James B. Faircloth is Assistant Professor of Marketing at the University of North Dakota. Dr. Faircloth received his DBA from Mississippi State University. He has published in such distinguished journals as the Journal of Food Products Marketing and Advances in Consumer Research.
Victoria P. Richard is Assistant Golf Course Superintendent at the Mississippi State University Golf Course. Ms. Richard received her B.S. degree from Mississippi State University. She is a member of the Golf Course Superintendents Association of America and has published in the Journal of Contemporary Business Issues and Sports Marketing Quarterly
Golf is a $20 billion industry in the U.S. and is expected to reach $40 billion by the year 2000 (Symonds 1989). While theNational Golf Foundation (NGF) predicts that 3650 new golf courses will have opened by the year 2000 [NGF1989], the nature of the golf industry is rapidly changing. The number of new membership golf courses is growing at a much slower rate than the number of new public or pay-as-you-play courses [Hick 1989].
Yet despite golf being the fastest growing sport in the U.S. [Nelson 1990] and recognizing that there are over 23 million golfers presently in the U.S. [NGF 1989], little is known about golf course choice intentions. In fact, no research appears to have examined in detail the attributes of golf courses or their operations which might be expected to influence choice intentions. This knowledge gap apparently applies to private as well as public courses. Thus, a significant amount of marketing is undertaken each year by golf course marketers in the absence of solid research on behavioral intentions. Consequently, marketers rely on intuition and subjective judgement when developing and implementing marketing campaigns for public courses.
We believe that relying on intuition and subjective judgement is not necessary, as the voluminous research on choice intentions for other product/service offerings and the general marketing literature provide a reasonable foundation for new research which models public golf course choice intentions. Research of this type is consistent with the calls by many researchers for extensions of the choice framework to other product/service categories [McFadden 1986].
The purpose of this study is to develop a model of public golf course choice intentions. In it regression analysis is used to investigate whether 17 golf course attributes exert a significant influence on public golf course choice intentions. Choice intentions are hypothesized to be a function of four sets of attributes:
Choice Intentions Models
While there are many possible specifications of choice intentions models, there are several generally accepted principles that are applied to the many possible specifications [Cooper and Nakanishi 1983].
Go to a review of the literature concerning choice intentions.
FIRST, choice offerings are described as combinations or bundles of attributes.
SECOND, consumers (e.g., golfers) evaluate offerings in terms of some or all of the attributes.
THIRD, these attributes are combined in some manner to form the utility of each offering for each consumer.
FOURTH, consumers are assumed to be utility maximizers choosing the offering with the highest utility. Finally, attributes of the consumer (e.g., demographics) can be included in the model to account for differences between consumers.
While these models suggest that many offerings have been the focus of choice intentions studies, there is a paucity of literature dealing with public golf course choice intentions. The only empirical studies on golf course consumers in general have been conducted by Piper  and the NGF .
Piper  investigated the relationship between the demand for public golf courses and several demographic variables. Income and age were found to have an inverse relationship with demand for public courses (as opposed to private, membership courses).
The NGF is responsible for the vast majority of empirical research on U.S. golfers. Much of these results are contained in the 1989 Golf Consumer Profile [NGF 1989]. Two-thirds of golfers surveyed cited price (i.e., greens fees) as the single most important factor in deciding on which golf course to play. Other attributes found to influence choice intentions were the number of other golfers on the course, the distance to the golf course from home, and the speed of play.
While the NGF study of U.S. golfers is useful, the results are only descriptive in nature. In addition, the NGF study focuses only on a subset of the attributes thought to influence golf course choice intentions. As such, the golf course general manager and the golf course professional are left with only a rudimentary understanding of the attributes thought to influence choice intentions.
The process suggested by Churchill  was used to generate a series of attributes thought to influence public golf course choice intentions. An initial pool of attributes was generated through discussions with golfers, golf professionals, and a review of the relevant literature. Churchill's  iterative process of attribute development, testing, and deletion was utilized to develop a relevant and manageable set of 17 attributes:
The speed of play is to my liking.
The price of play is reasonable for a course of this quality.
