information from a sample o f those who make inquiries and a separate sample o f visitors to Frankenmuth The visitors would also be asked whether, or not, they made a pre-trip inquiry. This information could be used to assess how well the models are able to predict percent o f visits and characteristics o f those who visit. If collected on a regular basis, these data could be used to adjust the models over time. 140 L IS T O F R E F E R E N C E S Aaker, D. A. & Carman, J. M. (1982). Are you over-advertising? Journal o f Advertising Research. 22(Aug/Sept). 57-70. Ahmed, Z. U. & Krohn, F. B. (1990). Reversing the United States' declining competitiveness in the marketing o f international tourism: a perspective on future policy. Journal o f Travel Research. 29(2), 23-29. Anderson, J. A. (1980). Logistic discrimination, In Krishnaiah and Kanal (Eds ). Handbook o f Statistics. Vol. 2: Classification Pattern Recognition and Reduction o f Dimension. New York, NY: North-Holland. Assael, H. & Keon, J (1982). Nonsampling vs sampling errors in survey research. Journal o f M arketing. 46(Spring). 114-123. Ballman, G., Burke, J., Blank, Ule, & Korte, D. (1984). Toward higher quality conversion studies: Refining the numbers game. Journal o f Travel Research. 22(4), 27-33. Barnett, N. L. (3 969). Beyond market segmentation . Harvard Business Review. 47(1), 152-166. Bull, S. B. and Donner, A. (1987). The efficiency o f multinomial logistic regression compared with multiple group discriminant analysis. Journal o f the American Statistical Association. 82, 1118 - 1122. Burke, J. F. (1989). Evaluation research: "Tips for conducting conversion studies". TTRA International Newsletter. 10(1), 3-4. Burke, J F. & Gitelson, R. (1990) Conversion studies: Assumption, application, accuracy, and abuse. Journal o f Travel Research. 28(3), 46-50. Burke, J. V. & Lindblom, L. (1989). Strategies for evaluating direct response tourism marketing. Journal o f Travel Research. 27(2), 33-37. Calantone, R. J. & Johar, J. (1984). Seasonal segmentation o f the tourism market using a benefit segmentation approach. Journal o f Travel Research. 23(2), 14-14. 141 Campbell, D. T., and Stanley, J. C. (1966). Experimental and Quasi-Experimental design for research. Chicago, IL: Rand McNally. Churchill, G A (1986). Marketing research Chicago, IL: The Dryden Press. Cleary, P. & Angel, R. (1984). The analysis o f relationships involving dichotom ous dependent variables. Journal o f Health & Social Behavior. 25(3), 334-348. Chon, Kye-Sung. (1992). The role o f destination image in tourism: An extension. The Tourist Review. 47(1), 2-8. Chon, Kye-Sung (1991). Tourist information-seeking behavior and its marketing implication. Hospitality Research Journal. 14(2), 485-490. Chon, Kye-Sung.(1990). The role o f destination image in tourism: a review and discussion. The Tourist Review. 45(2), 2-9. Chon, Kye-Sung. (1989). Understanding recreational traveler's motivation, attitude, and satisfaction. The Tourist Review. 44(1), 3-7 Cadez, G. & Hunt, J. D. (1987). A comparison between Port-of-Entry visitor center users and non-users. Logan, U T : Institute for the study o f outdoor recreation and tourism, Utah State University. Cox, D. R. (1970). The analysis o f binary data. London: Chapman and Hall. Coltman, M. (1989). Tourism marketing. New York, NY: Van Nostrand Reinhold Crompton, J L. (1979). Motivations for pleasure vacation. Annals o f Tourism Research. 6(4), 408-424. Crompton, J. L., Fakeye, P. C., & Lue, C. C. (1992). Positioning: The example o f the lower Rio Grande Valley in the winter long stay destination market. Journal o f Travel Research. 31(2), 20-26. Dann, G. (1976). The holiday was simply fantastic. The Tourist Review. 31(3), 19-23. Dann, G. (1977). Anomie, ego-enhancement and tourism. Annals o f Tourism Research. 4(4), 184-194. Dann, G. (1981). Tourism motivation: An appraisal. Annals o f Tourism R esearch. 8(2), 187-219. 142 Davis, D. & Sternquist, B. (1987). Appealing to the elusive tourist: An attribute cluster strategy. Journal o f Travel Research. 25(4), 25-31. Davidson, T. L. & W iethaupt, W. (1989). Accountability marketing research: An increasing vital tool for travel marketers Journal o f Travel Research. 27(4), 42-45 Dholakia, R.R., Johnson, J.L., Della Bitta, A.J., and Dholakia, N. (1993). Decision-making time in organizational buying behavior: an investigation o f its antecedents. Journal o f the Academy o f Marketing Science. 21(4), 281-292. Draper, N. and Smith, H. (1966). Applied Regression Analysis. New York, NY: Wiley. Ellerbock, M. J. (1981). Improving coupon conversion studies: A comment. Journal o f Travel Research. 19(4), 37-38. Engel, J. F., Blackwell, R. D. & Miniard, P. W. (19901. Consumer behavior ( 6 th Eds.) Chicago, IL: The Dryden Press. Ethridge, F. M. (1982). A migration model o f pleasure travel. Sociolouical Spectrum . 2, 99-121. Etzel, M. J. & Wahlers, R. G. (1985). The use o f requested promotional material by pleasure travelers. Journal o f Travel Research. 23(4), 2-6. Fakeye, P. C. & Crompton, J. R. (1991). Image differences between prospective, first-time, and repeat visitors to the lower Rio Grande Valley. Journal o f Travel Research. 30(2), 10-16. Fesenmaier, D. R. & Johnson, B. (1989). Involvement based segmentation: Implication for travel marketing in Texas. Tourism Management. 10(4), 293-300. Fesenmaier, D. R. & Vogt, C. A. (1993). Evaluating the economic impact o f travel information provided at Indiana welcome centers. Journal o f Travel Research. 31(3), 33-39. Fesenmaier, D. R., Vogt, C. A , & Stewart, W P. (1993). Investigating the influence o f welcome center information on travel behavior. Journal o f Travel Research. 3 1(3), 47-52. Fick, G. R. & Brend Ritchie, J R. (1991). Measuring service quality in the travel and tourism industry. Journal o f Travel Research. 30(2), 2-9. Fienberg, S. E. (1991). The analysis o f cross-classified categorical data. Cambridge: MIT Press 143 Fleming, W & Toepper, L (1990) Economic impact studies: relating the positive and negative impacts to tourism development. Journal o f Travel Research. 29(1), 35-41. Fodness, D. (1990). Consumer perceptions o f tourist attractions. Journal o f Travel Research. 28(4), 3-9 Fowleer, F. J. (1988). Survey research m ethods. Newbury Park, CA: Sage. France, L (1991). An application o f the tourism destination area life cycle to Barbados. Tourist Review. 