The Effect of Food Tourism Behavior on Food Festival Visitor’s Revisit Intention

In this paper, the effects of novelty-seeking and food involvement on food tourism behavior and the effect of food tourism behavior on intention to revisit the Jeonju Bibimbab Food Festival are studied. Additionally, the effects of attitude, anticipated emotions, perceived behavioral control, and subjective norms on intention to revisit through desire are identified. Three hundred and forty questionnaires are used for the study. These are analyzed with partial least squares structural equation model (PLS-SEM) using SmartPLS 2.0. The results show that two variables of novelty-seeking and food involvement have positive effects on food tourism behavior. It is also found that the effect of food involvement is higher on the food tourism behavior than that of novelty-seeking. This is because survey respondents are mostly domestic visitors, and for them, Bibimbab itself is familiar. In addition, the food tourism behavior has a significant effect on the intention to revisit the Jeonju Bibimbab festival. It implies that people who like to travel to enjoy food are also interested in the food festival.


Introduction
Among the cultural tourism resources that can represent the national or provincial culture and identity, food is an important tourism resource, because it can be easily accessed by tourists [1]. In fact, according to a survey of foreign tourists by the Korean Ministry of Culture, Sports, and Tourism [2], food accounted for the second-highest percentages (44.3%) after shopping (60.0%) of tourists' motives for visiting Korea. Even among the top shopping items of foreign visitors, food-related items such as groceries (36.2%), ginseng and herbal medicine (12.2%), and alcohol (5.8%) are included [3]. Especially, the food festival, of which food is the main subject, has a great value as a tourism resource that shows the identity of the region. Therefore, in each region, various forms of food festivals have been introduced in order to commercialize the local foods, containing its own food culture and ingredients [4].
Due to the local autonomy that began in the mid-1990s, local governments sought to promote regional development and economic revitalization. Since festivals can inform and introduce local culture and are an important element of the local economy, the influence of festivals on local governments, industries, and economies is enormous. Therefore, many local governments in Korea have been the main bodies to hold various events and festivals using foods and specialties representing their area, in order to preserve and promote local culture, and activate local economies through attracting tourists [5][6][7][8]. Examples include Gyungju's food festival using sake and rice cake, Jeollanam-do's food and culture festival, Gwangju's world Kimchi culture festival, Jeonju's Bibimbab festival, and so on. In particular, Jeonju is famous for being a world gourmet city and a well-preserved traditional Korean house and culture. So, it is the center of the project that the central government is promoting to introduce Korea to foreign countries. Therefore, the Jeonju Bibimbap festival has a great impact not only on the Jeonju region, but also on the national institutions that formulate the Korean tourism industry and tourism policy.
The Jeonju Bibimbab festival is held every autumn in Jeonju. The '2017 Jeonju Bibimbab festival' was held for four days from 26 October at the Korea Traditional Cultural Center with the theme of "Mix it Excitingly" and "Enjoy it Tastily". The festival was very successful, and the number of tourists visited was about 100,000, including foreigners. In the Jeonju Bibimbab festival, 38 programs were organized in five categories based on the excellent food culture of Jeonju, a global gastronomic city including Bibimbab. Bibimbab Performance, a representative program of the festival every year, attracted more than 10,000 people, including tourists from around the country and the world. Particularly, the festival was more successful by expanding the programs in which the visitors could participate.
As many food festivals such as Jeonju Bibimbab festival have become popular, local food festivals and food tourism fields are being studied as major research topics in academia. Some examples include empirical studies on the relationship between the physical environment of local food festivals and behavioral intentions [9,10], or between the motivation and satisfaction of visitors participating in local food festivals [11]. Sunwoo and Son [9] studied the influence of the service landscape of the food festival on visitor's behavioral intentions. According to their research, the aesthetic and functional factors of the food festival service landscape influence the behavioral intentions. Therefore, it can be seen that the physical environment of the food festival is a very important factor in forming visitors' behavioral intentions. Hwang and Lee [11] examined the effects of visitors' orientation on intention to revisit food festivals through satisfaction. They concluded that the revisitor has a tendency to find pleasure with many people and/or to expand himself in that. Min [12] conducted a study to investigate the effect of motivation on revisit intention to food festivals through satisfaction. He found that motivation was a significant factor to cause visitors to intend to revisit the food festival. Especially, if the motivation to visit food festivals is for the formation of social relations or for self-development, it can be seen that the intention to revisit is higher if they are satisfied through visits. In addition, research on service quality, satisfaction, and intention to revisit local food festivals has been actively conducted [13,14].
The prior studies on the food festivals revealed that the physical environment of the festival, motivation and tendency to visit, and the satisfaction factors of visitors such as experience are important factors. The environmental and experiential elements of festivals are important in food festivals. However, few studies have examined the relationship between food tourism behavior and food festivals, where foods are the most important elements of food festivals. Food tourism is becoming increasingly important as a new form of tourism. Therefore, the importance of food tourism in food festivals is also increasing. The tourism purpose of food tourists is the food-related experience itself, and they often have a high food involvement and a pursuit of novelty. Bibimbab is a universal food that is widely known to Koreans and foreigners, but bibimbab in Jeonju has a unique meaning for food tourists since it emphasizes the experience of local ingredients, cooking methods, ingestion methods, or the environment used in food. Therefore, since the Jeonju Bibimbab festival has elements of interest such as a festival experience and a local culture called the home of bibimbab, the Jeonju Bibimbab festival is a fascinating place for food tourists to visit again. However, despite this importance, there has been little research on the relationship between food festivals and food tourism behavior, as we have seen above. Therefore, in this study, we examine whether the novelty and the food involvement influencing the food tourism behavior are significant factors in the intentions to revisit the food festival, which is a fairly universal and well-known food theme of bibimbab. In addition, the impacts of attitude, subjective norms, perceived behavioral control, and anticipated emotions on intentions to revisit the Jeonju Bibimbab festival through desire are examined. These results could lay the foundations for developing effective strategies, programming planning, and the selection of foods for successful food festivals.

