A Discrete Choice Experimental Approach to Understand Sports Event Tourists’ In-Stadium Beer Consumption Preferences

Abstract

While drinking beer is an important component of sports event tourists’ ceremonial behaviors, there is surprisingly limited information regarding what features of beer service offerings at professional sporting events are most preferred. Using a sample of sports event tourists attending South Korean professional baseball games, the main purpose of this study is to provide an opportunity for improved knowledge regarding how spectators show their willingness to pay for in-stadium beer services. Applying a discrete choice experiment, we also intend to identify spectators’ heterogeneous preference systems, which vary depending on their levels of everyday alcohol use. Results suggest that respondents place great importance on serving temperatures and waiting time in line for purchasing beer at concession stands. The two drinker groups segmented based on the Alcohol Use Disorders Identification Test scores show different tastes for beer services. Several management implications are discussed to improve in-stadium entertainment and prevent alcohol-fueled misbehaviors.

Keywords

beer discrete choice experiment alcohol use disorders identification test professional sporting event sports event tourist willingness to pay

Introduction

Since sports event tourists often view alcohol, particularly beer, as a symbol of celebration, alcoholic beverages are available at stadium concession booths although some countries still prohibit in-stadium alcohol consumption due to their religious and cultural backgrounds (Lyne & Galloway, 2011). From the managerial perspectives, selling alcohol is an important revenue source for sports clubs (Mitchell & Montgomery, 2015). The Ohio State University was known to have generated US$1.35 million in total net revenue from beer sales at the Ohio Stadium in 2016 (Smola, 2017). There is no detailed figure regarding how much money major professional sports teams in the United States earned by selling alcoholic beverages due to their secret contracts with liquor suppliers (Bolluyt, 2018).

While beer consumption has become an important part of in-stadium services and entertainment packages (Arul, 2015), alcohol abuse is closely associated with a variety of public health problems, including serious psychiatric and behavioral disorders (Room et al., 2005). As a result, a sizable body of previous literature has focused on the significant interconnectedness between in-stadium alcohol consumption and subsequent problematic fan behaviors. For example, Rees and Schnepel (2009) provided empirical evidence supporting that binge drinking at sporting venues results in diverse alcohol-induced incidents including vandalism, assault, and disorderly conduct either during or after events. Despite the public concerns regarding those destructive behaviors, consuming alcoholic beverages in stadium settings has gradually become a universal phenomenon (Lenk et al., 2009). Analogous to the Major League Baseball stadiums where beer is a basic item of concession facilities, the unique drinking culture of “Chimaek”—a blended term for fried chicken and “Maekju,” the South Korean word for beer—can be witnessed effortlessly at every corner of professional baseball parks in the country (Lim, 2016).

Although the intake of alcoholic beverages is believed to be a ritual of sports consumption, earlier studies (e.g., Lyne & Galloway, 2011; Menaker & Connaughton, 2010) have paid much attention to developing different strategies for restricting sports event tourists’ alcohol consumption. Those studies have also made much effort to suggest useful policies for preventing spectators’ alcohol-related problematic behaviors during or after sporting events. The regulatory milieu for curbing alcohol use in stadium settings may result in an incomplete awareness of event attendees’ attitudes toward alcohol consumption. In particular, limited research has been conducted to examine sports event tourists’ preferences and willingness to pay for in-stadium alcohol service environment and their inherent preferences for alcohol control policies.

Within the context of the Korea Baseball Organization (KBO) league, we aim to better recognize what features representing beer service offerings at professional baseball parks are considered important by South Korean sports event tourists in order to reach their highest levels of attendance satisfaction. Using a discrete choice experimental (DCE) approach, which has been popularly applied for evaluating the inherent economic values of particular products and services (Hensher et al., 2015), we also attempt to assess sports event tourists’ willingness to pay for in-stadium beer services, which is known to vary depending on their levels of everyday alcohol use. Based on study findings, we will provide several management strategies to help sporting event administrators fulfill their customers’ game day entertainment needs and minimize problematic alcohol-related behaviors.

Literature Review

In-Stadium Beer Consumption

Beer is the most popular and cheapest alcoholic beverage consumed while attending diverse sporting venues (Arul, 2015). The drinking pattern is associated with the distinctive peer effects of beer, normally used with a group of friends and family members (Arnold, 2005). Read et al. (2003) argued that beside the social reinforcement motives for in-stadium alcohol consumption, drinking behaviors can be driven to stimulate users’ positive emotions or ameliorate their negative feelings. While consuming beer at stadiums is a highly controversial topic, most sports organizations have officially allowed in-stadium alcohol consumption. This is because sports event tourists who wanted to drink beer had to stay at home or go to a pub in order to watch televised games (Dodd, 2016).

Despite the omnipresence of beer at stadium settings, there is limited research effort to better understand how sports event tourists form their preferences for in-stadium beer service offerings. Alternatively, previous studies (e.g., Cebula, 2013; Chastain et al., 2017) have paid much attention to examine whether sporting event attendance is boosted by the availability of alcoholic beverages at concession booths. Using a panel data analysis, for example, Cebula (2013) found that a variety of promotional efforts including discounted beer prices served as a significant factor in encouraging attendance entertainment and attracting more spectators to baseball parks. Paul et al. (2009) also provided empirical evidence that the number of sports event tourists attending South Atlantic League baseball games was positively influenced by several in-stadium entertainment components such as free beer and food coupons. On the contrary, Chupp et al. (2007) failed to identify statistically significant correlations between alcohol consumption and sporting event attendance. In the context of minor league baseball, Paul et al. (2007) revealed that beer-related promotional policies characterized by free beers and reduced beer prices were not directly related to frequent game attendance.

A sizable body of previous work has focused on behavioral interplays between fan intoxication and subsequent disorderly conduct rather than several optimistic effects of beer drinking for game day entertainment. Nelson and Wechsler (2003) found different negative aspects of in-stadium alcohol consumption leading to a high likelihood of binge drinking at stadiums and problematic behaviors among collegiate sports fans. As a result, sporting event administrators have implemented alcohol control policies to prevent alcohol-related problems. For example, the National Basketball Association has limited beer serving size since the infamous Pacers-Pistons brawl happened in 2004, which was sparked by a Pistons fan’s throwing beer on Ron Artest (Filce et al., 2016). Based on the recommendations from the Techniques for Effective Alcohol Management, a nonprofit organization for responsible drinking at sporting facilities, most Major League Baseball concession stands also stop selling beer near the end of the seventh inning to curb alcohol-fueled fan misbehaviors (Fitzpatrick, 2015).