46(3), 25-31 Fridgen, J. D. (1990). Dimensions o f tourism . E.Lansing, MI: Educational Institute o f the American Hotel & Motel Association. Gilbert, E. S. (1969). The effect o f unwqual variance-covariance matrices on Fisher's linear discriminant function. Biometrics. 25, 505-516. Gitelson, R. J. (1986). 1985 North Carolina coupon conversion study: An analysis o f those individuals who requested the North Carolina information packet. North Carolina Division o f Tourism. Gitelson, R. J. & Crompton, J. (1983). The planning horizons and sources o f information used by pleasure vacationers. Journal o f Travel Research. 21(3), 2-7. Gitelson, R. J. & Crompton, J. (1984). Insights into the repeat vacation phenomena. Annals o f Tourism Research. 11(2), 199-218. Gitelson, R. J. & Perdue, R. R. (1987). Evaluating the role o f State W elcome Centers in disseminating travel related information in North Carolina. Journal o f Travel R esearch. 25(4), 15-19. Gyte, D. M. & Phelps, A. (1989). Patterns o f destination repeat business: British Tourists in Mallorca, Spain. Journal o f Travel Research. 28(1), 24-28. Hagemann, R. P. (1981). The determinants o f household vacation travel: some empirical evidence Applied Economics. 13(2), 225-234. Hanushek, E., & Jackson, J. (1977). Statistical methods for social scientists. New York, N Y : Academic Press. Harris, C. C., Tynon, J. F., & McLaughlin, W. J. (1990) A comprehensive method for study leisure travel Journal o f Travel Research. 29(2), 39-44 144 Hill, J B., McDonald, C. D., and Uysal, M. (1990). Resort motivations for different life cycle stages. Visions in Leisure and Business. 8(4), 18-27. Holecek, D. F. (1993). Trends in world-wide tourism. In van Lier, H. N. & Taylor, P. D (Eds ). New challenges in recreation and tourism planning (pp: 17-33). New York, NY: Elsevier Science Publishers. Holloway, J. C. & Plant, R. V. (1988). M arketing for tourism . London: Pitman. Hosmer, D. W. & Lemeshow, S. (1989). Applied logistic regression. New York, NY: John Wiley & Sons. H ouston, M. J. & Ford, N. M. (1976) Broadening the scope o f methodological research in mail surveys. Journal o f Marketina Research. 13 (November), 397-403. H ouston, M. J. & Nevin, J. (1977). The effects o f source and appeal on mail survey response patterns. Journal o f Marketing Research. 14(August), 374-378. H ow ard, D. & Gitelson, R. (1989). An analysis o f the differences between state welcome center uses and nonusers: A profile on Oregon vacations Journal o f Travel Research. 27(4), 38-40. H ow ard, D. R., Lankford, S. V., & Havitz, M. E. (1992). A response to Roehl's " A comment concerning 'A method for authenticating pleasure travel expenditure' ", Journal o f Travel Research. 30(4), 19-23. H ow ard, D. R., Lankford, S. V., & Havitz, M. E. (1991). A method for authenticating pleasure travel expenditure. Journal o f Travel Research. 29(4), 19-23. Hsieh, S. & O'Leary, J. (1993). Communication channels to segment pleasure travelers. Journal o f Travel & Tourism. 2(2/3),57-75. Hunt, J. D. & Dalton, M. J. (1983). Comparing mail and telephone for conducting coupon conversion studies. Journal o f Travel Research. 21(3), 16-18. Hunt, J. D. & Layne, D. (1991). Evolution o f travel and tourism terminology and definitions. Journal o f Travel Research. 29(4), 7-11. Iso-Aloha, S. E. (1982). Toward a social psychological theory o f tourism motivation: A rejoinder. Annals o f Tourism Research. 9(2), 256-262. Jefferson, A. & Lickorish, L. (1988). Marketing tourism - A practical guide. Longman, Harlow. 145 Johnson, R. R. & Messmer, D. J. (1991). The effect o f advertising on hierarchical stages in vacation destination choice. Journal o f Advertising Research. 31 (December), 18-24. Kanuk, L & Berenson, C. (1975). Mail surveys and response rates A literature review Journal o f M arketing Research. 12(November), 440-453. Kent, P. (1990) The role o f place image in the holiday choice process o f potential tourists. Unpublished PhD thesis, Department o f Geography, University o f Reading, Reading. Klecka, W. R. (1980). Discriminant analysis. Beverly Hill, CA: Sage Publications. Kotler, P. H. (1984). M arketing management - analysis, planning, control (5th eds ). Englewood Cliffs, NJ: Prentice-hall, Inc. Kreisman, R. (1982). 1982 Alaskan Meltdown: Tourism warms up / Louisiana's dream o f a campaign/ Rhode Islanders in a proud state. Advertising A ge. 53, 30-31 Krippendorf, J. (1987). The holiday makers. London, Heinemann. Laskey, H. A., Seaton, B., & Nicholls, J. A. F. (1994). Effects o f strategy and pictures in travel agency advertising. Journal o f Travel Research. 32(4), 13-19. Latham, J. (1991). Bias due to group size in visitor surveys. Journal o f Travel Research. 29(4), 32-35. Law, E. (1995). Tourist destination management: issue, analysis and police. New York, NY: Routledge. Lewis, R. & Chambers, R E. (1989). Marketing leadership in hospitality. New York, NY: Van Nostrand Reinhold. Linden, F. (1980). The business o f vacation traveling. Across the Board. 17(4), 72-75. Linsky, A. S. (1975). Stimulating responses to mail questionnaires: A review. Public Opinion Quarterly. 39 (Spring), 82-101. Loker, L. E. & Perdue, R. R. (1992). A benefit-based segmentation o f a nonresident summer travel market. Journal o f Travel Research. 31(1), 30-35. M anfredo, M. J. (1989). A test o f assumptions inherent in attribute-specific advertising. Journal o f Travel Research. 27(3), 8-13. 146 M anfredo, M. J., Bright, A. D., and Haas, G. E. (1992). Research in tourism advertising. In M. J. Manfredo (Ed.) Influencing human behavior, (pp. 327-368). Champaign, IL: Sagamore Publishing Inc. Mak, J., Moncur. J., & Yonamine, D. (1977). Determinants o f visitor expenditures and visitor length o f stay: A cross-section analysis o f U.S. visitors to Hawaii. Journal o f Travel Research. 15(3), 5-8. Mak, J., Moncur. J., & Yonamine, D. (1977). How or how not to measure visitor expenditures. Journal o f Travel Research. 16(1), 1-4. M cCool, S. F. (1991). Using probabilistic incentives to increase response rates to mail-return highway intercept diaries. Journal o f Travel Research. 30(2), 17-19 M cIntosh, R. W. & Goeldner, C. R. (1990). Tourism: principles, practices, philosophies ( 6 th eds.) New York, NY: John Wiley & Sons, Inc. Megyesy, T. A. (1982). The conversion from inquirer to traveler in the state o f Pennsylvania. Pennsylvania Travel Review.3(2). 1-5. Messmer, D. J. & Johnson, R. R. (1993). Inquiry conversion and travel advertising effectiveness. Journal o f Travel Research. 