Model of Goal-Directed Behavior
A number of theories on attitude-behavior relations have emerged in social psychological literature. One of salient theories is the TRA (theory of reasoned action) [15], which proposes that individuals' attitude toward a behavior and significant others' pressure lead them to perform the behavior determining the intention, and subsequently behavior. Ajzen [16] addressed that the effect of behavioral intention on behavior might be changed according to situation. To tackle the circumstantial limitations of the TRA, the theory of planned behavior (TPB) [16] was developed by adding a new variable, perceived behavioral control.
Perugini and Bagozzi [17] developed the model of goal-directed behavior (MGB), which includes three variables of affective, habitual, and motivational processes to the TPB to increase the prediction of behavior. These new variables are considered as significant factors in human decision-making processes [17]. Wilson and Arvai [18] highlighted that affective responses to a particular behavior may even sometimes overpower analytic computations in human decision formation. Hence, adding anticipated affective reactions to TPB is a useful way to heighten understandings of behavior. As such, anticipated emotions are adopted as a determinant to represent the affective aspect of behavior in MGB. A number of studies argue that the frequent enacting of a behavior may result in subsequent behavior in terms of habitual processes, although past behavior does not necessarily have a causal relationship with subsequent behavior [19]. Therefore, past behavior is incorporated as predictors of desire and intention into the MGB. Desire is also known as playing an important role in predicting human behavior from the motivational aspects of human decision formation. In MGB, desire is incorporated as a mediator to explain how reasons for behavior are consolidated and emerged as behavioral intentions, which would enhance the model in terms of predictive ability.
Recently, many researchers have modified the MGB by adding new constructs or altering existing variables to capture the unique properties in a certain context [1,[20][21][22]. For an example, Taylor [20] proposed a new model of predicting consumers' information search behaviors under uncertain and risky conditions by augmenting the MGB. In this model, he incorporated another emotion of anticipated regret to the MGB, while other emotions remained in the model, because the feeling of regret is common and significant in decision formation under risk. Moreover, the attitude factor that was originally in the MGB was divided into two exogenous variables, utilitarian and hedonic attitude toward information search, to specifically consider the cognitive and affective aspects of attitudes. Song et al. [21] incorporated the construct of tourism behaviors into the MGB to examine its effect on the festival visitors' intention at the Mud Festival in Boryeong. In this study, perceived customer and effective environmental concerns were incorporated as predictors of environmentally-friendly tourism behavior, which is a determinant of desire in the MGB. In another study by Song and Lee [22], the MGB was extended by adding the perception of responsible gambling strategy to explain its role on desire and casino visitors' intention to gamble. Lee et al. [1] addressed an extended model of MGB in the tourism context by introducing two constructs of non-pharmaceutical intervention and the perception of 2009 H1N1(Hemagglutinin type1 and Neuraminidase type1) influenza to examine their effects on the behavioral intentions of prospective international tourists. Therefore, the MGB could be employed to predict the food festival visitor's decision formation process.

