Bagozzi (2000) proposes an action theory model of consumption. Under this model, the consumer under­goes a rational process (conscious or unconscious) based on perceived stimuli, mentally appraising the anticipated outcomes. These outcomes are experienced as positive and negative anticipatory emotions. Desire is produced in a subsequent stage, where reasoning, emotional, and social processes are integrated and transformed into a decision to act or not to act. Finally, additional affective and reasoning processes are carried out, beginning in a stage called "trying to consume," in which decisions are planned and implemented and behaviors are directed toward achieving the desired goal.

In the context of food, Steenkamp (1997) highlights the influence of psychological factors in consumers’ food-choice processes, and Shepherd (2011) examines the influence of affective aspects on food choice. Several studies have demonstrated the existence of a direct relationship between emotions and the evaluations consumers make ( e.g., Oliver et al., 1997; Reinares‐Lara et al., 2016; Salazar-Ordóñez et al., 2018). This relationship has also been found in the context of wine ( e.g., Barreiro-Hulé, 2007; Olarte et al., 2017). To determine the influence of emotions in the "trying to consume" stage with wine, consideration has been given to three types of emotions:

(i) Influence of the emotions produced by the product (the offered wine) on purchase intent. Products create a mental image in people generated by emotions. This image, in turn, shapes their evaluation of the product (Mittal, 1994; Elliot, 1998; Laverie et al., 2002; Pelegrín-Borondo et al., 2016). In general, objects or events that produce a positive emotion are evaluated positively, while those that generate a negative emotion are evaluated negatively (Bagozzi et al., 1999; Mano, 2004). When the products and services a consumer evaluates generate positive emotions, these emotions increase the value for the consumer (Yeung & Wyer, 2004; Yeh et al., 2012), who thus becomes more favorably disposed to buy the product (Oliver et al., 1997; Bagozzi et al., 1999; Shiv & Fedorikhin, 1999). The opposite is true when negative emotions come into play (Elliot, 1998). In other words, consumers do not tend to purchase products and services that generate negative emotions in them (Schwarz, 2000; Han et al., 2007).

Often, consumers choose products without having tried them. These products produce emotions that influence the purchase decision. In this regard, the theory of bounded rationality holds that consumers make decisions about products in the absence of complete information (Kahneman, 2003; Thaler, 2008). Ariely (2008) argues that consumers mentally make their choice before trying a product.

With regard to food products, Siegrist (2008) in­cludes psychological factors in models for innova­ti­ve products. These factors include emotions, which are key to knowing why consumers choose to buy some food items but not others (King et al., 2010; Barrena & Sánchez, 2013). King & Meiselman (2010) show that emotional intensity is related to food acceptance. Lockie et al. (2004) demonstrate that the sensory and emotio­nal experiences involved are two of the determinant factors in the growing consumption of organic food products. Dalenberg et al. (2014) and Gutjar et al. (2015) observe that the emotions produced by food are a good predictor of its selection by consumers, while Organ et al. (2015) establish that positive emotions toward certain foods develop when the consumer associates positive experiences with them.

Several researchers have studied the influence of negative emotions on the acceptance of certain foods. Ronteltap et al. (2007) highlight the strong role pla­yed by negative emotions in consumer rejection of certain food innovations. This is quite clear in genetically modified food, where both perceived benefits and perceived risks have been found to influence the purchase decision-making process (Bredahl, 2001; Rodríguez et al., 2013). Barrena & Sánchez (2013) demonstrate the importance of emotions in neophobic people when it comes to tasting a new product. Neophobic people avoid new foods for fear they might make them sick or cause some other malady (Jaeger et al., 2003).

Specifically with regard to wine consumption, Horska et al. (2016) observe that people have positive and negative experiences during wine tastings. Ferrarini et al. (2010) demonstrate that wine consumption is more often associated with a positive experience than a negative one. Silva et al. (2016) compare the emotions associated with wine with the emotions associated with beer and find that wine is associated with positive low-arousal emotional responses, such as calm and love, whereas beer is associated with higher-arousal emotions. Dhar & Wertenborch (2000) show that affective aspects can lead a person to buy a wine to taste it, even when he or she thinks it may be a bad choice. Barrena & Sánchez (2009) note that the emotions generated by wine are a good criterion for segmenting wine consumers. Olarte et al. (2017) establish that the negative emotions produced during a wine tasting are detrimental to the future purchase decision.