31(4), 14-21. Mill, R. C. & Morrison, A. M. (1992). The tourism system: An introductory text (2nd Eds ). Englewood Cliffs, NY: Prentice Hall. M ok, H. M. (1990). A quasi-experimental measure o f the effectiveness o f destination advertising: Some evidence from Hawaii. Journal o f Travel Research. 24(1), 30-34. M orrison, M. A. (1989). Hospitality and Travel M arketing. Albany, NY: Delmar Publishers Inc. M outinho, L. (1987). Consumer behavior in tourism. European Journal o f M arketing , 21(10), 5-44. Muha, S. (1977). Who uses highway welcome centers? Journal o f Travel Research. 15(3), 1-4. Murray, E. J. (1964). Motivation and emotion. Englewood Cliffs, N.J.: Prentice-Hall Inc. Nichols, C. M. & Snepenger, D. J. (1988). Family decision making and tourism behavior and attitudes. Journal o f Travel Research. 26(4), 2-6, 147 Nolan, S. D. Jr. (1976) Tourists’ use and evaluation o f travel information sources: summary and conclusions. Journal o f Travel Research. 14(3), 6 - 8 . Pearce, P. L. and Caltabiano, M L (1983). Inferring travel motivation from travelers' experiences. Journal o f Travel Research. 22(2), 16-20. Perdue, R. R. (1986). The influence o f unplanned attraction visits on expenditures by travel -through visitors. Journal o f Travel Research. 25(1), 14-19. Perdue, R. R. (1985). The 1983 Nebraska visitor survey: Achieving a high response rate with a diary questionnaire. Journal o f Travel Research. 24(2), 23-26. Perdue, R. R. (1985). Segmenting state travel information inquirers by timing o f the destination decision and previous experience. Journal o f Travel Research. 23(3), 6 - 11. Perdue, R. R. & Botkin, M. R. (1988). Visitor survey versus conversion study. Annals o f Tourism Research. 15(1), 76-87. Perdue, R. R. & Gustke, L. D. (1992). The influence o f multiple trips on inquiry conversion research results. Journal o f Travel Research. 30(4), 27-30. Perdue, R. R. & Pitegoff, B. E. (1990). Methods o f accountability research for destination marketing. Journal o f Travel Research. 28(4), 45-49. Pitts, R. E. & W oodside, A. G. (1986). Personal values and travel decisions. Journal o f Travel Research. 25(1), 20-25. Pizam, A. (1990). Evaluating the effectiveness o f travel trade shows and other tourism sales-promotion techniques. Journal o f Travel Research. 29(1), 3-8. Pizam, A. (1978). Tourism's impacts: the social costs to the destination community as perceived by its residents. Journal o f Travel Research. 17(4), 8-12. Plog, S. C (1974). Why destination areas rise and fall in popularity. The Cornell H R A Quarterly. 14(4), 55-58. Plog, S. C (1990). A carpeter's tools: An answer to Stephen L. J. Smith's review o f psychocentrism/allocentrism. Journal o f Travel Research. 28(4), 40-45 Przeclawski, K. (1986). Humanistic foundations o f tourism . Warsaw: Institute o f Tourism. 148 Pyo, S. & Uysal, M (1990). Regional implications in tourism motivation. Unpublished manuscript, Department o f Parks, Recreation and Tourism M anagement, Clemson University, Clemson, S C Reid, L J. & Andereck, k. L. (1989). Statistical analyses use in tourism research. Journal o f Travel Research. 28(2), 21-25. Reid, L. J. & Reid, D. R. (1993). Communicating tourism supplier services: Building repeat visitor relationships. Journal o f Travel and Tourism M arketing. 2(2/3), 3-19. Reynolds, H. T. (1984). Analysis o f nominal data (2nd eds ). Newbury Park, CA: Sage Publications, Inc. Ricci, P. R. & Holland, S. M. (1992). Incentive travel: recreation as a motivational medium. Tourism M anagement. 13(3), 288-296. Ritchie, J. R. (1975). Some critical aspects o f measurement theory and practice in travel research. Journal o f Travel Reseach. 14(1), 1-10. Roehl, W. S. (1992). A comment concerning " A method for authenticating pleasure travel expenditures. Journal o f Travel Research. 30(4), 67-68. Roehl, W. S. & Fesenmaier, D. R. (1992). Risk perceptions and pleasure travel: An exploratory analysis. Journal o f Travel Research. 30(4), 17-26. Ronkainen, I. A. & Woodside, A. G. (1987). Advertising conversion studies. In Travel tourism and hospitality research, edited by Ritchie, J. R. B. & Goeldner, C. R. New York, NY: Wiley. Sawuer, A. G. & Ball, A. D. (1981). Statistical pow er and effect size in marketing research. Journal o f M arketing Research. 18(August), 275-290. Schroeder, T D (1983). Use o f multiple regression in recreation research: A discussion o f several issues Journal o f Leisure Research. 15(3), 247-250 Shih, D. (1982). Pennsylvania's toll free calls revisited. Pennsylvania Travel Review. 3(2), 6 - 8 . Shimp, T. A. (1990). Promotion management and marketing communication. Chicago, IL: Dryden Press. 149 Siegel, W. & ZifF-Levine, W. (1990). Evaluating tourism advertising campaigns: Conversion vs advertising tracking studies. Journal o f Travel Research. 28(3), 51-55. Silberman, J. & Klock, M. (1986) An alternative to conversion studies for measuring the impact o f travel ads. Journal o f Travel Research. 24(4), 12-16. Simon, J. L. & Arndt, J. (1980). The shape o f the advertising response function, Journal o f Advertising Research. 20(Aug/Sept), 11-28. Smith, L. J. S. (1983). Recreation geography. London and New York: Longman. Snepenger, D. (1987). Segmenting the vacation market by novelty seeking role. Journal o f Travel Research. 25(2), 8-12. Snepenger, D , Meged, M., Snelling, M , & Worrall, K. (1990). Information search strategies by destination-naive tourist. Journal o f Travel Research. 29(1), 13-16. Spotts, D. M. & Mahoney, E. M. (1991). Segmenting visitors to a destination region based on the volume o f their expenditures. Journal o f Travel Research. 29(4), 24-31. Sternthal, B. & Craig, C. S. (1982). Consumer behavior. Englewood Cliffs, NJ: Prentice-Hall. Stewart, W. P., Anderson, B. S., Fesenmaier, D. R., & Lue, C.C. (1993). Highway welcome center surveys: Problems with nonresponse bias. Journal o f Travel Research. 31(3), 53-57. Stewart, W. P., Lue, C.C., Fesenmaier, D. R., & Anderson, B. S. (1993). A comparison between welcome center visitors and general highway auto travelers. Journal o f Travel Research. 31(3), 40-46. Stynes, D. J. (1985). M arketing tourism. Journal o f Physical Education an D ance. 54(4), 21-23. Tierney, P. T. (1993). The influence o f state travel information centers on tourist length o f stay and expenditures. Journal o f Travel Research. 31(3), 28-32. Uni, S. & Crompton, J. L. (1992). The roles o f perceived inhibitors and facilitators in pleasure travel destination decisions. Journal o f Travel Research. 