The Relationships among Food Involvement, Novelty Needs, and Food Tourism Behavior
Hall and Mitchell [23] defined food tourism as a visit to a food fair, a food festival, or a particular area and restaurant to taste or experience food. In addition, food tourism not only involves tasting the food produced in the area, it also includes experiencing the local culture [24]. Boyne, Hall, and Williams [25] suggested that the region hosting the food tourism should strive to create exotic emotions and images in order to differentiate it from other regions. To this end, local food and services should be strengthened more distinctively and uniquely. That is, food is a very important factor in food tourism. Therefore, the higher the involvement of people in food, the more likely they prefer food tourism. Based on these assumptions, Bell and Marshall [26] developed a scale to measure the food involvement of each person. The model says that people with higher food involvement tend to show higher food tourism behavior.
Hall, Sharples, Mitchell, and Camboume [27] classified three types of tourism according to their level of food involvement. In other words, the type of tourism most favored by tourists who are most interested in food is gourmet tourism, those with moderate interest are classified within food tourism, and those with a low interest in food as a whole are classified within local tourism.
Shenoy [28] presented four personal attributes that directly affect the decision of tourists to choose food tourism. These attributes are all related to food, such as the fearfulness of new foods, the tendency to pursue diverse foods, the pleasurable consumption tendency toward food, and the involvement of food.
In the meantime, an example of a study on the relationship of food involvement and food tourism in Korea includes the study of Lee and Kwon [29]. They conducted a study on the food involvement and behavioral intention of food tourists in the Jeolla provinces, and found that tourists with a high interest in local specialties and food showed continuous and high involvement in food tourism.
In addition, Yoo and Seo [30] conducted a study on the tourism behaviors of gastronomists with high food involvement. The results showed that tourists who enjoy gastronomy assign a high value to food culture experiences in tourist areas, and therefore, there is a positive influence relationship between their gastronomic activity and intention to take part in gastronomic tourism.
In the meantime, it is said that novelty-seeking is one of the motivations of travel as an important factor influencing tourism selection behavior, and it induces motivation to move out of the present familiar place and environment to new and unfamiliar places of travel [31]. The degree of novelty-seeking is different for each person, and therefore, much research is being done on the type of tourism behavior according to the novelty-seeking [32].
Moon and Han [33] examined the relationship between novelty-seeking and food tourism behavior through the influence of novelty-seeking in intentions to visit ethnic restaurant. Through this study, they demonstrated that ethnic restaurants may be relatively unfamiliar to tourists, and that travel with such unfamiliar experiences would be preferred by tourists seeking novelty. However, little research has been done on the effects of novelty-seeking in general food tourism, i.e., all types of food tourism. In addition to the restaurants that offer some unfamiliar food to tourists such as ethnic food, we could assume that novelty-seeking affects all forms of food tourism behavior. This is because, even if food is generally well-known, the nature of food tourism may tend to experience newness such as differences in materials, manufacturing methods, and differences in the food culture of the region.
Therefore, in this study, we examine the effect of food involvement and novelty-seeking on food tourism behavior through the following hypothesis. Based on a review of the literature, the following hypotheses are suggested:

H1.
Novelty-seeking has a positive effect on food tourism behaviors. H2. Food involvement has a positive effect on food tourism behaviors.

The Relationship between Attitude, Subjective Norms, Perceived Behavior Control, and Desire
Most researchers claim that intention is a significant determinant of behavior, and mediates the relationship between attitude and behavior [34,35]. Therefore, a person's behavior, as in the TRA and the TPB, is predictable from his attitude toward that behavior if intention and behavior are highly correlated [36]. It follows that a person's attitude toward performing the act represents his evaluation of that act. That is, a person who evaluates the suggested act as positive has a higher intention.
Bagozzi [37] introduced the desire as a determinant of intention to address the predictive ability in the MGB. More specifically, the desire in the MGB is a function of attitude, perceived behavior control, subjective norms, and anticipated emotions, and in turn is a major predictor of intention [17]. In other words, a person's attitude, subjective norms, and anticipated emotions give indirect impacts on intention only through desire, and perceived behavior control affects intention directly or indirectly through desire in the MGB. Therefore, a person's attitude acts as a predictor of desire in the MGB. Based on a review of research, the proposed hypothesis is as followed: H3. Attitude has a positive effect on desire.