Beer Choice Decisions

Decisions for beer choice are closely associated with consumers’ distinctive tastes for particular types of beer and their attitudes toward the alcoholic beverage, which vary according to regional and cultural backgrounds (McCluskey & Shreay, 2011). Within the context of retail stores and restaurants, a number of previous studies in food and nutrition science areas have paid great attention to identifying different characteristics that affect beer choice decisions of ordinary consumers. Using several choice criteria for choosing specialty beers including malt types, adjuncts, ingredients, and retail price, Donadini et al. (2016) uncovered a heterogeneous preference system of beer consumers embedded across three European countries. While comparing a set of craft beer attributes such as aroma, color, calories, raw material, alcohol content, and perceived quality, Aquilani et al. (2015) also revealed that retail consumers put relatively heavier weight on the two features of aroma and perceived quality when purchasing a microbrew product. Gabrielyan et al. (2014) provided empirical evidence where diverse sensory attributes (e.g., tastes, aroma, temperature) play a critical role in the beer choice decision-making process.

General beer consumers tend to place great importance on brewery brand images (Gómez-Corona et al., 2016). Speece et al. (1994) noted that typical imported beer consumers in Hong Kong were less sensitive to prices but more responsive to other attributes like brand images and product quality. Adopting a conjoint rating experiment, Donadini and Porretta (2017) also found that Italian beer consumers consider brewery types and brewing technology as important criteria in determining the quality perception of craft beers. While only a handful of studies have been implemented to recognize what factors contribute to beer choice decisions at stadiums, sports event tourists attending professional sporting events are likely to show an identical preference pattern with general beer consumers. Muret (2014) indicated that spectators were more in favor of specialty brews with a variety of flavors than mass-produced macro beers while they were most unsatisfied with a long wait in line to purchase the alcoholic beverage at concession stores.

Beer preferences of general consumers vary depending upon their distinctive levels of everyday alcohol use (Rogers & Greenfield, 1999). An international comparison of alcohol consumption conducted by Bloomfield et al. (2003) revealed that the formation of alcohol preferences was significantly affected by a variety of antecedent variables including drinking frequency and serving size. Along the same lines, Courtney and Polich (2009) disclosed that heavy drinkers pursue beverages with higher alcohol content and a longer time of alcohol consumption. Sports fans are engaged in binge drinking more frequently than general alcohol consumers (Nelson & Wechsler, 2003), so there may be a close association between sports event tourists’ preferences for beer consumption and their ordinary alcohol use levels.

Discrete Choice Experiment

The DCE is acknowledged to be a useful microeconomic tool for assessing consumers’ willingness to pay for proposed products and services with an application of the stated preference theory that derives individuals’ utility from their responses to hypothetical market situations (Louviere et al., 2000). The key theoretical background of the DCE approach involves the random utility theory, assuming that ordinary consumers purchase a product or service offering while maximizing their perceived utility (Manski, 1977). The random utility theory also implies that individuals’ utility is composed of two different aspects: observable deterministic and unobservable stochastic terms. Given that sports event tourist i choose in-stadium beer service offering j from J alternatives, the indirect utility function can be presented as follows:

U_{i j} = V_{i j} + ε_{i j},

where V_ij denotes the observable part of utility to be estimated and ε_ij indicates the unobservable error component.

With an assumption that the stochastic error element is independently and identically Gumbel-distributed, the predicted probability of sports event tourist i purchasing beer service offering j can be expressed as follows:

P_{i j} = \exp (V_{i j}) / \sum_{j = J} e x p (V_{i j}) .

According to McFadden (1974), this probability can be estimated by employing the multinomial logit model. Despite the concise estimation algorithms, the multinomial logit model violates the indispensable DCE assumption of the independence of irrelevant alternatives (IIA) and fails to incorporate respondents’ heterogeneous preference systems into the indirect utility function (Lyu & Lee, 2018). Alternatively, the random parameter logit (RPL) model has been applied to avoid the stringent IIA assumption and embrace consumers’ preference heterogeneity in the utility function (McFadden & Train, 2000). To estimate accurate parameters in the RPL model as well as derive useful information regarding respondents’ willingness to pay, researchers specify the distribution of nonprice attributes as normal, holding the price-related attribute constant (Hensher et al., 2015).

Method

Attributes and Hypotheses

We extensively reviewed previous literature focusing on alcohol consumption in professional and collegiate sporting events (e.g., Chupp et al., 2007; Paul et al., 2007) and preference orientations of beer consumers (e.g., Aquilani et al., 2015; Donadini et al., 2016) to generate an initial pool of prospective attributes encompassing in-stadium beer service offerings. To incorporate diverse stadium alcohol control policies into the attribute pool, we explored internet homepages and news articles featuring alcohol consumption at sporting events. For screening the content validity of the original attributes, a consultation meeting with five KBO league club marketers was held in September 2017. As a result, we chose eight attributes comprising the hypothetical professional ballpark beer service offerings (i.e., BRAND, TEMP, WAIT, CLOSE, TASTE, ALCOH, LIMIT, and PRICE) and then allocated two or three levels to each attribute. Supplemental Table S1 (available online) shows detailed information on attributes and levels.

Explanations regarding the hypothesized coefficient signs of each attribute are as follows:

BRAND represents the number of beer brands available at concession stands. This attribute involves three levels (i.e., “1 brand,” “3 brands,” and “5 brands”). We hypothesized a positive coefficient sign because most beer-consuming spectators are likely to prefer more beer brands (Donadini et al., 2016).

Hypothesis 1₀: Diverse beer brands would not positively affect spectators’ beer preferences.

Hypothesis 1_a: Diverse beer brands would positively affect spectators’ beer preferences.

TEMP indicates the beer temperatures served by concession sellers. This attribute has three different levels (i.e., “4 °C,” “8 °C,” and “10 °C”). We hypothesized a negative sign since previous studies (Arnold, 2005) supported that general beer users show stronger tastes for a cool beer temperature.

Hypothesis 2₀: High beer temperature would not negatively affect spectators’ beer preferences.

Hypothesis 2_a: High beer temperature would negatively affect spectators’ beer preferences.

WAIT denotes the waiting time for ordering and purchasing beer at concession booths. This attribute is comprised of three levels (i.e., “1 minute,” “3 minutes,” and “5 minutes”). Based on Muret’s (2014) findings, we hypothesized a negative sign because no one wants to spend a longer time buying beer at concession stores.

Hypothesis 3₀: Long waiting time would not negatively affect spectators’ beer preferences.

Hypothesis 3_a: Long waiting time would negatively affect spectators’ beer preferences.

CLOSE indicates the closing time of beer booths. This attribute consists of two levels (i.e., “End of the 7th inning,” and “End of the 8th inning”). We hypothesized different signs according to respondents’ degrees of everyday alcohol consumption. More specifically, we anticipated that heavy drinkers were likely to reveal their desires for a longer time of beer intake, but light drinkers were likely to show their distastes for binge drinking (Courtney & Polich, 2009).