30(3), 18-25. Uni, S. & Crompton, J. L. (1990). Attitude determinants in tourism destination choice. Annals o f Tourism Research. 17(3), 432-448. 150 U. S. T ravel D ata C enter. (1995). Survey o f state travel o ffices 1 9 9 4 -1 9 9 5 . W ashington, DC: A uthor. U. S. Travel Data Center. (1991). The 1990-1991 economic review o f travel in America. W ashington, DC: Author. U. S. Travel Data Center. (1987). The 1986-1987 economic review o f travel in America. W ashington, DC: Author Uysal, M. & Crompton, J. (1985). An overview o f approaches used to forecast tourism demand. Journal o f Travel Research. 23(4), 7-15. Vincent. V. C & Santos, G. (1990). Winter Texans: two segments o f the senior travel market. Journal o f Travel Research. 29(1), 9-12. Wall, G. (1992). Towards a tourism typology. In J.G. Nelson, R. Butler, & G. Wall (Eds.) Tourism and sustainable development: monitoring, planning, managing, (pp.45 -58) London, ONT: University o f Waterloo. Walsh, A. (1987). Teaching understanding and interpretation o f logit regression. Teaching Sociology. 15(April), 178-183. W aters, S. (1988). Travel industry world yearbook: The big picture. New York, NY: Child and W aters Inc. W eaver, P. & McCleary, K. W. (1984). A market segmentation study to determine the appropriate Ad/Model format for travel advertising. Journal o f Travel Research. 23(1), 12-16. Wicks, B. E., & Schuett, M. A (1991). Examining the role o f tourism promotion through the use o f brochures. Tourism M anagement. 12(4), 301-321. Wind, Y (1978). Issues and advances in segmentation theory. Journal o f M arketing Research. 15(Aug), 317-337. Witt, S. F., Gammon, S. & White, J. (1992). Incentive travel: overview and case study o f Canada as a destination for the UK market. Tourism M anagement. 13(3), 275-287. W oodside, A. G. (1990). Measuring advertising effectiveness in destination marketing strategies. Journal o f Travel Research. 29(2), 3-8. W oodside, A. G. (1985). Theoretical and applied relevance o f consumer limits to processing advertising Journal o f advertising. 14(1), 57-58 151 W oodside, A. G. (1981). M easuring the conversion o f advertising coupon inquiries into visitors. Journal o f Travel Research. 19(4), 38-41 W oodside, A G & Carr, J. A (1988) Consumer decision making and competitive marketing strategies: Applications for tourism planning. Journal o f Travel Research. 26(3), 2-7 W oodside, A. G. & Glenesk, G. B. (1984). Thought processing o f advertisements in low versus high noise conditions. Journal o f Advertising. 13(2), 4-11 33. W oodside, A. G. & Jacobs, L. W. (1985). Step two in benefit segmentation: Learning the benefits realized by major travel markets, Journal o f Travel Research. 24(1), 7-13. W oodside, A. G. & Lysonski, S (1989) A general model o f traveler destination choice Journal o f Travel Research. 27(4), 8-14 W oodside, A. G., Moore, E., Bonn, M. A., & Wizeman, D. (1986). Segmenting the timeshare resort market, Journal o f Travel Research. 24(3), 6-11. W oodside, A. G. & Motes, W H (1982) Image vs direct-response advertising. Journal o f Advertisinu Research. 22(Dec./Jan ), 31-37. W oodside, A. G. & Motes, W. H. (1981). Sensitivities o f market segments to separate advertising strategies. Journal o f M arketing. 45(Winter), 63-73. W oodside, A. G. & Motes, W. H. (1979). Evaluating the consumer profiles o f general response, direct response, and reader service inquirers from magazine advertising. Journal o f Advertising . 8(1), 19-25. W oodside, A G & Reed, D. (1974). Tourism profiles versus audience profiles: Are upscale magazines really upscale? Journal o f Travel Research. 12(4), 17-23. W oodside, A. G. & Ronkainen, I. A. (1984). How serious is nonreponse bias in advertising conversion research. Journal o f Travel Research. 22(4), 34-37. W oodside, A. G. & Ronkainen, I. A. (1982). Travel advertising: newspapers versus magazines. Journal o f Advertising Research. 22(June/July), 39-43. W oodside, A. G. & Sherrell, D. (1977). Traveler evoked, inept, and inert sets o f vacation destination. Journal o f Travel Research. 16(1), 14-18. W oodside, A. G. & Soni, P. K (1991). Direct-response advertising information: profiling heavy, light, and nonusers. Journal o f Advertising Research. 3 l(D ecem ber), 26-36. 152 W oodside, A G. & Soni, P. K. (1988). Assessing the quality o f advertising inquiries by mode o f response. Journal o f Advertising Research. 28(August/September), 31-37. W oodside, A G & Trappey, R J III (1992) Finding out why customers shop your store and buy your brand: automatic cognitive processing models o f primary choice. Journal o f Advertising Research. 32(N ov./D ec), 59-78. W oodside, A. G. & Wilson, E. J. (1985). Effects o f consumer awareness o f brand advertising on preference. Journal o f Advertising Research. 25(Aug./Sept ), 41-48. Yochum, G. R. (1985). The economics o f travel advertising revisited. Journal o f Travel Research. 24(2), 9-12. Yuan, S. & McDonald, C. D. (1990). Motivational determinate o f international pleasure time. Journal o f Travel Research. 29(1), 42-44. Yuan, T., Fridgen, J. D , Hsieh, S. & O'Leary, J. T. (1995). Visiting friends and relatives travel market: The Dutch case. Journal o f Tourism Studies. In press. APPENDICES APPENDIX A SAMPLE OF FRANKENMUTH INQUIRERS QRUESTIONNAIRE 153 QUESTIONNAIRE 1994 Survey of Frankenmuth Inquirers I Travel, Tourism, and Recreation Resource C enter M ichigan State University 154 19 94 SURVEY OP FRANKENMUTH INQUIRERS T ha nk y o u v a r y n u c h f o r p a r t i c i p a t i n g i n o u r a c u d y l P l a a s a a n s v a r t h a f o l l o w i n g q u e s t i o n s s e q u e n t i a l l y , u n l e s s d i r e c t e d t o s k ip o v e r a b lo c k o f one o r a o r e q u e s t i o n s by th a s ta te m e n t "GO TO QUESTION. . . . " P l e a s e f o l l o w t h e s e "GO TO" d i r e c t i o n s v e r y c a r e f u l l y . 