Getting convenient tee times is not a major problem
The course is conveniently located for regular play.
The course is not too long in terms of yardage.
The course layout fits my style of play (e.g., width, hazards, etc.).
The condition of the greens are excellent.
The condition of the fairways are excellent.
The course difficulty is appropriate for my level of play.
The attitude of the other golfers is friendly.
The course was designed by a well known
designer (e.g., Pete Dye, Jack Nicklaus, etc.).
The golf pro has helped me improve my game.
The course was recommended by someone I respect.
The restaurant or food service is excellent.
The course has several amenities such as a
pool, tennis courts, etc.
I enjoy shopping at the pro shop.
The practice facilities are excellent.
Ratings on each of the above attributes were assessed using ten point response scales anchored by "Strongly Disagree" (0) and "Strongly Agree" (9). Choice intention was measured by the item: "If given the opportunity, how likely would you be to play the course again?" The choice intentions item was anchored by "Very Unlikely" (0) and "Very Likely" (9).
The final version of the questionnaire was administered by personal interviews at a large Southeastern airport. Care was taken to ensure that the interviewers went to their designated locations on different days of the week and different times of the day to attempt to obtain a representative cross-section of the market.
A systematic sampling procedure was employed by the interviewers. The interviewers were instructed to survey every tenth person seated in the interviewers' designated areas. Response rates averaged 85 percent. This seemingly high response rate is not surprising given the interviewing environment. Most of the respondents were waiting for a flight or to meet someone at the airport. Several respondents expressed the view that they would complete the questionnaire since they had little else to do but wait. The personal interviews yielded 241 useable questionnaires.
Respondents were asked to fill out a four page questionnaire. One section consisted of a series of items designed to gather the name and location of the last public golf course played, choice intentions, and a set of 17 evaluative attributes concerning the named golf course. The evaluative attributes in this section were scrambled. The second section consisted of a standard demographic profile of respondents. The self-administered questionnaire required approximately 10 minutes to complete.
Results: Reliability And Factor Structure
Following Parasuraman, Zeithaml, and Berry , the 17 attributes thought to influence public golf course choice intentions were subjected to a Common Factor Analysis and an oblique rotation. The criterion used to generate factors was the average communality extracted as suggested by Hair, Anderson, Tatham, and Black . The results of the Factor Analysis appear in Table 2. The resulting four factor solution explained 77.60 percent of the variance. No cross loadings equaled or exceeded 0.3000 providing support for convergent and discriminant validity of the attributes.
While there was no a priori determination of how the attributes should load, the resulting solution was intuitively appealing. The factor names and interpretations were suggested by how the attributes loaded on the factors. The ACCESS factor appears to represent the ability of the golfer to play the course. The COURSE factor appears to represent the physical characteristics of the golf course itself. The PEOPLE factor appears to represent other individuals who directly or indirectly influence the golfer. Finally, the EXTRAS factor appears to represent the "associated facilities" of the golf course.
Go to Table One
The alpha values associated with each factor that appear in Table One all exceeded 0.7000. These findings lend support that the attributes comprising each factor share a common core [Churchill 1979]. The item-to-total correlation did not suggest further deletion. While it was hoped that the alpha values might be higher, the attributes did exhibit acceptable levels of reliability for an exploratory study [Nunnally 1978].
Regression analysis was employed to investigate whether a group of explanatory variables exerts a significant influence on public golf course choice intentions. The 17 attributes serve as explanatory variables in the regression model, and they are hypothesized to exert a positive impact on choice intentions.
With the use of such a large number of possibly highly correlated explanatory variables, the possibility of multicollinearity exists. Multicollinearity can produce misleading results as to the importance of the individual explanatory variables. An initial examination of the Condition Indexes suggest that multicollinearity should be further investigated (the largest and only index in excess of 30 being 30.8331). However, the proportion of variance of two or more parameters associated with each high Condition Index (i.e., Variance Decomposition Proportion) failed to exceed the 0.5000 level [Belsley, Kuh, and Welsch 1980]. This suggests that multicollinearity is not severe enough to confound the results [Belsley, Kuh, and Welsch 1980].