The Relationship between Desire and Subjective Norm
Subjective norms prevent or do not prevent specific behaviors [36]. In the TPB, subjective norm acts as a determinant of intention in the decision-making process. When undertaking certain specific actions, individuals have the potential to follow or consider the opinions of nearby people (friends, family, and colleagues) [35]. Therefore, people tend to undertake the particular behavior that their significant others expect them to perform. That is, if an individual thinks that other people expect him to do an act, then he has higher intention to do that act, and in turn he has a higher tendency to do so. The role of a person's behavioral intention in the TPB is revised in the MGB, where subjective norms have an effect on the behavioral intention through desire. Specifically, Taylor [20] found that, under uncertain, risky conditions, subjective norm affects desire. In the study of the behavior of visiting the complex resort casino, Song and Lee [36] reported that the desire of their own to visit a casino is higher when nearby people who have an important influence on them give positive messages about their casino-visiting behavior. Therefore, this study hypothesized that the desire to revisit the Jeonju Bibimbab festival will be influenced by the groups of people who are meaningful to them.
H4. Subjective norm has a positive influence on desire.

The Relationship between Anticipated Emotions and Desire
The TPB excludes the emotional aspects experienced by consumers because it consists of cognitive factors such as attitude, perceived behavior control, and subjective norms. However, the goal-oriented behavioral model suggests positive/negative anticipated emotions for goal achievement or failure as predecessors of the decision-making process. These anticipated emotions are predictive of aspiration as TPB variables, and are important determinants of behavioral intentions [38].
Perugini and Bagozzi [17] found that anticipatory emotions increase the explanatory variance of behavior, and are important predictors of decision making. They argued that positive anticipatory emotions can lead to decisions to keep the current plan in relation to achieving goals, while negative anticipatory emotions occur when the ongoing plan has problems and the desired goals are not achieved.
For example, Parker, Manstead, and Stradling [39] found that the anticipated regret regulated the expectations for behavior involved in transgression violations; thus, negative emotions affected behavioral intentions. Sheeran and Orbell [40] found that the anticipated regret was affecting the intention to purchase the lottery. Their research focused only on negative anticipatory emotions, and based on these findings, it is possible to form emotions by acting pre-factual on the outcome of achieving or not achieving their goals.
Thus, anticipated emotions involve a dynamic self-regulatory process involving an evaluation of future goals (success and failure) and an emotional evaluation of self. Also, anticipatory predictive and reflective counterfactuals can be synchronized to avoid negative anticipated emotions and promote positive anticipatory emotions, affecting intent and behavior. Therefore, in this study, the following hypotheses were established to examine the effects of anticipated emotions (positive and negative) on desire.

H5.
Positive anticipated emotion has a positive effect on desire. H6. Negative anticipated emotion has a positive effect on desire.

The Relationship among Perceived Behavioral Control, Desire, and Behavior Intention
Perceived behavioral control involves external factors and internal factors. The external factors include time, financial allowance, and other material resource factors, whereas the internal factors imply individual ability such as self-efficacy or self-confidence [34].
For example, if a person has the necessary resources (time or transportation) for him/her to go on vacation, there is more possibility that he/she will form the intention to perform the act. In other words, the actor has a tendency to consider whether they have the ability and it is the situation to fully utilize the necessary resources, opportunities, and time before executing the target action. If there are sufficient resources, opportunities, and time for the target action, the action subject will increase the willingness to perform the target action. If not, then the actor will act in a direction to lower the target behavior [41].
Therefore, in this study, the following hypothesis was set up to examine the effects of perceived behavioral control on aspiration and behavioral intention H7. Perceived behavior control has a positive influence on desire. H8. Perceived behavior control has a positive influence on behavior intention.

The Relationship between Food Tourism Behavior and Behavior Intention
Lee and Kwon [29] found five components of food tourism in the study of the relationship between food tourism and revisit intention, and explained the effect of each component on revisit intention. Five components include 'local specialties', 'local foods', 'local drinks', 'familiar foods in the region', and 'local fine food'. Among the five factors, 'familiar food', 'local food', and 'high-quality food in the region' had positive effects on revisit intention. Based on the research of Lee and Kwon [29], Lee [30] demonstrated the relationship between these factors and revisit intentions. The results of the study showed that four factors that are related to participation in food tourism are 'food culture experience', 'knowledge/information-seeking of food', 'health orientation', and 'food-learning seeking'. In addition, among these factors, the 'food culture experience', 'health orientation', and 'knowledge/information-seeking of food' factors showed significant influence on revisit intention. From these results, it can be seen that tourists tend to pursue cultural experiences, interactions, and health through food. Therefore, it can be expected that the behavior of food tourism to be studied in this study will have an important influence on the formation of revisit intentions about the Jeonju Bibimbab Food Festival. The following hypothesis is suggested.
H9. Food tourism behaviors have a positive influence on behavior intention.