1. How d i d y o u o b t a i n t h e p h o n e num be r o r a d d r e s s o f t h a F r a n k e n m u t h C h am b er o f C om m e rc e/ V i s i t o r s Bureau? } ) ] j j ] 2. N ewspaper a r t i c l e N ewspaper a d v e r t i s e m e n t M agazine a r t i c l e M agazine a d v e r t i s e m e n t R adio a d v e r tis e m e n t T e le v is io n ad v ertisem en t P r i o r t o r e q u e s t i n g i n f o r m a t i o n o n _____ / ______ / ( ) Yes | ) No ] j ) j F rien d /re letiv e/c o -w o rk er Frankenm uth a r e a b u s i n e s s A t a t r a v e l show T ravel agent R egional t o u r i s t a sso c iatio n j M ic h ig a n T r a v e l B ureau T elep h o n e d i r e c t o r y D ire c to ry a s s is ta n c e B ro ch u re O th er; p le a se s p e c i f y : ________________ , h ad you e v e r v i s i t e d F rankenm uth? » CO TO QUESTION 3 I 2 a . A p p r o x im a te ly when d i d th e m o s t r e c e n t o f t h e s e v i s i t s d a t e s f o r up Co t h r e e v i s i t s . ) M o n th 2b. Year M on th Y ear occur? (P lease p ro v id e M onth Y e ar D id y o u r e q u e s t i n f o r m a t i o n f r o m t h e F r a n k e n m u t h C ham ber o f C o m m e r c e / V i s i t o r s B ureau to h e l p p l a n any o f th e s e v i s i t s ? [ ] Yes [ ] No ( J D o n ' t re m e m b e r GO TO QUESTION 3 3. Was t h e i n f o r m a t i o n y o u r e q u e s t e d fr o m t h e F r a n k e n m u t h C ha m be r o f C o m m e r c e / V i s i t o r s B u r e a u o n ______ / ______ / ______ f o r YOUR USE o r f o r USE BY SOMEONE ELSE? [ 4. ] F o r my u s e ( ] F o r u s e b y so m eone e l s e When y o u r e q u e s t e d i n f o r m a t i o n , was a v i s i t ( { [ t o F rankenm uth. ] A lre a d y d e c i d e d upon ] B eing c o n s id e re d j Not b e in g c o n s id e re d ,b u t d e s ir e d in fo rm a tio n 2 an yway » GO TO QUESTION 26 . . . 155 5. D i d y o u v i s i t F r a n k e n m u t h AFTER REQUESTING I n f o m a t i o n F ra n k e n m u th C ham ber o f C o m m e r c e /V is ito r s B u re a u ? ] Yes ) No- 5a. A re you c o n s i d e r i n g a v i s i t ( ) Yes CO TO QUES­ TIO N 6 ON NEXT PAGE f F rankenm uth? ) N o -------- ►GO TO QUESTION 5 c 5 b . When w i l l [ [ j [ to from on th i s v i s i t m ost l i k e l y ] V lth in th a ] 6 - 6 m o nths ) 7 -9 m onths ] 10 o r is o re next 3 in th e In th e m o n th s occur? m o nths fu tu re fu tu re In th e f u tu r e 5 c * D id y o u r e c e i v e a n y I n f o r m a t i o n from t h e F r a n k e n m u th Cham ber o f C o m m e r c e /V is ito r s B u re a u i n r e s p o n s e t o your req u est? I * GO TO QUESTION 26 [ ) Yes 5 d . Vas th e I n f o r m a t i o n y ou r e c e i v e d t h e I n f o r m a t i o n you req u ested ? [ ] Yes [ ) No [ 1 P a rtially 5 e . A p p r o x i m a t e l y when d i d y o u r e c e i v e D ay D id y o u r e a d t h e b r o c h u r e s / No - Yes In fo rm atio n ? / __________ / M onth 5f. th e Y ear t h a t w ere 5g. sent to you? U hat d i d you do w ith th e m ? ( ] K ep t them f o r f u t u r e referen ce I n Q u e s t i o n s 5h and ( ] D i s c a r d e d them 51 s c a l e s f r o m 1 t o 7 j ) C ave th em t o someone a r e s h o w n . On t h e s e [ ) O th er a c a le s 1 and 7 r e p r e ­ GO TO QUESTION 28 s e n t ex trem es t h a t have been la b e le d in th e c a s e o f e ach q u e s ti o n ; th e v a lu e s b etw een 1 and 7 r e p r e ­ s e n t d e g re e s b etw een th e s e e x tre m e s. P le a s e a n sw e r each q u e s t i o n b y c i r c l i n g ONE n u m b e r I n t h e c a a e o f e a c h s c a l e . 5h. To w h a t e x t e n t d i d t h e b r o c h u r e s d e c r e a s e I n t e r e s t I n v i s i t i n g Fran k en m u th 7 1 or 2 G reatly in c re a se d my i n t e r e s t G reatly decreased my I n t e r e s t 51. In crease your How w o u l d y o u r a t e th e o v e r a ll q u a lity o f th e brochures? 1 E x cel­ le n t T e rrib le CQ TO QUESTION 28 th e 156 6 . A p p r o x i m a t e l y w ha n d i d y o u d e p a r t o n y o u r f l r a t t r i p I n v o l v i n g a v i s i t t o F r a n k e n i s u t h a f t e r r e q u e s t i n g I n f o r m a t i o n f r o m t h e F r a n k a n n u t h C ham ber o f C o m m e r c e / V i s i t o r s B u r e a u , o n ______ / ______ / ______ ? / Month / Day Year _ MOTE: IM THE FOLLOWING QUESTIONS, "THIS TRIP" REFERS TO THE TRIP THAT BEGAN ON THIS DATE— * 7. P r i o r t o t h i s t r i p , d i d y ou o b t a i n any i n f o r m a t i o n a b o u t F ran k e n m u th o t h e r t h a n t h a F r a n k e n m u t h C ham ber o f C o m m e r c e / V i s i t o r s B u r e a u ? [ ] Yes [ ] No from any s o u r c e s >-G0 TO QUESTION 0 7 a . W hat o t h e r s o u r c e ( s ) ? ( ( ( ( 8. ] ) ) ) j j ( ( [ j ] T r a v e l show j Travel agent ] M ich ig an T r a v e l B ureau j R egional t o u r i s t a s s o c ia t io n j T elephone d i r e c t o r y ) B rochure j O ther; p le a se sp ecify : N ewspaper a r t i c l e N ewspaper a d v e r t i s e m e n t M agazine a r t i c l e M agazine a d v e r t i s e m e n t R adio a d v e r tis e m e n t T e le v is io n advertisem en t F r le n d /ra la tlv e /c o 'W o rk e r F rankenm uth a r e a b u s in e s s ) ) Was F r a n k e n m u t h t h a PRIMARY DESTINATION o f t h i s I J Yes | | No- 8 a . W hat was t h e p r i m a r y d e s t i n a t i o n ? 9 . What w as t h e PRIMARY PURPOSE o f t h i s ( ( { j trip ? trip ? ] R ec re a tio n /p leasu re ] B usin ess o r co n v en tio n /m eetin g j C o m b in e d b u s i n e s s a n d r e c r e a t i o n / p l e a s u r e J Other; please specify: _________________________ 1 0 . D i d y o u s p e n d a n y NICHTS AUAY FROM HOME o n t h i s trip 7 GO TO QUESTION 11 10a. D id y o u s p e n d a n y n i g h t s IN FRANKENMUTH o n t h l a 10b. Yes In v h at c i t i e s trip ? d i d you sp en d t h e n i g h t ? CO TO QUESTION 11 1 0 c . How many n i g h t s d i d y o u s p e n d IN FRANKENMUTH? ___________ lO d . Where w e r e t h e s e n i g h t s IN FRANKENMUTH s p e n t ? [ [ j (C heck a l l t h a t a p p ly ) ] H o te l o r m otel ( ] F r i e n d ' s / r e l a t i v e ' s home ] Bed & B r e a k f a s t ( j S e c o n d home y o u own j C am pgro und [ ) O t h e r ; p l e a s e s p e c i f y : ______________________ CO TO QUESTION 11 ________ _________________ 4 157 1 1 . When y o u v i s i t e d F r a n k e n m u t h on t h i s such a s a m otor coach t o u r group? I trip , w e re y o u a member o f a n o r g a n i z e d t o u r g r o u p , 1 1 a . Did a n y o n e a cc o m p a n y y o u o n t h i s t r i p ? } Yes [ ] Yes ( ) No > GO TO QUESTION 12 l i b . How many p e r s o n s ( n o t I n c l u d i n g y o u r s e l f ) a c c o m p a n i e d y o u ? ______________ l i e . Were t h e s e p e r s o n a : [ i | [ [ ] F riends j R elativ es ] F r ie n d s and r e l a t i v e s j B u sin ess a s s o c i a t e s ) O ther; p le a s e s p e c ify : CO TO QUESTION 12 U e ' d a l s o l l l c e t o know a p p r o x i i a a t e l y how much ( I f a n y ) money y o u r " s p e n d i n g u n i t " s p e n t i n F r a n k e n m u t h o n t h i s t r i p , a " s p e n d i n g u n i t " i s a n i n d i v i d u a l who p u r c h a s e s t h i n g s f o r h i m s e l f o r h e r s e l f , o r a g r o u p t h a t p u r c h a s e s t h i n g s a s a g r o u p . An I n d i v i d u a l who p a y s h i s o r h e r own e x p e n s e s I s o n e s p e n d i n g u n i t w h e t h e r h e o r s h e i s t r a v e l i n g a l o n e o r w i t h o t h e r s . Two m a r r i e d c o u p l e s t r a v e l i n g t o g e t h e r a r e two s p e n d i n g u n i t s I f t h e y p u r c h a s e t h i n g s s e p a r a t e l y , a l t h o u g h t h e y may s p l i t some e x p e n s e s . 