Watson et al. (1988) establish a two-dimensional model of emotions that can be measured by the Positi­ve and Negative Affect Schedule (PANAS). These au­thors consider that positive and negative affect have consistently emerged as two dominant and relatively independent dimensions. Pelegrín-Borondo et al. (2017b) find that the PANAS scale has three dimensions: positive, negative, and anxiety. Anxiety is related to attentiveness toward the object that is producing the emotion. These authors establish that the emotions in this dimension are a subset of the negative emotions of the scale by Watson et al. (1988), and they are considered to have a negative effect on purchase intent. In this regard, Venkatesh et al. (2003) find evidence of a dimension related to anxiety that could influence purchase intention with regard to new technology. As for the emotions produced by wine, Olarte et al. (2017) likewise find that the PANAS sale can be separated into three dimensions: positive emotions, negative emotions, and attentiveness. These authors establish that the latter dimension refers to the emotion of feeling attentive or watchful toward the new type of wine that the consumer is evaluating (in their case, a new natural sparkling red wine).

In light of the foregoing arguments, the following hypotheses were proposed: H1) The positive emotions produced by a wine offer positively affect the intention to purchase it; H2) the negative emotions produced by a wine offer negatively affect the intention to purchase it; H3) the attentiveness emotions produced by a wine offer negatively affect the intention to purchase it.

(ii) Influence of the emotions produced by the environment (the winery visit) on purchase intent. The literature has demonstrated the influence of the emotions generated in the environment where a product is sold on its purchase (Machleit & Eroglu, 2000). Bagozzi et al. (1999) observe that stores play background music and use other stimuli to generate emotions that will influence their customers’ purchase decisions. Penz & Hogg (2011) demonstrate that the sales environment influences the enthusiasm cus­tomers show toward products. Positive emotions produced in virtual environments on e-commerce sites have likewise been found to generate emotions that influence customers’ behavior. For example, Wang et al. (2011) show that the order, legibility, and simplicity of a website decrease the level of arousal, while aspects such as aesthetics increase it. They further observe that the level of arousal influences customers’ responses.

Regarding negative emotions, Kim & Lennon (2011) find that difficulty finding products in a store generates negative emotions in customers, which, in turn, negatively affect their image of the store. Mano (2004) demonstrates that certain environmental conditions may cause negative emotions ( e.g., sadness) that reinforce the desire to buy so as not to feel the negative emotion.

From a wine tourism perspective, Bruwer et al. (2013) observe that the pleasing physical and natural aspects of the landscapes around wineries provide wine tourists with hedonic experiences that are associated with positive emotions. Referring to the environment cultivated at wineries, Yuan et al. (2008) find that positive emotions generated by the services offered to visitors increase their satisfaction. This increased satisfaction then generates positive emotions that influence their purchase decision and future visits. Charters et al. (2009) demonstrate that emotions generated by satisfaction with the winery visit are capable of mitigating the negative emotions caused by feeling obliged to buy wine at the end of the tour. This is because satisfied wine tourists feel grateful for their experience at the winery and thus feel the need to buy wine or a souvenir (Kolyensikova & Dodd, 2008).

In light of the above, the following hypotheses are proposed: H4) the positive emotions produced by the winery visit positively affect the intention to purchase the wine on offer at the winery; H5) the negative emotions produced by the winery visit negatively affect the intention to purchase the wine on offer at the winery; H6) the attentiveness emotions produced by the winery visit negatively affect the intention to purchase it.

(iii) Influence of induced emotions (wine news) on purchase intent. Induced emotions refer to emotions that are not generated by the product or the environment and are independent of the subject (Pelegrín-Borondo et al., 2015). Emotions induced through a text, news item, or video, among other things, have been shown to influence behavior. Griskevicius et al. (2010) conduct a set of experiments in which participants are induced to feel the emotions of pride and contentment. They find that pride generates interest in products that are used publicly ( e.g., clothes that are worn in public) but does not increase the desire for products used at home. In contrast, contentment increases the desire for products used inside the house. Kahn & Isen (1993) demonstrate that when customers are induced to feel a positive emotion, it increases their positive perception of the product’s benefits.

Labroo & Patrick (2008) demonstrate the influence that induced states of mind have on consumers’ judgment and ability to process information. It has likewise been shown that inducing individuals to feel emotions can affect their attention span, reducing their cognitive capacity (Mano, 2004) and resulting in a more heuristic decision-making process (Kahn & Isen, 1993).

Some researchers have failed to find evidence that induced emotions influence behavior. Garg et al. (2005) show that customers induced to feel sadness are not less likely to maintain their original choice, even when it is more difficult to evaluate the alternatives.

Several authors have demonstrated the capacity of news to produce emotions in consumers ( e.g., Nie et al., 2015; Yu et al., 2015; Myrick & Wojdynski, 2016). Park (2015) finds that negative news generates negative emotions that influence people’s behavior. Wilkinson et al. (2005) show that fear of suffering an illness can result in acceptance of food items that might diminish the risk of the ailment. Thus, news about the health benefits of certain foods generates a tendency to consume them.