The parameters of the regression model were estimated using the data of 241 respondents. Parameters were estimated by ordinary-least-squares. Table Two presents the parameter estimates and goodness-of-fit statistics for the regression model. The model achieved a significant overall level of goodness-of-fit as measured by the F-test. In other words, at least one of the attributes is important for explaining choice intentions. In addition, the R2 indicates that approximately 71 percent of the variance in choice intentions can be explained by the attributes.
Go to Table Two
The true test of a model's diagnostic usefulness is evidenced by its predictive accuracy. Therefore, a second set of 54 respondents (from a second study) served as a holdout sample to assess the predictive accuracy of the model. The parameters of the regression model estimated from the estimation data set were used to predict choice in the holdout data set. Predicted choice is correlated with actual choice in the holdout data set to obtain a measure called a "cross validity correlation coefficient" [Green and Srinivasan 1978]. The cross validity correlation coefficient is 0.6901. As such, the regression model predicts choice intentions reasonably well in the holdout data set.
The t-test provided by the regression output was employed to assess the statistical significance of each of the attributes hypothesized to influence public golf course choice intentions. The regression model has 10 significant attributes, not including the intercept. In addition, all of the attributes agree with their a priori signs.
All four sets of attributes (i.e., factors) have at least one significant attribute. It appears that ACCESS, COURSE, PEOPLE, and EXTRAS attributes are important for explaining choice intentions.
All four of the ACCESS attributes are significant. These results suggest that speed of play (SPEED), reasonable price (PRICE), convenient tee times (TEETIME), and convenient location (LOCATE) are important explanatory variables of golf course choice intentions.
Two of the COURSE attributes are significant. As such, course layout (LAYOUT) and condition of fairways (FWAYS) appear to be important explanatory variables of golf course choice intentions.
Three of the PEOPLE attributes are significant. This suggests that other golfers' friendliness (OTHERS), the course designer (DESIGNER), and the golf pro (PRO) are important explanatory variables of golf course choice intentions.
One of the EXTRAS attributes is significant. This suggests that the practice facilities (PRACTICE) is an important explanatory variable of golf course choice intentions.
The rank order from largest parameter=1 to smallest parameter=17 also appears in Table Two. No one factor has a preponderance of large or small ranking attributes (the only exception might be for the factor EXTRAS). The largest parameters are for PRICE, TEETIME, and PRO. The smallest parameters are for AMENITY, GREENS, and FOOD.
This study has some limitations. First, the analysis was performed on a diverse group of public golf courses. The evaluation of the attributes influencing public golf course choice intentions may vary by "type" of public golf course (e.g., high priced courses versus low priced municipal courses).
In addition, this study assumes a homogeneous group of respondents. No attempt was made to examine possible differences between various demographic segments of the market in their evaluation of the attributes thought to influence public golf choice intentions.
Several significant conclusions can be drawn from this research. It appears that golfers utilize multiple factors and attributes when choosing a public golf course. In addition, there appears to be support for concluding that four factors influence public golf course choice intentions.
Consistent with what is hypothesized about consumers of public golf courses, the variables PRICE and TEETIME had the largest parameter estimates. The importance of these ACCESS factor attributes suggest that public golf courses are more likely to be utilized by those relatively more interested in playing golf than consuming other services offered by the course. Since the sample consisted mostly of traveling business people, it can be conjectured that this group has less time to play than many other golfers (e.g., retired people) and that this group is very interested in value and convenience.
It should be pointed out that PRICE has a positive sign, which is expected since the question was in reference to the "reasonableness" of price. Thus, the more reasonable the price the more likely the golfer is to return to play the course.
The importance of the accessibility factor (i.e., ACCESS) is further supported by the high rankings of the SPEED and LOCATE attributes, which may indicate the time pressures of the specific sample utilized. Golfers appear to choose public courses which are compatible with the golfer's desired speed of play and the location of the course relative to the golfer's home. Again, this may be indicative of those who utilize public courses principally to play a round of golf, rather than to consume other services offered by the course.