1 2 . U h i l e y o u w e r e I n F r a n k e n m u t h on t h i s t r i p , UNIT s p e n d a n y money? ( ) Y es t ) No- 1 2 a . How many p e r s o n s . 12b. d i d y o u o r a n y o t h e r members o f y o u r SPENDING -CO TO QUESTION 13 I n c l u d i n g y o u r s e l f , w e r e I n y o u r SPENDING UNIT o n t h i s t r i p ? A p p r o x i m a t e l y how much money d i d y o u r SPENDING UNIT s p e n d FRANKENMUTH o n e a c h o f t h e f o l l o w i n g i t e m s ? on th is trip IN PLEASE DO NOT LEAVE ANY SPACES BLANK; WRITE "0 " TO INDICATE NO EXPENDITURES ON A GIVEN ITEM. INCLUDE ONLY EXPENDITURES MADE IN FRANKENMUTH. EXCLUDE FEES PAID TO MOTORCOACH OPERATORS. L o d g i n g ............................................................................................................................S. C am pin g f e e s ................................................................................................................$. G if ts , c r a f t s , so u v en irs, c lo th in g , an d /o r s p e c ia lty f o o d i t e m s ............................................................................................................... $, C r o c e r y a n d c o n v e n i e n c e s t o r e f o o d a n d b e v e r a g e s ......................$ R e s t a u r a n t a n d b a r m e a l s a n d d r i n k s ..................................................... G a s o lin e , o i l , r e p a i r s , and o th e r v e h i c l S 'r e l a t e d i t e m s . . $ R ental fe e s ( fo r g o lf c a r ts , cro ss-co u n try s k is , e t c . ) . . . $ G u i d e d t o u r s .............................................................................................................. A l l o t h e r i t e m s ...................... $ TOTAL............................................................................................................................... GO TO QUESTION 13 158 11. Was y o u r e x p e r i e n c e i n F r a n k e n m u t h o n t h i s t r i p . ( I I Much v o r a e th an e x ­ pected ( ] ) About w h a t y ou axpactad Somewhat v o r ia th an ex p ected i i I ) Som ew hat b e t t e r th an ex p ected Much b e t ­ t e r th an ex p ected L ik ely to v i s i t again C e r ta in to v ls L t again 16. How l i k e l y a r e y o u t o v i s i t F r a n k e n m u t h a g a i n ? I ) 1 5. A f t e r t h i s ( I } i i U n c e rta in w hether I w ill v i s i t again U n lik ely to v i s i t again C ertain to not v i s i t again trip , d i d you re c o m se n d a v i s i t t o F rankenm uth t o anyone? ) Yes [ J Ho ( ) D on’ t r e a e n b e r 1 6 . O v e r a l l , how w o u l d y o u r a t a F r a n k e n m u t h a s a t o u r i s t d e s t i n a t i o n o n a s c a l e f r o * 1 t o 7, w here 1 l a " t e r r i b l e * and 7 i s " e x c e l l e n t " ? P l e a s e c i r c l e one n u n b e r on t h e s c a l e b elow . E x cellen t Terrible Now w e ' d l i k e t o l e a r n a l i t t l e s o r e a b o u t y o u r r e q u e s t f o r C ham ber o f C o m m e r c e / V i s i t o r s B u r e a u o n ______/ ______ / ______ . In fo n aa tlo n from th e Frankenm uth 17. O ld y o u r e c e i v e an y I n f o r m a t i o n i n r e s p o n s e t o y o u r r e q u e s t ? [ ) Y es I ► C O TO QUESTION 28 ] No 1 7 a . Was t h e i n f o n a a t l o n y o u r e c e i v e d t h e i n f o n a a t l o n y ou r e q u e s t e d ? | ] Y es | 1 No [ ] P a rtially 1 7 b . A p p r o x i m a t e l y w he n d i d y o u r e c e i v e t h e i n f o r m a t i o n ? Month 17c. / Day / ________ Year D id y o u r e c e i v e I n f o r m a t i o n fr o m t h e F r a n k e n m u t h C h am b er o f C o m m e rc e/ V i s i t o r s B u r e a u BEFORE t h i s t r i p o r AFTER t h i s t r i p ? [ ( ) B efore t h i s t r i p j A fter th is tr i p — *»G0 TO QUESTION 16 ►CO TO QUESTION 26 1 6. D id y o u r e a d t h e b r o c h u r e s t h a t w e re s e n t t o y o u ? 1 8 a . What d i d y o u do w i t h th e m ? ( [ ] K e p t th e m f o r f u t u r e ( re fe re n c e ) D i s c a r d e d th e m ( CO TO QUESTION 28 1 9. D i d y o u c o n s u l t t h e b r o c h u r e s w h i l e v i s i t i n g F r a n k e n m u t h o n t h i s [ ] Yes [ J No ( } D o n ' t rem em ber 6 J C a v e th e m t o so m eone ] O ther trip ? 159 I n Q u e s t i o n s 20 t h r o u g h 2 7 , s c a l e s fr o m 1 t o 7 a r c show n. On t h e s e s c a l e s 1 a n d 7 r e p r e s e n t e x tre m e s t h a t have been l a b e l e d In th e c a s e o f each q u e s t i o n ; th e v a l u e s b e tw ee n 1 and 7 r e p r e s e n t d e g r e e s b e t w e e n t h e s e e x t r e m e s . P l e a s e a n s w e r e a c h q u e s t i o n b y c i r c l i n g ONE num ber on each s c a l e . 20. How u a e f u l was t h e In fo rm a tio n In th e b ro c h u re s? 1 Not a t a l l useful 21. 2 3 4 To w h a t e x t e n t w e r e t h e b r o c h u r e s 1 2 V e r y un* In te restin g 4 2 2 . To w h a t e x t e n t w e r e t h o b r o c h u r e s a t t r a c t i v e 23. To w h a t e x t e n t Frankenm uth? 1 G reatly decreased my i n t e r e s t 3 did 2 th e 6 7 E xtrem ely useful 5 6 7 V e ry In terestin g 5 6 i n t e r e s t i n g to read? 3 1 2 V e ry u n a ttra c tiv e 5 in design? 4 brochures decrease 3 or in crease 4 your 5 1 Had n o in flu en c e at a ll 25. 2 3 4 To w h a t e x t e n t d i d t h e b r o c h u r e s c a u s e t r i p th a n y ou would have o th e r w is e ? 1 N ot a t all 2 3 26. B a a e d o n y o u r e x p e r i e n c e o n t h i s 1 C om p letely In a c cu rate 27. How w o u ld y o u r o t e 1 T errib le 2 3 you t o S s p e n d m ore 4 5 in te re st 6 24. To w h a t e x t e n t d i d t h e b r o c h u r e s i n f l u e n c e y o u r d e c i s i o n t o v i s i t 7 Very attra c tiv e in v isitin g 7 G reatly increased ny i n t e r e s t Frankenm uth? 6 mone y I n 6 7 A ctu ally c a u s e d me to v i s i t F r a n k e n m u t h on t h i s 7 A great deal t r i p , how a c c u r a t e was t h e i n f o r m a t i o n i n t h e b r o c h u r e s ? 