In the field of tourism, Labroo & Rucker (2010) induce negative emotions in participants by encouraging them to remember sad, angry, shameful, or anxious situations. They find that participants in a sad or angry emotional state are more likely to associate a ski resort with happiness than calmness, while participants in a state of anxiety or shame are more likely to associate it with calmness than happiness. Gnoth et al. (2000) demonstrate that emotions resulting from problems at work influence the decision to travel as a way of forgetting those problems (the work-related emotions are unrelated to the evaluated and chosen destinations).

Based on this literature review, the following hypo­theses were proposed: H7) the positive emotions pro­duced by a wine-related news item positively affect the intention to purchase the wine on offer at a winery; H8) the negative emotions produced by a wine-related news item negatively affect the intention to purchase the wine on offer at a winery; H9) the attentiveness emotions produced by a wine-related news item negatively affect the intention to purchase it.

The conceptual model shown in Figure 1 reflects the comparative influence of the positive and negative emotions produced by the wine offer vs. the winery visit vs. wine news on wine purchase intent among tourists to a winery.

The study was carried out at two wineries located in two distinct wine regions: La Rioja in Spain (Old World wine country) and Baja California in Mexico (New World wine country). A survey was conducted with a sample of 600 wine tourists (300 in each region). The data were analyzed using variance-based structural equation modelling (SEM) by means of partial least squares path modelling (PLS). G* Power software was used for the power analysis. This software uses post-hoc power analysis to compute achieved power based on the level of significance, sample size, and number of predictors. For the purposes of the study, the same type of wine was offered at both wineries, with the sole exception of the winery brand. To this end, the wine tourists were told that the wine they were being offered had been produced at the winery but did not sample it. They were then asked about the emotions generated by the wine (product), the visit (environment), and the wine news (induced).

To obtain a more diverse sample and mitigate the destination effect, two prestigious wineries were sampled in two different wine regions: Bodegas Franco Españolas, in La Rioja (Spain), and Bodegas de Santo Tomás, in Baja California (Mexico). The wineries share certain similarities. Bodegas Franco Españolas was founded in 1890 by Frederick Anglade Saurat, of Bordeaux, on the banks of the Ebro River in the city of Logroño. In 1983, it was acquired by the Elguizábal family. Over the years, it has increasingly internationalized, and today its wines are sold in 65 countries. It has moreover cultivated a strong wine tourism culture. In 2013, it was named Best Wine Tourism Destination in the area of Innovative Experiences by Great Wine Capitals (Bodegas Franco Españolas, 2017). Bodegas de Santo Tomás was founded in 1888 by Francisco Andonegui and Miguel Ormart in Santo Tomás Valley, south of Ensenada, along the region’s historical wine route (Quiñónez et al., 2012). Today the company, which has been under the management of the Pando Group since 1962, exports about 14% of its production. Its top export destinations are the USA, Canada, Spain, and China. Already one of the most visited wineries in the area, it is currently developing wine tourism projects to attract new customers (Cervantes, 2012). Thus, both companies are historical wineries that were founded at around the same time, are oriented toward wine tourism, and export their wines.

The same procedure was followed to gather the data in both cases: upon completing the tour, tourists were invited to take part in a survey. The interview was conducted in the winery itself. Once a participant had agreed to take part, he or she was shown the wine offer and the corresponding documentation. The same offer was used at both wineries, with the sole exception of the winery brand name.

The documentation included only the following information, intended to elicit an emotional response: VARIETAL COMPOSITION: Tempranillo 100%. TAS­TING NOTE: Red wine, cherry colored with a ruby edging. A nuanced nose of red fruits and herbs over a licorice bottom. ALCOHOL: 13% Vol. SUGGESTED SERVING TEMPERATURE: 14° to 16° C [57° to 61° F]. TIME IN BARREL: 12 months in American and French oak. PRICE: €8.90/bottle (La Rioja) or $268.00 Mexican pesos/bottle (Baja California).

At the end of a wine tour, tourists are often invited into the wine store, where the entire range of wines is available, including some they would have tasted as part of the tour and others they would not. As mentioned above, in this case, the participants were shown the wine offer, but did not taste the wine.

The participant was then read the following news item: According to a study published in the Annals of Internal Medicine, drinking one glass of wine with dinner may be ideal. Scientists selected 224 people who did not drink wine daily and divided them into three groups. The first group was told to drink one glass of red wine with dinner for two years; the second group was to drink one glass of white wine with dinner for two years; and the third group was told to drink mineral water. At the end of the study, the members of the first group had better cholesterol levels and decreased risk factors for diabetes. The participants who drank wine (red or white) had better triglyceride levels and better sleep quality than those that drank water.