The Relationship between Desire and Behavior Intention
In the goal-directed behavior model, various evaluation factors such as attitude, subjective norms, perceived behavior control, and anticipated emotions directly affect the desire of an emotional driving factor. This desire is said to have an effect on behavioral intention through the influence of the evaluative factors [1,17,20,42].
In other words, if tourists are positive about their intentions to revisit the Jeonju Bibimbab festival, and the people around them also agree with these thoughts, and they have enough ability and opportunity to revisit or have positive feelings about revisiting the Jeonju Bibimbab festival, it can be expected that they are likely to have the desire to revisit the festival. In addition, it seems that this desire may affect the actual intention to revisit. Therefore, in this study, the following hypothesis was set up to examine the influence of the desire to revisit the food festival on the behavioral intention.

H10.
Desire has a positive effect on behavior intention.
In summary, this extended MGB can serve as a new theoretical framework to explain the behavior intention of the food-based festival behaviors. Therefore, the theoretical model that is proposed in this study is shown in Figure 1.

Measurements
Based on prior research [1,17,34,[43][44][45][46][47], the questionnaire was prepared to measure the variables that are utilized in this study. Multiple items were used to assess all of the constructs (i.e., novelty-seeking, food involvement, food tourism behaviors, attitude, the subjective norms, anticipated emotions, perceived behavior control, desire, and revisit intention) with a five-point Likert scale (1 = strongly disagree, 5 = strongly agree).
The original latent variables of MGB were modified to reflect the context of the Jeonju Bibimbab festival as follows. The respondents' attitudes toward the Jeonju Bibimbab festival were measured with six items (e.g., "I think that visiting the Jeonju Bibimbab festival is a [desirable, pleasant, attractive, exciting, enjoyable, valuable] behavior"). Similarly, subjective norms were assessed with four items (e.g., "Most people who are important to me [agree with, recommend, support, and understand] that I visit the Jeonju Bibimbab festival"). Anticipated emotions were evaluated with six items (three items on positive anticipated emotions and three items on negative anticipated emotions) (e.g., "If I will visit the Jeonju Bibimbab festival, I will be [delighted, pleasant, happy"] and "If I can't visit the Jeonju Bibimbab festival, I will be [angry, unpleasant, disappointed"]). Lastly, the perceived behavior control was assessed with four items (e.g., "I am confident that if I want, I can visit the Jeonju Bibimbab festival," "I am capable of attending the Jeonju Bibimbab festival," "I have enough resources (money) to visit the Jeonju Bibimbab festival"). Desire is operationalized with three items (e.g., "I [would like to, hope to] visit the Jeonju Bibimbab festival in the near future," and "I [want to have fun] when I visit the Jeonju Bibimbab festival in the near future"). Revisit intention was measured with two items (e.g., "I am planning to revisit the Jeonju Bibimbab food festival" and "I am willing to revisit the Jeonju Bibimbab food festival").
Three newly added constructs, novelty-seeking about food, food involvement, and food tourism behaviors were derived from the tourism literature [26,28]. Novelty-seeking was assessed with eight items [48].
Food involvement was measured with seven items [26]. Four items were used to assess food tourism behavior [28].