4 5 6 5 6 7 C om pletely a cc u ra te th e o v e r a l l q u a l i t y o f th e b ro c h u re s? 2 3 4 7 E x cellen t 160 P l e a s e a n s w e r t h e r e m a i n i n g q u e s t i o n s s o t h a t we w i l l be a b l e t o d e v e l o p a p r o f i l e o f t h e t y p e s o f p e o p l e who r e q u e s t I n f o r m a t i o n fr o m t h e F r a n k a n n u t h C ham be r o f C o m m e r c e / V i s i t o r s B u r e a u . Y o u r r e s p o n s e s w i l l , o f c o u r s e , r e m a i n s t r i c t l y c o n f i d e n t i a l ; t h e y w i l l s i m p l y be c o m b i n e d w i t h t h o s e o f o t h e r r e s p o n d e n t s t o com pute p e r c e n t a g e s a n d a v e r a g e s . l e t h e 21P CODE o f y o u r p e r m a n e n t la your gender? I 10 . H ale What I s y o u r ( [ [ 31. ] I r e s i d e n c e ? ___________________________ _ _ 28. What 2 9 . What ] F e m a le p r e s e n t e m p lo y m e n t s i t u a t i o n ? } W o r k in g f u l l t i m e j W orking p a r t tim e j T e m p o r a r il y unem ployed P lease c i r c l e [ ) Homemaker ( ) R etired j j S tudent [ ] O ther; p le a s e s p e c i f y : ____________ th e h i g h e s t y e a r o f fo rm al s c h o o li n g you have c o m p le ted . 1 2 3 4 5 6 7 8 G rade Sch o o l 1 2 3 4 H ig h S c h o o l 1 2 3 4 U ndergraduate c o lle g e e d u catio n 1 2 3 4 5 6 7+ G raduate c o l l e g e ed u catio n 3 2 . I n w h a t y e a r w e re y o u b o r n ? _________________ 3 3. How many p e o p l e r e s i d e i n y o u r h o u s e h o l d ( i n c l u d i n g y o u r s e l f ) ? _____ 36. How many EMPLOYED p e r s o n s a g e 18 o r o l d e r r e s i d e i n y o u r 35 . household? Do a n y c h i l d r e n u n d e r a g e 18 r e s i d e I n y o u r h o u s e h o l d ? { ] Yes [ ] No --•- CO TO QUESTION 36 3 3 a . What I s t h e ACE o f t h e o l d e s t c h i l d l i v i n g a t home? _ 3 5 b . What i s t h e ACE o f t h e y o u n g e s t c h i l d l i v i n g a t home? GO TO QUESTION 36 36. What was y o u r t o t a l 1993 h o u s e h o l d Inc ome b e f o r e t a x e s ? ( ( [ [ ) Under $ 1 5 ,0 0 0 ) $ 1 5,000 to $19,999 ] $20,000 to $24,999 J $25,000 to $34,999 ( ( | 1 ] $ 3 5 ,0 0 0 to $49,999 ) $ 5 0 ,0 0 0 to $74,999 ) $ 7 5 ,0 0 0 to $104,999 )$ 1 0 5 , 0 0 0 t o $ 1 1 9 , 9 9 9 { ( ( [ j ) ) ] $120,000 $135,000 $150,000 $300,000 to to to or $134,999 $149,999 $299,999 more THANKS FOE YOUR HELPI P l e a s e r e t u r n y o u r q u e s t i o n n a i r e I n t h e p o s t a g e - p a i d e n v e l o p e p r o v i d e d . I f you m is p la c e d t h i s e n v e lo p e , p le a s e r e t u r n th e q u e s t i o n n a i r e to : 1994 S u r v e y o f F r a n k e n m u t h I n q u i r e r s T r a v e l , T ourism , and R e c r e a t i o n R esource C e n te r 172 N a t u r a l R e s o u r c e s B u i l d i n g M ichigan S t a t e U n iv e r s ity E a s t L a n s i n g , MI 48824-1222 APPENDIX B RESULTS OF THE SURVEY OF FRANKENMUTH INQUIRERS 161 RESULTS OF THE 1994 SURVEY OF FRAN K E NM UT H IN Q U IR E R S Q1 H ow did you obtain the phone number or address o f the Frankenmuth Chamber o f Commerce/Visitors Bureau 9 (N=573) [ | [ [ [ | 2% 5% 10% 18% 1% 1% 1N ew spaper article 1N ew spaper advertisem ent ] M agazine article 1M agazine advertisem ent 1Radio advertisem ent ]Television advertisem ent | [ | [ [ 10% 1% 2% 2% 2% ] Fricnd/rclative/co-w orkcr [ ] Frankenm uth area business [ ) At a travel show [ ) Travel agent [ ] Regional tourist association! 12% 3 % 19% 8% 4% ] | ] ] ] M ic h ig a n T ra v e l B u re a u T e le p h o n e d ir e c to r y D ire c to ry a s s is ta n c e B ro c h u re O th e r Q2. Had you ever visited Frankenmuth? (N=594) [ 55% ] Yes [ 45% ] No Q2b. Did you requesting information from the Frankenmuth Chamber o f Commerce/Visitors Bureau to help plan any o f these visits? (N=320) [2 1 % ] Yes [70% ] No [9% ] Don't remember Q3. W as the information you requested from the Frankenmuth Chamber o f Commerce/Visitors Bureau for your use or for use by someone else? (N=586) [98% ] For my use [ 2% ] For use by someone else Q4. When you requested information, was a visit to Frankmuth. ..(N=574) [ 30% ] Already decided upon [ 65% ] Being considered [ 5 % ] Not being considered, but desired information anyway Q5. Did you visit Frankenmuth alter requesting information? (N=585) [ 44% ] Yes [ 56% ] No Q5a. Are you considering a visit to Frankenmuth? (N=328) [ 93% ] Yes [ 7% ] No Q5b. When will this visit most likely occur? (N=296) [ 47% ] Within the next 3 month [10 % ] 7-9 months in the future [ 3 1 % ] 4-6 months in the future [ 12% ] 10 or more months in the future 1 62 Q5c. Did you receive any information from the Frankenmuth Chamber o f Commerce/Visitors Bureau in response to your request 9 (N=326) [97% ] Yes [ 3% ] No Q 5 d. W as the information you received the information you requested? (N=31 5) [90% ] Yes [ 2% ] No [ 8 % ] Partially Q5f. Did you read the brochures that were sent to you? (N = 317) [ 96% ] Yes [ 4% ] No Q5g. What did you do with them if you didn't read the brochures that were sent to you? (N =l 2) [100% ] Keep for future reference [ 0% ] Gave them to someone [ 0% ] Discarded them [0% ] Other Q5h. To what extent did the brochures decrease or increase your interest in visiting Frankenmuth? (N=304; Mean= 5.4) [ 1% ] 1 [ 0% ] 2 [ 3% ] 3 [ 20% ] 4 [29% ] 5 [23% ] 6 G re a tly [24% ] 7 G ic .tlh d c u c .ts e d iiK ic .tM .- d m y in te re s t it» y i n t e r e s t Q5i. Flow would you rate the overall quality o f the brochures? (N= 305; Mean= 5.7 ) [ 0% ] 1 [ 2% ] 2 [ 2% ] 3 [ 1 1% ] 4 [22% ] 5 T c m lilc [30% ] 6 [33% ] 7 ly ic -llc n t Q7. Prior to this trip, did you obtain any information about Frankenmuth from any sources other than the Frankenmuth Chamber o f Commerce/Visitors Bureau? (N=252) [49%] Yes [51% ] No Q7A. What other sources? (N =l 19) [ 0% ] Newspaper article [ 0% ] N ewspaper advertisement [ 7%] M agazine article [ 7% ] M agazine advertisement [ 0% ] Radio advertisement [ 2% ] Television advertisement [33%] Friend/relative/co-worker [ 6 % ] Frankenmuth area business [ 1%] Travel show [ 5%] Travel agent [ 9%] Michigan Travel Bureau [ 3%] Regional tourist association [ 0%] Telephone directory [14%] Brochure [ 5% ] Other 163 Q 8 . Was Frankenmuth the PRIMARY DESTINATION o f this trip 9 (N=253) [8 3 % ] Yes [1 7 % ] No Q9 W hat was the PRIM ARY PURPOSE o f this trip? (N=255) [90%] Recreation/pleasure [ 4% ] Business or convention/meeting [ 5% ] Combined business and recreation/pleasure [ 1% ] Other Q10. Did you spend any NIGHTS AWAY FROM HOME on this trip 9 (N=255) 72% ] Yes [28%] No QlOa. Did you spend any nights in FRANKENMUTH on this trip? (N=184) [64% ]Yes [36%] No QlOc. How many nights did you spend IN FRANKENMUTH? (N =l 16; M ean= 1.84) QlOd. W here were these nights IN FRANKENMUTH spent? (N = l 17) [96% ] Hotel or motel [0% ] Friend's/relative's home [ 1% ] Bed & Breakfast [0% ] Second home you own [ 3% ] Campground [0% ] Other Q 1 1. When you visited Frankenmuth on this trip, were you a member o f an organized tour group, such as a motor coach tour group? (N=255) [4% ] Yes [96% ] No Q1 la. Did anyone accompany you on this trip? (N=241) [95%] Yes [ 5% ] No Q1 lb. H ow many persons (not including yourself) accompanied you?