Unexpectedly, LONG, GREENS, and HARD were not significant. It was hypothesized that these attributes should influence how one plays and, thus, intentions to return. However, the other two COURSE factor attributes, LAYOUT and FWAYS, were significant. These results may suggest that the sample was not very concerned with hazards (which are not as common on public courses in comparison to private courses) or other difficulty of play attributes like length, but was concerned with whether the ball could be kept in play on quality fairways with a reasonable layout.
As for the PEOPLE factor, all attributes except for RESPECT were significant. In an environment where golf is the principle activity and justification for consumption, it is not surprising that the attitude of other golfers would be important (i.e., OTHERS). Since oftentimes in public courses the other golfers are not social friends and are less likely to be bound by friendship courtesies, a friendly atmosphere may assume greater importance.
The course designer attribute (DESIGNER) was also significant. Who the course was designed by may provide the golfer with valuable information prior to playing a course for the first time. Many course designers have developed reputations for including certain attributes in the design of a course. For example, Pete Dye is known to design very difficult courses containing many hazards. Knowledge of the course designer might allow the golfer to assess certain attributes of the course prior to actual play.
Note should also be taken of the highly ranked PRO attribute. Since many who use public golf courses may be fairly new to the game and perhaps require more instruction, it is expected that the PRO attribute would be very important. The assistance of an expert is critical to those new to any activity such as golf.
Normally, the advice of someone who is respected is important in making purchasing decisions. Nonetheless, the failure of the RESPECT attribute to achieve significance might not be so unexpected if one considers the ease of trial-and-error in selecting a public course on which to play and the reasonably small risk involved. This reduces the need for a positive recommendation.
The failure of all of the EXTRAS factor attributes except for PRACTICE to be significant appears to be consistent with the other findings of the study. While marketers would be ill advised to ignore the pro shop (SHOP), restaurant (FOOD), and other amenities (AMENITY) as influencing choice intentions of a public golf course, the results suggest that many other attributes are more important.
As previously indicated, the golfers in the sample appeared to be more concerned with playing a round of golf and less concerned with other social/recreational activities. This conclusion is supported by the fact that only EXTRAS factor attribute which was significant was PRACTICE. (The practice tee/green is a crucial amenity for golfers concerned with improving their level of play and is important for someone with insufficient time to play a round of golf.)
It appears that there are a number of important attributes that influence the choice intentions for public golf courses. This knowledge should assist marketers to segment markets and to develop positioning strategies relative to competition.
The reliability and validity of the 17 attributes thought to influence public golf course choice intentions should encourage academicians and golf course managers to pursue further research on this topic. To the extent that those engaged in marketing golf courses find the 17 attribute questionnaire diagnostically useful, we offer four guidelines which should assist in the development and adjustment of their marketing strategies:
(1) Course-Type Analysis: The analysis described in this article was performed on public golf courses in general. Therefore, its results should not be generalized to apply to other types of golf courses. Specifically, differences in attribute import ance may exist between public and private courses, high-priced public courses and low priced courses, etc. However, the same questionnaire and methodology can be easily applied to other types of courses. As a result, marketers can gain insights from this study as to the most important attributes that influence choice intentions.
(2) Segment-Specific Analysis: The 17-attribute questionnaire can be administered to several segments of golfers based on differences in demographics, psychographics, etc. Segment-specific analysis allows you to investigate the differential influence of the attributes across segments of golfers.
(3) Competitive Analysis: By asking golfers to complete the 17-attribute questionnaire on several competiting public golf courses, the results can be used to evaluate and compare strategies relative to several competiting golf courses.
(4) Temporal Analysis: The 17-attribute questionnaire can be administered periodically to track change (or lack of change) in golfer evaluations over time. As such, you can evaluate the effectiveness of strategies over time.
Clearly, public golf course users represent an identifiable segment of the golf market who consider certain attributes when choosing a course. The findings of this study suggest that those marketing public golf course should be concerned with several significant attributes in their marketing effort, but should primarily concentrate on providing the maximum convenience, value, and a challenging quality course if they are to meet the needs of the market.