Immediately after, they were asked a series of questions about the emotions produced by the wine offer, the winery visit, and the news item. There is, unfortunately, no agreement regarding how many emotions exist. The idea that there is a set of basic emotions possibly invariant across cultures has been a prevailing view (Prescott, 2017). To measure how strongly these emotions were felt, the PANAS scale was used (Watson et al., 1988), with scores ranging from 0 (not at all) to 10 (intensely). This scale includes ten positive emotions and ten negative emotions: interested, distressed, excited, upset, determined, guilty, scared, hostile, enthusiastic, proud, irritable, alert, ashamed, inspired, nervous, strong, attentive, jittery, active, and afraid. The scale was applied to each type of emotion ( i.e., "Think about how the wine offer makes you feel. Rate the degree to which you feel the following adjectives on a scale of 0 (not at all) to 10 (intensely)."). They were also asked about their intention to purchase the offered wine. This intention was measured using the scale developed by Venkatesh & Davis (2000) ( i.e., "If I could, I would try to buy the wine on offer"; "If I could, I anticipate I would buy the wine on offer") with an 11-point Likert scale. The survey ended with a few questions designed to deter­mine the participants’ socio-demographic classification. Subse­quently, three structural equation models were run linking the type of emotion to the tourists’ wine purchase intent: (i) the emotions produced by the wine and purchase intent; (ii) the emotions produced by the winery visit and purchase intent; and (iii) the emotions produced by the news item and purchase intent.

The sample obtained in La Rioja consisted of 300 validated surveys, and the socio-demographic classification was similar to that obtained in the same region by Molina et al. (2013), which consisted of 598 wine tourists surveyed in five Spanish wine regions. In that study 48.5% of the respondents were female, and 51.5% were male. In the present study, the respective percentages were 55.7% and 44.3%. In Molina Collado et al. (2013), 44% of those surveyed were aged 34 and under, while 85% were aged 44 and under. In the present study, 46.3% were aged 34 and under, and 80% were aged 44 and under. With regard to educational attainment, in Molina Collado et al. (2013), 64.2% of the respondents had a college degree, and 30.6% had a high school diploma. In the present study, these figures were 65.7% and 28.7%, respectively.

The sample for the survey conducted in Baja California also consisted of 300 valid participants. The age bracket was similar to that obtained in the same region by Orta et al. (2016), who sought to understand the importance of satisfaction in the wine tourism experience. In that study, 340 people were surveyed, with respondents aged 39 and under accounting for 50% of the total. In the current study, respondents aged 39 and under accounted for 66.3%. It should be noted that only limited information is available on the sample used in Orta et al. (2016), as the authors do not provide a breakdown of the sample by gender or educational attainment level. In the present study, 51.3% of the respondents were female and 48.7% male. Furthermore, 81.7% had a college degree, while 17.3% held a high school diploma.

SEM, specifically, the consistent partial least squares (PLSc) SEM technique, was used to assess the measurement model. PLSc is less sensitive than PLS to Type I and Type II errors and should be applied to models in which all the constructs are reflective (Dijkstra & Henseler, 2015), as in the present case. PLS tends to skew factor loadings upwards and underestimate regression coefficients (Gefen et al., 2011). PLS was chosen as it is less sensitive to the violation of assumptions of data normality (Chin, 1998a; Ram et al., 2014).

Exploratory factor analysis (maximum likelihood) was used to analyse the factors formed from the observable variables of the three scales (emotions produced by the wine on offer, emotions produced by the visit to the winery, and emotions produced by a news item on wine). Which allowed us to make a first exploration of how the theoretical framework is translated into the empirical data.

All three cases gave rise to the same factors comprising the same emotions. Together, these three factors explained 51.63% of the variance in the emotions produced by the wine, 45.25% in the emotions produced by the winery visit, and 46.94% in the emotions produced by a news item about wine. The first factor, called "positive emotions", included the following positive emotions from the Watson et al. (1988) scale: enthusiastic, proud, determined, excited, active, inspired, interested, and strong. The second factor, called "negative emotions", included the following negative emotions from the Watson et al. (1988) scale: scared, upset, hostile, ashamed, guilty, distressed, irritable, afraid, and nervous. The third factor, called "attentiveness", included the following emotions: attentive, alert, and jittery.

For all three scales, the Bartlett’s test of sphericity coefficient reflected a significance level less than 0.001 and showed good results for the KMO (0.92 for the emotions produced by wine, 0.91 for the emotions produced by the visit to the winery, and 0.90 for the emotions produced by the wine-related news item).