Hypothesis 4-1₀: Delayed closing time of beer booths would not negatively affect harmless drinking spectators’ beer preferences.

Hypothesis 4-1_a: Delayed closing time of beer booths would negatively affect harmless drinking spectators’ beer preferences.

Hypothesis 4-2₀: Delayed closing time of beer booths would not positively affect hazardous drinking spectators’ beer preferences.

Hypothesis 4-2_a: Delayed closing time of beer booths would positively affect hazardous drinking spectators’ beer preferences.

TASTE represents the number of beer flavors. This attribute involves three levels (i.e., “1 type,” “3 types,” and “5 types”). We hypothesized a positive coefficient sign because previous studies (e.g., Donadini & Porretta, 2017; Speece et al., 1994) provided supportive evidence that most beer consumers want to enjoy more diversified beer tastes.

Hypothesis 5₀: Diverse beer tastes would not positively affect spectators’ beer preferences.

Hypothesis 5_a: Diverse beer tastes would positively affect spectators’ beer preferences.

ALCOH denotes the highest alcohol content of beer served. This attribute has three levels (i.e., “4%,” “6%,” and “8%”). We hypothesized different coefficient signs according to respondents’ usual alcohol use behaviors. In other words, more frequent drinkers were expected to consume beers with higher alcohol content (Aquilani et al., 2015).

Hypothesis 6-1₀: High alcohol content would not negatively affect harmless drinking spectators’ beer preferences.

Hypothesis 6-1_a: High alcohol content would negatively affect harmless drinking spectators’ beer preferences.

Hypothesis 6-2₀: High alcohol content would not positively affect hazardous drinking spectators’ beer preferences.

Hypothesis 6-2_a: High alcohol content would positively affect hazardous drinking spectators’ beer preferences.

LIMIT indicates the maximum number of beer cups per purchase. This attribute comprises three levels (i.e., “2 cups,” “3 cups,” and “4 cups”). We hypothesized different signs depending upon sports event tourists’ levels of everyday alcohol use. Specifically, we anticipated that heavy beer drinkers were likely to prefer more cups of beer per purchase (Courtney & Polich, 2009).

Hypothesis 7-1₀: More number of beer cups per purchase would not negatively affect harmless drinking spectators’ beer preferences.

Hypothesis 7-1_a: More number of beer cups per purchase would negatively affect harmless drinking spectators’ beer preferences.

Hypothesis 7-2₀: More number of beer cups per purchase would not positively affect hazardous drinking spectators’ beer preferences.

Hypothesis 7-2_a: More number of beer cups per purchase would positively affect hazardous drinking spectators’ beer preferences.

PRICE denotes beer prices for a regular-size cup at concession stores. This attribute consists of three levels (i.e., “Korean Republic won [KR₩]4,000,” “KR₩5,000,” and “KR₩6,000”). We hypothesized a negative coefficient sign because of the common distastes for higher product prices.

Hypothesis 8₀: High beer prices would not negatively affect spectators’ beer preferences.

Hypothesis 8_a: High beer prices would negatively affect spectators’ beer preferences.

Experiment Design

A full factorial design including all possible combinations of attributes and levels results in an uncontrollable number of choice sets (i.e., 3⁷ × 2¹ = 4374). Alternatively, we utilized a fractional factorial design with main effects to derive a minimum set of choice comparisons. Following the guidance of Kuhfeld (2005), a D-optimal design was implemented to accomplish the maximized (i.e., 100%) efficiency of parameter estimates. This experimental design process yielded a total of 18 choice comparisons. We also utilized a blocking strategy to improve response rates; therefore, our choice comparisons were equally distributed into six different questionnaire versions and each respondent was asked to answer only three choice sets. Supplemental Figure 1 (available online) demonstrates an example of choice comparisons presented to respondents.

To measure respondents’ levels of everyday alcohol use, we utilized the Alcohol Use Disorders Identification Test (AUDIT) questionnaire developed by the World Health Organization. The 10-item measurement tool is applied globally to identify hazardous and harmful alcohol consumption (Reinert & Allen, 2007). Each item of the AUDIT is scored on a scale from 0 to 4 to assess respondents’ average quantity and frequency of drinking, presence or absence of binge drinking, dependence symptoms, and alcohol-induced problems (Saunders et al., 1993). Despite the diversified influences of alcohol intake depending on body weight and metabolism, a total score of 8 or higher is commonly considered an indication of hazardous and harmful alcohol use, as well as possible alcohol dependence. Supplemental Table 2 (available online) provides descriptive information regarding the AUDIT used for this study.

Data Collection

Collaborating with a South Korean market research company, we conducted an online survey to collect data for a week in October 2017. Using a proportionate random sampling procedure, we selected slightly fewer than 10,000 respondents from approximately 1 million panel members managed by the research company. A web address for retrieving our questionnaire homepages was delivered to each respondent with a screening question asking whether they had consumed beer at a KBO league ballpark during the previous 12 months. As a result, 438 respondents completed the survey. After excluding 24 respondents with incomplete answers, we analyzed the DCE data derived from 414 participants.

Results and Discussion

Participant Profiles

Males held a majority (n = 223, 53.9%) and the average age of respondents was 34.4 years (SD = 9.3). The greatest portion was in their 30s (n = 163, 39.4%), followed by 20s (n = 125, 30.2%) and 40s (n = 92, 22.2%). Almost seven out of 10 participants reported holding a bachelor’s degree or higher (n = 289, 69.8%). The largest proportion indicated that their average annual household income ranges from KR₩40 million to KR₩50 million (n = 78, 18.8%). Respondents also reported attending 4.7 KBO league games on average (SD = 4.3) during the past 12 months.

We further segmented respondents into two beer user groups based on the AUDIT scores. As a result, 162 participants (39.1%) were classified as harmless drinkers with a total score of less than 8, while 252 respondents were categorized as hazardous drinkers. A series of difference tests were conducted to detect group differences in several sociodemographic and attendance characteristics. The results of difference tests indicated that harmless drinkers were less likely to be males (χ² = 23.710, p < .001) and attended fewer KBO league games (t = 2.264, p = .024).

RPL Model Estimations for Aggregate Sample

Using the NLOGIT, a software package for econometrics, we estimated several RPL models, which are free from the strict IIA assumption of the DCE algorithms. Following the recommendation of Train (2003), the Halton draws, a popular simulated maximum likelihood estimation method, was employed to derive robust parameter estimates. To identify the effects of sociodemographic features on respondents’ choice decisions, we also estimated the RPL model with several interaction terms. Table 1 shows the estimation results of the RPL models with and without interaction terms.