Jeonju Bibimbab Festival
In recent years, local government agencies and organizations have made their best efforts to attract tourists by holding various events and festivals containing their own unique cultures and foods to promote the local economy and traditional culture [5]. Among the various regional festivals, food festivals using local foods are popular, since foods are representative of the unique image of the region and easily accessible to tourists.
Since local food festivals should reflect the cultural characteristics of the region and promote the local economy, foods are, apparently, the most important components of the food festivals [49].
There are three types of local foods that are used as important materials in local food festivals. First, there are foods that are cooked using local or regional produce as their main ingredients. Second, there are foods that are cooked by the unique recipes of the region that have been handed down for a long time. Third, there are the foods that have been enjoyed by the local residents for a long time. These local foods are a temporal cultural heritage with a cultural experience value that has been accumulated for a long time in the local residents, and has a regional characteristic that is distinct from other regions [49]. One of the reasons why local food festivals using local foods and specialties have recently attracted interests from tourists is that their income levels have improved. In other words, because of the increase in the income level, people are increasingly desirous to taste various foods and enjoy the taste of food [50]. In addition, local food is a cultural tourism resource that can be easily accessed by tourists visiting the area, and that allows tourists to experience the long tradition and culture of the area.
Among these local foods, Bibimbab is a traditional food representing the culture of Jeonju province, and has become a representative food of Korea. Korean Air's Bibimbab was selected as the "Mercury Gold Award" by the International Travel Catering Association (ITCA) in 1998, and it was the opportunity to promote Bibimbab as a representative food in Korea [51].
Also, in 2012, Jeonju city was selected as the first food craft city designated by UNESCO in Korea. Thanks to this background, Jeonju Bibimbab festival, which is held every autumn in Jeonju City, has become one of the most successful festivals among the various regional food festivals that are held in Korea. Jeonju Bibimbab festival was based on Jeonju's representative food, 'Bibimbab', which is characterized by harmony and individuality. The theme of taste and health has been held every year in Jeonju as an annual event since it started in 2007, except for being canceled in 2009 due to influenza. The Jeonju Bibimbab festival was originally held under the name of 'Jeonju's 1,000-year banquet feast', and it has been called the present name since the third festival. The Jeonju Bibimbab festival consists of five themes: 'Delicious Bibim', 'Pleasant Bibim', 'Different Bibim', 'Healthy Bibim', and 'Together Bibimb'. Food festivals, cooking concerts, artistic craftsmanship, and 'Bibimb' performances provide tourists and participants with enjoyable sightseeing and fun. In particular, the 'Daedong Bibimb' performance, where citizens and tourists participate together, is the most popular program.

Data Collection
The survey was conducted for the visitors who participated in the 10th Jeonju Bibimbab Festival by five researchers who were fully trained in the purpose of the study and the survey method. The questionnaires were collected during the four-day festival period (26-29 October 2017).
A total of 472 valid samples out of the 500 surveys that were collected were used for the empirical analysis after deleting 28 questionnaires whose responses were judged to be somewhat unfaithful or missing. Table 1 shows the demographic profile of the respondents. The male (231, 48.9%) and female (241, 51.5%) sex ratios of the samples are almost the same. In terms of age distribution, 298 respondents in their 20s accounted for 63.1% of the total, while 16.7% (79 people) were in their 30s, 10.8% (51 people) were in their 40s, and 9.2% (44 people) were in their 50s. Most-69.3% (327 persons) of the total-were married, and most of the respondents were highly educated college graduates (179, 37.9%) and university graduates (185, 39.2%). The amount of money spent by festival visitors at the festival was found to vary (less than $50 = 40.7%; $50-100 = 37.3%; $101-200 = 17.8%; more than $200 = 4.2%).

Measurement Model
Statistical analysis of the collected data was conducted using smart partial least squares (Smart PLS). The adequacy of the measurement model was tested by examining the indicator reliability, internal consistency, convergence validity, and discriminant validity of the indicator variables [52]. The initial result of the partial least squares structural equation model (PLS-SEM) showed that the factor loadings of one item of the novelty-seeking and one item of the perceived behavior control were less than 0.7. By removing these two items, the indicator reliability was secured [53]. Internal consistency was assessed by the composite reliability, which is known to be more suitable than Cronbach's alpha in PLS-SEM [54]. Table 2 showed that the latent variables' composite reliabilities varied from 0.823 to 0.973, all of which were greater than the recommended threshold of 0.70 [48]. Thus, the modified measurement model had good internal consistency.
Convergent validity was evaluated by using the average variance extracted (AVE) for each latent variable. From Table 2, the results showed that the AVE varied from 0.608 to 0.931, all of which were higher than the suggested threshold of 0.5 [53]. Therefore, the suggested measurement model showed adequate convergent validity. To show the discriminant validity, the square root of AVE for each latent variable was evaluated. As shown in Table 3, the square root of AVE for each variable is greater than its correlation with any other latent variables [53]. Therefore, the Fornell-Larcker criterion was satisfied.