(N=224; M ean= 2.4) Q1 lc. W ere these persons: (N=226) [19%] Friend [ 0% ] Business associates [63%] Relatives [ 0% ] Other [18%] Friends and relatives 164 Q12. While you were in Frankenmuth on this trip, did you or any other members o f your SPEN D IN G UNIT spend any money‘s (N=253) [ 97% ] Yes [ 3% ] No Q12a. H ow many persons, including yourself, were in your SPENDING UNIT on this trip? (N=243 Mean=4 3) Q12b Approximately how much did your SPENDING UNIT spend on this trip in FRANKENM UTH on each o f the following items? (N=241) (M ean= (M ean= (M ean= (M ean= (M ean= (M ean= (M ean= (M ean= (M ean= (M ean= $ $ $ $ $ $ $ $ $ $ 81) 0) 147) 12) 90) 13) 0) 3) 10) 356) Lodging Camping fee Gifts, crafts, souvenirs, clothing, and/or specialty food items Grocery and convenience store food and beverages Restaurant and bar meals and drinks Gasoline, oil, repairs, and other vehicle-related items Rental fees (for golf carts, cross-country skies, etc.) Guided tours All other items TOTAL Q13 W as your experience in Frankenmuth on this trip.. (N=251, Mean=3.8) [ 0% ] 1. [ 4 % ] 2. [ 36%] 3. [ 32%] 4. [ 28%] 5. Much worse than expected Somewhat worse than expected About what you expected Somewhat better than expected Much better than expected Q14. How likely are you to visit Frankenmuth again? (N=253; Mean= 4.4) [1 % ] 1. [1 % ] 2. [ 9% ] 3. [ 40%] 4. [ 49%] 5. Certain to not visit again Unlikely to visit again Uncertain whether I will visit again Likely to visit again Certain to visit again Q15. After this trip, did you recommend a visit to Frankenmuth to anyone? (N -2 5 2 ) [ 8 6 % ] Yes [ 8 % ] No [6 % ] Don't know 165 Q16. Overall, how would you rate Frankenmuth as a tourist destination on a scale from 1 to 7, W here 1 is "terrible " and 7 is "excellent"? (N=251;M ean=5.9) [0% ] 1 [1% ] 2 [1% ] 3 [4%] 4 [29%] 5 [32%] 6 [33%] 7 T e rrib le E x c e lle n t Q17 Did you receive any information in response to your request? (N=252) [9 9 % ] Yes [l%]No Q17a. W as the information you received the information you requested? (N=242) [96%] Yes [0%] No [ 4%] Partially Q17c. Did you receive information from the Frankenmuth Chamber o f Commerce/Visitors Bureau BEFORE this trip or AFTER this trip? (N=243) [93%] Before this trip [ 7% ] After this trip Q18. Did you read the brochures that were sent to you? (N=232) [100% ] Yes [0% ] No Q18a. What did you do with them? ( N=l ) [ 0%] [ 0%] [ 100%] [ 0 %] Kept them for the future reference Discarded them Gave them to someone Other Q19 Did you consult the brochures while visiting Frankenmuth on this trip? (N=229) [82%] Yes [13%] No [5% ] Don't remember In Question 20 through 27, scales from 1 to 7 are shown. On these scales 1 and 7 represent extremes that have been labeled in the case o f each question; the values between 1 and 7 represent degrees between these extremes. Q20. How useful was the information in the brochures? (N=232; M ean=5.8) [ 0% ] 1 [0% ]2 [ 1% ] 3 [10% ] 4 [ 30%] 5 [25% ] 6 [ 34% ] 7 N o t at a l l E x tre m e ly u sefu l u sefu l Q21 To what extent were the brochures interesting to read? (N=232, Mean=5 6 ) [ 0% ] 1 V e ry u n in te r e s tin g [0% ]2 [ 3% ] 3 [ 12% ] 4 [ 31%] 5 [27% ] 6 [ 27% ] 7 V e ry m te ie s lin g 166 Q22. To what extent were the brochures attractive in design? (N = 231; M ean=5.8) [ 0% ] 1 [0% ] 2 [ 1% ] 3 [ 9% ] 4 [2 7 % ] 5 [33% ] 6 V crv unattractive [ 30% ] 7 Very attractive Q23. To what extent did the brochures decrease or increase your interest in visiting Frankenmuth? (N=232, Mean=5.7) [ 0% ] 1 [0% ] 2 [ 0% ] 3 [ 15% ] 4 [ 28%] 5 [26% ] 6 G reatly decreased m y interest [ 31% ] 7 G reatly i n creased my interest Q24. To what extent did the brochures influence your decision to visit Frankenmuth? (N=230; M ean=4.5) [ 16% ] 1 [3% ] 2 [ 7% ] 3 [16%] 4 [ 21%] 5 [ 2 1 %] 6 [ 1 6 % ] 7 1lad no inllucnce at all Actually caused me to visit Q25. To what extent did the brochures cause you to spend more money in Frankenmuth on this trip than you would have otherwise? (N=230; Mean=3.2) [31%] 1 [8 % ] 2 [ 1 2 % ] 3 [2 4 % ] 4 [ 13%] 5 [9% ] 6 [ 3% ] 7 Not at all A great deal Q26. Based on your experience on this trip, how accurate was the information in the brochures? (N=232; Mean = 6 1) [ 1% ] 1 [0% ] 2 [ 0% ] 3 [7% ] 4 [16%] 5 [29% ] 6 C om pletely inaccurate [ 47% ] 7 C om pletely accurate Q27. How would you rate the overall quality o f the brochures? (N=232; M ean= 6 . 1) [ 0% ] 1 [0% ] 2 [ 0% ] 3 [7% ] 4 [ 19%] 5 T errible Q29. What is your gender? (N=-594) [29%] Male [ 71%] Female Q30. What is your present employment situation? (N=594) [ 53%] W orking full time [ 13%] W orking part time [ 1% ] Temporarily unemployed [ 12% ] Homemaker [ 16% ] Retired [ 3% ] Student [ 2% ] Other [35% ] 6 [ 39% ] 7 Excellent 167 Q 3 1 What is the highest year o f formal schooling have you completed? (N=590; Mean=14 2) Q32. What is your age 9 (N=587; Mean=46.2) Q33. How many people reside in your household (including yourself )9 (N=593; M ean=2.8) [0%] 0 [10%] 1 [41%] 2 [19%] 3 [20%] 4 [ 8 %] 5 [1%] 6 [1%] 7 Q34 How many EM PLOYED persons age 18 reside in your household? (N=588; M ean=l .6 ) [17%] 0 [27%] 1 [43%] 2 [10%] 3 [3%] 4 Q35. Do any children under age 18 reside in your household? (N= 592; M ean=l .6 ) [36% ] 1 [64%] 2 Q35a What is the AGE o f the oldest child living at home? (N=208; M ean=10) Q35b. What is the AGE o f the youngest child living at home 9 (N=204; M ean=7.7) Q36. What was your total 1993 household income before taxes? (N=524) | 4% | | 5% | | 6% | 115%] Under $15,000 $15,000 to $19,999 $20,000 to $24,999 $25,000 to $34,999 [24%) [30%] 112%| | 3% ] $35,000 to $49,999 $50,000 to $74,999 $75,000 to $ 104,999 $105,000 to $119,999 [ 1%| |0%| [0%| |0%[ $120,000 to $134,999 $135,000 to $149,999 $ 150,000 to $299,999 $300,000 or more