Table 1

Results of RPL Model Estimations Using Aggregate Sample

Attributes	RPL Without Interactions		RPL With Interactions
Attributes	M Coefficient (SE)	SD Coefficient (SE)	M Coefficient (SE)	SD Coefficient (SE)
ASC	9.244*** (0.882)	3.219*** (0.504)	9.510*** (1.628)	3.257*** (0.548)
BRAND	0.244*** (0.036)	0.028 (0.157)	0.245*** (0.038)	0.110 (0.144)
TEMP	−0.146*** (0.023)	0.240*** (0.038)	−0.152*** (0.025)	0.235*** (0.039)
WAIT	−0.287*** (0.044)	−0.244*** (0.077)	−0.291*** (0.047)	0.270*** (0.084)
CLOSE	0.164*** (0.051)	0.018 (0.110)	0.164*** (0.052)	0.025 (0.108)
TASTE	0.183*** (0.032)	0.023 (0.111)	0.186*** (0.034)	0.072 (0.081)
ALCOH	−0.080*** (0.030)	0.041 (0.068)	−0.080** (0.030)	0.025 (0.140)
LIMIT	0.392*** (0.075)	−0.087 (0.178)	0.401*** (0.078)	0.192 (0.204)
PRICE	−0.957*** (0.097)		−0.973*** (0.105)
ASC × Gender			−0.586 (0.699)
ASC × Age			−0.025 (0.031)
ASC × Income			0.201* (0.122)
Log likelihood	−955.798		−954.538
N	414		414

Note: RPL = random parameter logit; ASC = alternative specific constant; BRAND = number of beer brands; TEMP = beer temperatures; WAIT = waiting time for ordering and purchasing beer; CLOSE = closing time of beer booths; TASTE = number of beer flavors; ALCOH = highest alcohol content of beer served; LIMIT = maximum number of beer cups per purchase; PRICE = beer prices for a regular-size cup.

p < .1. **p < .05. ***p < .01.

The results of the RPL model estimation without interaction terms indicate that all mean coefficients for eight ballpark beer service attributes were statistically significant. More specifically, the alternative specific constant (ASC) revealed a significant positive mean coefficient, suggesting that our respondents showed their inclinations to choose an alternative from the two ballpark beer service options rather than the forgoing alternative. The negative mean coefficient of the PRICE attribute indicated that the null hypothesis (Hypothesis 8₀) was rejected. This result provided evidence supporting the contention that higher beer prices were less favored by sports event tourists, which is consistent with the well-known demand theory.

The BRAND and TASTE attributes showed positive mean coefficients suggesting that the null hypotheses (Hypothesis 1₀; Hypothesis 5₀) were rejected. These results indicate respondents’ preferences for more varied breweries and beer flavors encountered at ballpark concession booths. The negative coefficient signs of the two attributes TEMP and WAIT revealed that respondents disliked tepid beer and a longer waiting time for ordering beer, which meant the rejections of the null hypotheses (Hypothesis 2₀; Hypothesis 3₀). We hypothesized different coefficient signs for the three attributes of the CLOSE, ALCOH, and LIMIT, which vary according to sports event tourists’ levels of usual alcohol usage. While we did not develop the null and alternative hypotheses for the three attributes, each mean coefficient sign may provide useful implications for better understanding respondents’ comprehensive preference systems. The CLOSE attribute debunked a positive mean coefficient, indicating that the extended beer sale deadline from the end of the seventh inning to the end of the eighth inning was preferred by KBO league spectators. The negative mean coefficient of the ALCOH attribute demonstrated all respondents’ aversion to beer with higher alcohol content, while the positive coefficient of the LIMIT attribute denoted their preferences for a more flexible limit on beer purchase.

The statistical significance pattern of the standard deviation coefficients was different from that of the mean coefficients. The standard deviation coefficients suggest important information regarding whether respondents show heterogeneous preferences for particular attributes. The estimation results of the RPL model without interaction terms indicated that our respondents held heterogeneous preferences for the two ballpark beer service attributes (i.e., TEMP and WAIT) as well as the ASC term.

The mean coefficients in the RPL model with interaction terms revealed an identical significance pattern with the RPL model without interactions. Among three different interaction terms used to better understand the cross effects between sociodemographic features and the ASC, only the interaction term of ASC × INCOME was statistically significant. The positive coefficient sign of the interaction term suggests that our respondents were more likely to purchase beer at concession stands as their household income levels increased. In order to compare the fit of the two RPL models, we implemented a likelihood ratio (LR) test, as recommended by Louviere et al. (2000). The LR test statistic of 2.520 (i.e., −2[−955.798 − −954.538]) failed to reject the null hypothesis of identical preferences between the two models when taking the critical value (i.e., $χ_{3, 0.05}^{2}$ = 7.815) into consideration. As a result, we chose the RPL model without interaction terms as the final model for our segmentation strategy because of its parsimony.

RPL Model Estimations for Segmented Samples

The estimation results of the RPL models using the segmented sports event tourist groups based on the AUDIT scores are shown in Table 2. Prior to interpreting different parameter estimates, this study performed an LR test to examine taste variations between aggregate sample and segmented groups. Ben-Akiva and Lerman (1985) noted that the LR test statistic can be derived using −2(LL_a − ∑LL_g), where LL_a denotes the log likelihood of the aggregate model and LL_g represents the log likelihood of the segmented models. The test statistic of 18.178 (i.e., −2[−955.798 − (−375.145 + −571.564)]) rejected the null hypothesis of identical coefficients across the two models when taking the critical value (i.e., $χ_{9, 0.05}^{2}$ = 16.919) into account. Stated differently, there is a significant difference in ballpark beer service preferences between aggregate sample and the two drinker groups.

Table 2

Results of RPL Model Estimations Using Segmented Samples

Attributes	Harmless Drinkers (AUDIT < 8)		Hazardous Drinkers (AUDIT ≥ 8)
Attributes	M Coefficient (SE)	SD Coefficient (SE)	M Coefficient (SE)	SD Coefficient (SE)
ASC	10.240*** (1.848)	−2.692** (1.308)	8.958*** (1.233)	3.296*** (0.802)
BRAND	0.298*** (0.068)	0.104 (0.158)	0.222*** (0.045)	0.086 (0.209)
TEMP	−0.218*** (0.050)	0.285*** (0.064)	−0.113*** (0.028)	0.203*** (0.044)
WAIT	−0.388*** (0.087)	0.386*** (0.122)	−0.234*** (0.051)	0.068 (0.191)
CLOSE	0.184** (0.090)	−0.034 (0.224)	0.163** (0.064)	0.007 (0.125)
TASTE	0.173*** (0.061)	−0.165 (0.131)	0.192*** (0.040)	0.016 (0.102)
ALCOH	−0.086 (0.054)	−0.047 (0.118)	−0.074** (0.037)	0.005 (0.068)
LIMIT	0.391*** (0.140)	0.335 (0.243)	0.412*** (0.093)	0.272* (0.161)
PRICE	−1.079*** (0.183)		−0.916*** (0.123)
Log likelihood	−375.145		−571.564
N	162		252

Note: RPL = random parameter logit; AUDIT = Alcohol Use Disorders Identification Test; ASC = alternative specific constant; BRAND = number of beer brands; TEMP = beer temperatures; WAIT = waiting time for ordering and purchasing beer; CLOSE = closing time of beer booths; TASTE = number of beer flavors; ALCOH = highest alcohol content of beer served; LIMIT = maximum number of beer cups per purchase; PRICE = beer prices for a regular-size cup.

p < .1. **p < .05. ***p < .01.