Structural Equation Model
The structural model and proposed hypotheses were tested by using SmartPLS 2.0 [54]. Since PLS does not provide overall goodness of fit indices, the variance explained (R2) and the goodness-of-fit (GoF) index were used to assess the model fit [55,56]. The R2 for each endogenous latent variable is a primary criterion to evaluate the explanatory power of the structural model [55]. The results showed that the proposed model explains 43.3%, 53.4%, and 44.1% of the variance of food tourism behavior, desire, and revisit intention, respectively. Thus, we could conclude that the R2 value represented a moderate explanatory power of the proposed model [51]. The model fit was assessed by using the GoF [56]. The GoF can be generated by the geometric mean of the average R2 for endogenous latent variables and average AVE. For the proposed model, a GoF value of 0.61 was larger than the cut-off value (GoFlarge = 0.36). Therefore, the GoF value of the model indicates a very good model fit [57]. As shown in Table 4, the results of hypotheses testing are presented. Hypotheses 1 and 2 posit that novelty-seeking and food involvement have positive effects on food tourism behavior. The results show that novelty-seeking (coefficient = 0.282, t = 3.596, p = 0.00) and food involvement (coefficient = 0.413, t = 6.278, p = 0.00) positively influence food tourism behavior, supporting H1 and H2. Furthermore, the results show that all of the predictors of desire are statistically significant, with linkages between attitude and desire (H3: coefficient = 0.216, t = 5.084, p = 0.00); subjective norms and desire (H4: coefficient = 0.422, t = 8.929, p = 0.00); positive anticipated emotions and desire (H5: coefficient = 0.134, t = 3.709, p = 0.00); negative anticipated emotions and desire (H6: coefficient = 0.111, t = 3.218, p = 0.018); and perceived behavioral control and desire (H7: coefficient = 0.142, t = 3.972, p = 0.00). Thus, H3, H4, H5, H6, and H7 are supported. However, the linkage between perceived behavioral control and revisit intention (coefficient = 0.012, t = 0.320, p = 0.749) is not significant. Thus, H8 is not supported. Food tourism behavior and desire are found to be significant predictors of revisit intentions. The results show that food tourism behavior and desire positively influence revisit intentions. Thus, both H9 and 10 are supported. Overall, five original latent variables of MGB are significant in predicting the festival visitor's desire formation to revisit the food festival. Two latent variables (food tourism behavior and desire) played an important role as essential antecedents in explaining the festival visitor's intention to revisit the food festival.

Results and Implications
This study proposed the extended goal-oriented behavior model (EMGB) to predict the behavioral intention to revisit the Jeonju Bibimbab festival. That is, food involvement, novelty-seeking, and food tourism behavior factors are added to the MGB model to examine the influence of these special forms of tourism-related constructs on the intention to revisit the food festival. The results of the analysis can be summarized as follows.
First, food involvement and novelty-seeking have a positive effect on food tourism behavior. In other words, the more involved in food people are, the more likely they prefer food tourism. It also shows that people who seek novelty in food prefer food tourism. This means that even if people have familiarity with Bibimbab, novelty-seeking people feel newness if the food is cooked with a distinct recipe or ingredients from the region. Therefore, it is better to conserve and develop the distinct food culture of the region. Also, the food tourism behavior has a positive influence on the intention to revisit the food festival. In other words, food tourism behavior is a significant antecedent in predicting the visitor's intention to revisit the food festival.
Second, attitudes toward the Jeonju Bibimbab festival showed a positive effect on desire. The cognitive and emotional attitudes toward the Jeonju Bibimbab festival can be said to form a desire to revisit the food festival. This means that tourists visiting the Jeonju Bibimbab festival have a strong desire to revisit the festival when they feel pleasant, interesting, and valuable toward the festival. Therefore, it is important to focus on the development of programs that can be enjoyed by the direct participation of Bibim performances, which are especially popular with tourists visiting the Jeonju Bibimbab festival. Therefore, festival managers should emphasize the development of programs that visitors can participate and enjoy, such as Bibim performance, which is especially popular with tourists visiting the festival.
Third, subjective norms have a positive effect on desire. This means that the evaluations of the Jeonju Bibimbab festival by the reference group that is important to the visitor, such as the family, friends, and neighbors of the visitor, have a great influence on the formation of visitor's desire to revisit the food festival. Therefore, it can be concluded that the positive evaluation of the festival by using diverse social media such as blog and Twitter, and the word-of-mouth marketing activities for the people around the tourists, will be very important.
Fourth, anticipated emotions (positive and negative) about visiting the Jeonju Bibimbab festival showed a positive effect on desire. This means that when revisiting the food festival, the higher the positive or negative emotion expected, the greater the desire to revisit. In other words, it implies that festival visitors' emotions are important both when they achieve goals and when they do not. Therefore, it is suggested that efforts should be made in the operation planning and marketing activities of the festival to ensure that festival visitors have positive anticipated emotions and negative anticipated emotions.
Fifth, perceived behavioral control showed a positive effect on the desire to revisit the festival. This shows that visitors have a desire to revisit the food festival based on their internal factors, such as personal ability, health status, or external factors such as money and time. This means that these internal and external factors play an important role in the formation of the desire to revisit the festival. On the other hand, perceived behavioral control did not have a direct effect on behavioral intentions to revisit the festival. However, the relationship between perceived behavioral control and behavioral intention was positively correlated, although it was not statistically significant.
Sixth, the desire to revisit the festival showed a positive effect on the behavioral intention to revisit the food festival. This shows that the desire to revisit the Jeonju Bibimbab festival is an important motivating factor for behavioral intention. These desires are influenced by emotional factors such as positive and negative anticipated emotions, attitudes, and perceived behavioral control. In other words, if the favorability of festivals, anticipated emotions, and temporal and financial ability become possible, the desire to revisit the Jeonju Bibimbab festival is formed, and this leads to the behavioral intention of action. Therefore, in order for the food festival visitors to revisit the Jeonju Bibimbab festival, it is important to make efforts to form visitors' desires to revisit the festival.
Finally, it can be seen that the food tourism behavior has a positive influence on the behavioral intention to revisit the festival. In other words, those who make the food itself as a purpose for their travel are also interested in revisiting the food festival. Their tendency is that they want to experience something unique in the food itself, but they also want to experience the food produced in the area or cooked in the region's own recipe. Furthermore, they also want to have diverse experiences by participating in a food-themed festival. Therefore, it is important for the festival managers to understand the visitor's propensity of the food and plan the recreation of local recipes, the development of local food ingredients, and the preparation of participatory programs on food, so that they can come back.