Excepting for the ALCOH attribute in the harmless drinker model, all mean coefficients were statistically significant. More specifically, the CLOSE and LIMIT attributes in the hazardous drinker model revealed a significant positive mean coefficient sign demonstrating that the null hypotheses (Hypotheses 4-2₀ and 7-2₀) were rejected, but the ALCOH attribute indicated a significant negative sign, indicating that the null hypothesis (Hypothesis 6-2₀) was failed to reject. The CLOSE and LIMIT attributes in the harmless drinker model indicated a significant positive mean coefficient sign, which displayed the failure to reject the null hypotheses (Hypotheses 4-1₀ and 7-1₀). The standard deviation coefficients also showed a similar significance configuration with the aggregate model.

Relative Importance

We calculated the relative importance of each attribute to be better aware of which attributes representing ballpark beer service offerings are considered substantial by respondents. According to Hensher et al. (2015), the relative importance of a particular attribute k can be measured as follows:

R I_{k} = P a r t - w o r t h_{k} / \sum_{k = K} P a r t - w o r t h_{k},

while the part-worth utility can be computed by multiplying the mean coefficient of each attribute by the interval of its levels. Supplemental Figure 2 (available online) shows the calculation results of the relative importance of our eight attributes.

The results revealed that our respondents put the greatest importance on the PRICE attribute (26.0%) when purchasing beer at professional ballpark concession booths, followed by the TEMP (15.9%), WAIT (15.6%), and BRAND attributes (13.3%). The two spectator groups segmented based on the AUDIT scores showed a different pattern regarding the relative importance. Harmless drinkers placed heavier weight on the TEMP (19.8%) and WAIT attributes (17.6%), whereas hazardous drinkers put greater importance on the LIMIT (12.1%) and TASTE attributes (11.4%) compared to their counterpart.

Marginal Willingness-to-Pay Values

The marginal willingness-to-pay (MWTP) values deliver practical information regarding how much respondents are ready to pay for maintaining their utility when the level of a particular attribute is changed by a unit (Hensher et al., 2015). The MWTP values can be calculated by using −β_k /β_price, where β_k denotes the mean coefficients of several nonprice random attributes k; β_price indicates the fixed coefficient of the price-related attribute (Louviere et al., 2000). Table 3 illustrates several computation results of the MWTP values derived from the aggregate model as well as the two segmented models.

Table 3

Results of MWTP Value Computations (Unit: KR₩)

Attribute	Marginal Change	MWTP Value
Attribute	Marginal Change	Aggregate	Harmless Drinkers	Hazardous Drinkers
BRAND	1 brewery	255	276	243
TEMP	1 °C	−153	−202	−124
WAIT	1 minute	−300	−360	−255
CLOSE	End of 7th inning → End of 8th inning	172	171	178
TASTE	1 flavor type	191	161	210
ALCOH	1%	−84	−79	−81
LIMIT	1 cup (350 ml)	410	363	449

Note: MWTP = marginal willingness-to-pay; KR₩ = Korean Republic won; BRAND = number of beer brands; TEMP = beer temperatures; WAIT = waiting time for ordering and purchasing beer; CLOSE = closing time of beer booths; TASTE = number of beer flavors; ALCOH = highest alcohol content of beer served; LIMIT = maximum number of beer cups per purchase.

The results indicated that our respondents were willing to pay KR₩410 if the purchase limit was relaxed by one more cup of beer. They also revealed their intentions to pay KR₩191 given that ballpark concession stores added one new-flavored beer. The negative MWTP values suggest spectators’ distaste for undesirable changes. Accordingly, the monetary value of the WAIT attribute indicates that our sports event tourists were willing to pay KR₩300 when waiting time for purchasing beer could be lessened by 1 minute. The two segmented groups showed different MWTP values for each beer service attribute. Harmless drinkers were willing to pay KR₩105 more than hazardous drinkers if waiting time at concession stands could be decreased by 1 minute. On the other hand, hazardous drinkers were willing to pay KR₩7 more than harmless drinkers in order to extend the beer selling deadline.

Scenario Analyses

Using the predicted probabilities derived from the RPL model estimation results, we implemented a series of scenario analyses to better understand what proportion of sports event tourists are expected to attend hypothetical ballparks delivering distinctive beer services. We designed three different scenarios: current status, drinking-friendly, and most restrictive. Table 4 indicates the results of our scenario analyses.

Table 4

Results of Feasible Scenario Analyses

Attribute	Scenario 1 (Current Status)	Scenario 2 (Drinking-Friendly)	Scenario 3 (Most Restrictive)
BRAND	1 brewery	5 breweries	1 brewery
TEMP	6 °C	4 °C	10 °C
WAIT	3 minutes	1 minute	5 minutes
CLOSE	End of the 7th inning	End of the 8th inning	End of the 7th inning
TASTE	2 types	5 types	1 type
ALCOH	6%	4%	4%
LIMIT	4 cups	3 cups	2 cups
PRICE	KR₩4,000	KR₩6,000	KR₩5,000
Potential market shares
Aggregate	39.1%	58.1%	2.8%
Harmless	33.1%	65.1%	1.8%
Hazardous	42.8%	53.9%	3.3%

Note: KR₩ = Korean Republic won; BRAND = number of beer brands; TEMP = beer temperatures; WAIT = waiting time for ordering and purchasing beer; CLOSE = closing time of beer booths; TASTE = number of beer flavors; ALCOH = highest alcohol content of beer served; LIMIT = maximum number of beer cups per purchase; PRICE = beer prices for a regular-size cup.

The scenario analysis results revealed that our respondents most preferred the drinking-friendly service environment indicating a potential market share of 58.1%, while they least favored the most restrictive beer service atmosphere with an anticipated market share of 2.8%. The potential market shares also showed that the drinking-friendly stadium environment was more attractive to harmless drinkers than to hazardous drinkers. On the other hand, hazardous drinkers were more interested than harmless drinkers in the current in-park beer service offering. The most restrictive situation was least preferred by both drinker groups, but harmless drinkers showed slightly stronger distaste for the strict drinking environment.

Conclusion

Study Implications

Although consuming beer is an enjoyable game day experience for sports event tourists, little research effort has been exerted to better understand how spectators show their preferences when choosing a favorite beer service in stadiums. The main purpose of this study was to examine what attributes comprising in-stadium beer service offerings are considered important by professional sports event tourists. Using a DCE approach, we also contributed to an improved insight into how tourist groups segmented based on levels of the AUDIT scores indicate their heterogeneous preferences for each ballpark beer service attribute.

Several results from the relative importance analysis revealed that waiting time for purchasing beer was regarded by sports tourists attending KBO league games a substantial concern. A recent fan survey presented an analogous result indicating that the largest proportion of spectators visiting professional and collegiate sporting event venues in the United States (38%) requested eliminating long waits in line at concession amenities for an enjoyable in-stadium experience (Muret, 2014). Our respondents also indicated a relatively high willingness-to-pay value of KR₩300 when they could save 1 minute at beer concession counters. A management implication derived from this finding is that sporting event administrators need to develop appropriate numbers and locations of beer concession stands for enhancing attendance satisfaction. Furthermore, adopting diverse cutting-edge information and communication technologies including in-seat ordering and delivering mobile applications may be helpful in reducing sports event tourists’ efforts for purchasing the alcoholic beverage (Shin & Lyu, 2019).

It is notable that the BRAND attribute played a substantial role in shaping beer preferences at professional sporting events. The computation results of several MWTP values showed that respondents were willing to pay KR₩255 for an additional brewery brand encountered at beer concession booths. This finding may not be surprising because a number of previous studies (e.g., Aquilani et al., 2015; Speece et al., 1994) provided empirical evidence supporting the idea that beer consumers are sensitive to brewery brands when purchasing the alcoholic beverage at retail stores. Most KBO league ballparks have sold one or two macro brew brands due to the exclusive contracts with their domestic beer sponsors since the outset of stadium beer sales at the dawn of the 21st century. Nevertheless, the imported beer market in the country is rapidly growing and the value of total beer imports almost tripled from US$37.16 million in 2009 to US$116.9 million in 2014 (Sohn, 2016). This trend manifests domestic consumers’ soaring demand for higher quality products (Park, 2016). The preference shifts to imported beers may encourage sporting event administrators to consider adding more diversified beer brands into their concession booths.

Emerging demand for more varied beer flavors beyond the lager-dominated domestic brands was substantiated by the result of the relative importance analysis. A growing percentage of South Korean people have shown their strong preferences for craft beer despite its high prices (Sohn, 2016). Beer consumption in stadiums cannot be an exception to this preference changes. More professional sports stadiums in the United States are selling different types of craft beers because spectators are willing to pay more for those specialty brews than for several traditional macro beers (Muret, 2014). Apart from the production of new concession revenue, selling diverse specialty beers may promote professional sports teams to develop closer ties with local microbreweries.

It may be an interesting finding that respondents indicated their aversion to beer with higher alcohol content. The negative MWTP value on the ALCOH attribute revealed that sports event tourists attending KBO league games were willing to pay KR₩84 if the alcohol content would decrease by 1%. The vast majority of previous studies (e.g., Lyne & Galloway, 2011; Merlo et al., 2010) have raised a public concern that consuming beverages with a higher alcohol percentage increase the likelihood of engaging in destructive fan behaviors. Nevertheless, our respondents exhibited their desires for sober attendance experience while drinking beer with lower alcohol. This finding indicates that several policies for strictly limiting beverages with high alcohol content can be an effective way to prevent alcohol-fueled misbehaviors and improve attendance entertainment.

The segmentation approach based on distinctive levels of everyday alcohol consumption measured using the AUDIT provided useful opportunities for an enhanced awareness of sports event tourists’ embedded heterogeneity in in-stadium beer service preferences. According to the results of the relative importance analysis, harmless drinkers put heavier weight on several service-related aspects including the BRAND, WAIT, and TEMP attributes. More specifically, this light drinker group was less patient with long waiting in lines to purchase beer at concession stands and preferred more diversified brewery brands than did the heavy drinker group. Furthermore, harmless alcohol users were more sensitive than hazardous drinkers to beer serving temperatures, indicating their behavioral intentions to spend KR₩78 more as the temperature decreased by 1 °C. These findings seem to be attributed to the fact that harmless drinkers are less familiar with alcohol consumption at sporting event venues. This interpretation can be supported by the results of difference tests indicating significantly fewer game attendances among harmless drinkers. In other words, respondents affiliated with this light drinker group were more likely to pursue a hospitable service atmosphere as in pubs or restaurants when attending a KBO league game.

Unlike harmless beer users, hazardous drinkers were more interested in different fundamental features of beer including the TASTE attribute. This may not be surprising because previous studies (e.g., Holdstock et al., 2000) found heavy drinkers’ preferences for diverse alcohol flavors. Contrary to our hypotheses, the hazardous drinker group indicated their preferences for lower alcohol beer. Specifically, our respondents affiliated with the heavy drinker group were willing to pay KR₩81 as alcohol content decreased by 1%. Although sports event administrators have made various efforts to prevent alcohol-induced problems through reducing alcohol consumption among heavy users, this finding revealed that hazardous drinkers were vigilant about beer with higher alcohol content while attending professional sporting events. The results of these scenario analyses also provided supportive evidence that hazardous beer users were less interested than harmless drinkers in the drinking-friendly stadium settings (i.e., Scenario 2).

The computation results of several MWTP values indicated that hazardous drinkers were willing to pay KR₩7 more for the extended closing time of beer stands. The meager amount of monetary value suggests that heavy-drinking sports event tourists were not very dependent on alcohol consumption at sporting events. This finding is inconsistent with Nelson and Wechsler’s (2003) assertion that sports fans are more likely to cling to alcohol consumption and engage in binge drinking than general alcohol users. Nevertheless, hazardous drinkers exhibited a strong taste for a lenient limit on beer purchasing showing their willingness to pay KR₩86 more for an additional cup of beer compared to harmless users. In this sense, stricter enforcement of the alcohol purchase limit can be the most successful strategy for preventing alcohol-related problem behaviors.

Limitations and Future Studies

The DCE settings used for this study involved beer concession stands in professional baseball parks only. A considerable number of sports event tourists attending KBO league games purchase the alcoholic beverage at their seats using the “beer boy” systems—vendors carrying a barrel of beer and pouring it into cups upon orders—that contribute to the elimination of waiting in lines at concession booths (Lim, 2016). A variety of foods and snacks including fried chicken and pork barbecue nicely matched with beer are also available at professional ballparks in the country. Future studies using several attributes including different food and beverage options may provide deeper knowledge on sports event tourists’ preferences for in-stadium service offerings.