Conclusions
The results of this study can be applied to preparing marketing strategies and the local government's planning of tourism policy related to local food festivals, and can also contribute to academic development. From the marketing perspective, to make food travelers loyal, they should grasp the tendencies of food travelers who enjoy food tourism. In our research, it is shown that food tourists are willing to revisit the Jeonju Bibimbab festival. Although food tourists seek newness, the high intent to revisit a relatively well-known food-themed festival suggests a great deal for marketers who are planning food festivals. In other words, food does not have to be unfamiliar or not widely known when selecting the right food for the food festival. If the food should be unfamiliar, the organizer who prepares the food festival must prepare new food every time, so the success of the festival may not be continuous. Therefore, it may be more effective for the marketing planners to concentrate on the performance program or theme development, where everyone can participate by utilizing existing food rather than discovering new food. The same implications can be drawn from the effects of the attitude toward the food festival on the desire to revisit it. That is, food tourists who visit the Jeonju Bibimbab festival desire to revisit it if they have a pleasant time at the festival. It is known that tourists enjoyed the Bibim performances because they can participate the performance. It is also important for the markers to know that the reference group to the visitor plays a great role in forming the visitor's desire to revisit the festival. Thus, markers should strive to create the positive image or evaluation of the festival by utilizing diverse social media and word-of-mouth.
On the other hand, research shows that even if the food is same, food tourists take it as a unique and new food if it has different local ingredients and recipes. Therefore, in order to lead these food tourists to return to food festivals, local governments' efforts to produce and excavate local specialties or materials, research new recipes, and train chefs are needed [58]. The academic contribution of this study is that the effect of food tourism behavior on the intention to revisit the Jeonju bibimbab festival, one of the most successful food festivals in Korea, is investigated for the first time. In addition, it is also meaningful that in this study, it is empirically verified that the pursuit of novelty in food and the involvement of food have a direct effect on food tourism behavior. Therefore, the founding and results of this study can be used for the study of food festivals in different regions in many different ways.
This study has some limitations that might be dealt with in future research. First, since the data used in the study were collected from only one specific food festival in Korea, the findings of the study might not be applied to other food-based festivals. Thus, further studies utilizing the proposed extended MGB in other food-based festivals is expected to generalize the results from this study. Second, in this study, the original MGB is extended to fit in a food-based festival setting by introducing food tourism behavior constructs. The development of other extended MGB, including some variables that might be more essential for predicting the food-based festival visitor's intention-formation process, might be a good direction for future research.