Levels of alcohol consumption are dependent on users’ perception of everyday stress, which serves as a substantial factor in predicting alcohol-related problematic behaviors (Brady & Sonne, 1999). Although we utilized a segmentation strategy based on the AUDIT scores, future research focusing on spectators’ degrees of daily stress may bring an alternative perspective on their beer service preferences in a stadium atmosphere. Incorporating the emotional state as a moderating variable can be advantageous in that escaping from usual routine is known to be a predominant motivator for attending professional sporting events. Another drawback is associated with the limited generalizability of this study since we employed the case of South Korean sports event tourists attending professional baseball fans. As McCluskey and Shreay (2011) recommended, more research applying a cross-cultural context may be useful in better recognizing how sports event tourists’ distinctive attitudes toward drinking play a central role in shaping their preferences for in-stadium alcohol consumption. Accordingly, a multiple case approach can provide valuable opportunities for better understanding of sports event tourists’ heterogeneous tastes for beer service offerings. Future studies are also needed to examine how alcohol-fueled problematic behaviors affect the images of sports event tourist destinations.

Concluding Summary

This study provided useful opportunities for an enhanced awareness of how sports event tourists make in-stadium beer choice decisions and which beer service elements are considered important to increase their levels of attendance entertainment. Using a DCE approach, we also proposed several management strategies for preventing spectators’ alcohol-fueled misbehaviors and improving their attendance satisfaction. In-stadium beer selling is regarded by sports event tourism administrators as an effective means for offering entertaining game day experiences to spectators and keeping their customers inside the venues longer. We hope that this study will assist sports event tourism professionals to implement different policies for developing enjoyable attendance environment and preventing destructive fan behaviors.

Supplemental Material

sj-pdf-1-jht-10.1177_1096348021992099 – Supplemental material for A Discrete Choice Experimental Approach to Understand Sports Event Tourists’ In-Stadium Beer Consumption Preferences

Supplemental material, sj-pdf-1-jht-10.1177_1096348021992099 for A Discrete Choice Experimental Approach to Understand Sports Event Tourists’ In-Stadium Beer Consumption Preferences by Seong Ok Lyu and Jinsoo Hwang in Journal of Hospitality & Tourism Research

Footnotes

Authors’ Note:

This work was supported by the Korea University Future Research Grant K1823691.

ORCID iD

Jinsoo Hwang

Seong Ok Lyu (e-mail: watchdog88@korea.ac.kr) is an associate professor in the School of Global Sport Studies, Korea University, Sejong, South Korea. Jinsoo Hwang (corresponding author; e-mail: jhwang@sejong.ac.kr) is an associate professor in the College of Hospitality and Tourism Management, Sejong University, Seoul, South Korea.

References

Aquilani

Laureti

Poponi

Secondi

(2015). Beer choice and consumption determinants when craft beers are tasted: An exploratory study of consumer preferences. Food Quality and Preference, 41, 214-224. https://doi.org/10.1016/j.foodqual.2014.12.005

Arnold

J. P.

(2005). Origin and history of beer and brewing: From prehistoric times to the beginning of brewing science and technology. Beer Books.

Arul

(2015). Drinking and game day: The expansion and solution to alcohol abuse in collegiate sporting events. The Compass, 1(2), Article 7. https://scholarworks.arcadia.edu/cgi/viewcontent.cgi?article=1005&context=thecompass

Ben-Akiva

M. E.

Lerman

S. R.

(1985). Discrete choice analysis: Theory and application to travel demand. MIT Press.

Bloomfield

Stockwell

Gmel

Rehn

(2003). International comparisons of alcohol consumption. Alcohol Research & Health, 2 7(1), 95-109.

Bolluyt

(2018, February 1). These are the most overpriced foods for sale at sports stadiums. CheatSheet. https://www.cheatsheet.com/culture/the-most-overpriced-foods-at-sports-stadiums.html/?a=viewall

Brady

K. T.

Sonne

S. C.

(1999). The role of stress in alcohol use, alcoholism treatment, and relapse. Alcohol Research & Health, 23(4), 263-263. https://doi.org/10.1097/00000374-199902000-00011

Cebula

R. J.

(2013). A panel data analysis of the impacts of regional economic factors, marketing and promotions, and team performance on minor league baseball attendance. Annals of Regional Science, 51(3), 695-710. https://doi.org/10.1007/s00168-013-0558-0

Chastain

A. J.

Gohmann

S. F.

Stephenson

E. F.

(2017). Beer availability and college football attendance: Evidence from mid-major conferences. Journal of Sports Economics, 18(6), 592-600. https://doi.org/10.1177/1527002515589351

10.

Chupp

Stephenson

E. F.

Taylor

(2007). Stadium alcohol availability and baseball attendance: Evidence from a natural experiment. International Journal of Sport Finance, 2(1), 36-44.

11.

Courtney

K. E.

Polich

. (2009). Binge drinking in young adults: Data, definitions, and determinants. Psychological Bulletin, 135(1), 142-156. https://doi.org/10.1037/a0014414

12.

Dodd

(2016, January 27). Alcohol: Coming soon to a college football stadium near you. CBSSports.com. https://www.cbssports.com/college-football/news/alcohol-coming-soon-to-your-college-football-stadium-if-its-not-there-already/

13.

Donadini

Fumi

M. D.

Kordialik-Bogacka

Maggi

Lambri

Sckokai

(2016). Consumer interest in specialty beers in three European markets. Food Research International, 85, 301-314. https://doi.org/10.1016/j.foodres.2016.04.029

14.

Donadini

Porretta

(2017). Uncovering patterns of consumers’ interest for beer: A case study with craft beers. Food Research International, 91, 183-198. https://doi.org/10.1016/j.foodres.2016.11.043

15.

Filce

Hall

S. A.

Phillips

(2016). Stadium alcohol management: A best practices approach. International Journal of Sport Management, Recreation & Tourism, 21(c), 48-65. https://doi.org/10.5199/ijsmart-1791-874X-21c

16.

Fitzpatrick

(2015). Broken bats and broken bones: Holding stadium owners accountable for alcohol-fueled fan-on-fan violence. Jeffrey S. Moorad Sports Law Journal, 22(2), 663-691.

17.

Gabrielyan

McCluskey

J. G.

Marsh

T. L.

Ross

F. C.

(2014). Willingness to pay for sensory attributes in beer. Agricultural and Resource Economics Review, 43(1), 125-139. https://doi.org/10.1017/S1068280500006948

18.

Gómez-Corona

Escalona-Buendía

H. B.

García

Chollet

Valentin

(2016). Craft vs. industrial: Habits, attitudes and motivations towards beer consumption in Mexico. Appetite, 96, 358-367. https://doi.org/10.1016/j.appet.2015.10.002

19.

Hensher

D. A.

Rose

J. M.

Greene

W. H.

(2015). Applied choice analysis. Cambridge University Press. https://doi.org/10.1017/CBO9781316136232

20.

Holdstock

King

A. C.

Wit

(2000). Subjective and objective responses to ethanol in moderate/heavy and light social drinkers. Alcoholism: Clinical and Experimental Research, 24(6), 789-794. https://doi.org/10.1111/j.1530-0277.2000.tb02057.x

21.

Kuhfeld

W. F.

(2005). Marketing research methods in SAS: Experimental design, efficiency, coding and choice designs. SAS Institute.

22.

Lenk

K. M.

Toomey

T. L.

Erickson

D. J.

(2009). Alcohol-related problems and enforcement at professional sports stadiums. Drugs: Education, Prevention and Policy, 16(5), 451-462. https://doi.org/10.1080/09687630802278478

23.

Lim

(2016, October 14). Delectable wonders at baseball stadiums in Korea. The Korea Herald. http://www.koreaherald.com/view.php?ud=20161014000748

24.

Louviere

Hensher

D. A.

Swait

(2000). Stated choice methods: Analysis and applications. Cambridge University Press.

25.

Lyne

Galloway

(2011). Strategies for managing alcohol at large events in New Zealand. Drugs: Education, Prevention and Policy, 18(6), 438-446. https://doi.org/10.3109/09687637.2011.608392

26.

Lyu

S. O.

Lee

(2018). How do golf tourists manage golfing constraints? A choice modeling approach. Journal of Hospitality & Tourism Research, 42(2), 295-318. https://doi.org/10.1177/1096348015597036

27.

Manski

C. F.

(1977). The structure of random utility models. Theory and Decision, 8(3), 229-254. https://doi.org/10.1007/BF00133443

28.

McCluskey

J. J.

Shreay

(2011). Culture and beer preferences. In Swinnen

J. F. M.

(Ed.), The economics of beer (pp. 161-170). Oxford University Press.

29.

McFadden

(1974). The measurement of urban travel demand. Journal of Public Economics, 3(4), 303-328. https://doi.org/10.1016/0047-2727(74)90003-6

30.

McFadden

Train

(2000). Mixed MNL models for discrete response. Journal of Applied Econometrics, 15(5), 447-470. https://doi.org/10.1002/1099-1255(200009/10)15:5%3C447::AID-JAE570%3E3.0.CO;2-1

31.

Menaker

B. E.

Connaughton

D. P.

(2010). Stadium alcohol policies: A comparison of policies available on college athletic department web sites. International Journal of Sport Communication, 3(2), 151-162. https://doi.org/10.1123/ijsc.3.2.151

32.

Merlo

L. J.

Hong

Cottler

L. B.

(2010). The association between alcohol-related arrests and college football game days. Drug and Alcohol Dependence, 106(1), 69-71. https://doi.org/10.1016/j.drugalcdep.2009.07.012

33.

Mitchell

Montgomery

R. D.

(2015). Beer and ball on campus? The issue of in-stadium alcohol sales. The Sport Journal, 18. https://www.cabdirect.org/cabdirect/abstract/20163109641

34.

Muret

(2014, January 13). Tapping new revenue: Sports venues give a toast to the trend toward craft beers. SportsBusiness Journal. https://www.sportsbusinessdaily.com/Journal/Issues/2014/01/13/In-Depth/Craft-beer.aspx

35.

Nelson

T. F.

Wechsler

(2003). School spirits: Alcohol and collegiate sports fans. Addictive Behaviors, 28(1), 1-11. https://doi.org/10.1016/S0306-4603(01)00296-9

36.

Park

(2016, October 24). Imported beer brands win out. The Korea Times. http://www.koreatimes.co.kr/www/news/biz/2016/10/123_216684.html

37.

Paul

R. J.

Paul

K. K.

Toma

Brennan

(2007). Attendance in the NY-Penn Baseball League: Effects of performance, demographics, and promotions. New York Economic Review, 38(1), 72-81.

38.

Paul

R. J.

Toma

M. T.

Weinbach

A. P.

(2009). The minor league experience: What drives attendance at South Atlantic league baseball games. Coastal Business Journal, 8, 70-84.

39.

Read

J. P.

Wood

M. D.

Kahler

C. W.

Maddock

J. E.

Palfai

T. P.

(2003). Examining the role of drinking motives in college student alcohol use and problems. Psychology of Addictive Behaviors, 17(1), 13-23. https://doi.org/10.1037/0893-164X.17.1.13

40.

Rees

D. I.

Schnepel

K. T.

(2009). College football and crime. Journal of Sports Economics, 70(1), 68-87. https://doi.org/10.1177/1527002508327389

41.

Reinert

D. F.

Allen

J. P.

(2007). The alcohol use disorders identification test: An update of research findings. Alcoholism: Clinical and Experimental Research, 31(2), 185-199. https://doi.org/10.1111/j.1530-0277.2006.00295.x

42.

Rogers

J. D.

Greenfield

T. K.

(1999). Beer drinking accounts for most of the hazardous alcohol consumption reported in the United States. Journal of Studies on Alcohol, 60(6), 732-739. https://doi.org/10.15288/jsa.1999.60.732

43.

Room

Babor

Rehm

(2005). Alcohol and public health. The Lancet, 365(9458), 519-530. https://doi.org/10.1016/S0140-6736(05)17870-2

44.

Saunders

J. B.

Aasland

O. G.

Babor

T. F.

de la Fuente

J. R.

Grant

(1993). Development of the alcohol use disorders identification test (AUDIT): WHO collaborative project on early detection of persons with harmful alcohol consumption-II. Addiction, 88(6), 791-804. https://doi.org/10.1111/j.1360-0443.1993.tb02093.x

45.

Shin

J. H.

Lyu

S. O.

(2019). Using a discrete choice experiment to estimate spectators’ willingness to pay for professional baseball park sportscape. Sport Management Review, 22(4), 502-512. https://doi.org/10.1016/j.smr.2018.06.009

46.

Smola

(2017, December 10). Ohio Stadium beer profits top $1.35 million. The Columbus Dispatch. http://www.dispatch.com/news/20171210/ohio-stadium-beer-profits-top-135-million

47.

Sohn

(2016, January 7). Beer emerging as Korea’s top-selling alcohol. The Korea Herald. http://www.koreaherald.com/view.php?ud=20160107001091

48.

Speece

M. W.

Kawahara

S. L.

(1994). Imported beer in the Hong Kong market. British Food Journal, 96(1), 10-18. https://doi.org/10.1108/00070709410050584

49.

Train

K. E.

(2003). Discrete choice methods with simulation. Cambridge University Press.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.20 MB