The impact of beer type, pizza spiciness and gender on match perceptions
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
Vol. 6 Nº 2 págs. 173-188. 2008 Special Issue – Número Especial www.pasosonline.org The impact of beer type, pizza spiciness and gender on match perceptions Robert J. Harrington i University of Arkansas (USA) Daniel C. Miszczakii University of Guelph (USA) Michael C. Ottenbacher iii San Diego State University (USA) Abstract: This exploratory study surveys preferences of participants towards pairing three categories of beer (lager, ale and stout) with a non-spicy and spicy pizza. The goals of this study are to determine the level of a ‘just right’ match of pizza style with each beer type, determine any differences by gender, and to explore if spice has an impact on participants’ beer selection and beer preference. Implications of this research apply to restaurateurs’ ability to appropriately cater their beer and pizza offerings in terms of menu design and pro-duct delivery. Keywords: Food and drink pairing; Spice; Beer; Pizza; Gender differences Resumen: Este estudio exploratorio examina las preferencias de los consumidores hacia la relación existente entre tres tipos de categorías de cerveza (cerveza dorada (lager), cerveza inglesa (ale) y cerveza de malta (stout)) con una pizza no-picante y picante. Los objetivos del estudio son determinar el nivel de ajuste entre el estilo de pizza con cada tipo de cervezas analizadas, y explorar las diferencias por género, así como, sí una pizza picante tiene o no impacto en la elección y preferencia del consumidor de cerveza. Las implicaciones de esta investigación se aplican a las capacidades de los restaurantes a la hora de esta- blecer un stock de cerveza y tipo de pizza, y como ello influye en la creación de experiencias, bebida versus comida, por parte del turista gastronómico. Keywords: Relación comida versus bebida; Picante; Pizza; Diferencias de género i • Robert J. Harrington is Associate Professor and Endowed Chair in Hospitality & Restaurant Management Pro- gram. University of Arkansas (USA). E-mail: rharring@uark.edu ii • Daniel C. Miszczak is a graduate student in the School of Hospitality & Tourism Management. University of Guelph (Canada). E-mail: dmiszcza@uoguelph.ca iii • Michael C. Ottenbacher is Associate Professor in the School of Hospitality & Tourism Management. San Diego State University (USA). E-mail: mottenba@sdsu.edu © PASOS. Revista de Turismo y Patrimonio Cultural. ISSN 1695-7121
174 The impact of beer type, pizza spiciness... Introduction culinary tourism product includes a con- scious pairing of food and drink experiences While one often thinks of food and wine with other travel activities (Billups, 2007). when considering pairing possibilities, re- These food and drink pairings create me- cent interest has been shown in the beer morable experiences when tied to the ga- and food pairing arena (Bellamy, 2005; stronomic identity of a region or locale. Shriver, 2006). Recent articles in journals The concept of gastronomic identity illu- and the popular press point to opportuni- strates the influences of the environment ties for restaurateurs as a method for in- (geography and climate) and culture (histo- creasing guest satisfaction and interest in ry and ethnic influences) on prevailing this area (Beaumont, 2006; Charters & taste components, textures and flavors in Pettigrew, 2005). The topic is contentious food and drink. This identity has great in that most literature written on the topic is subjective in nature and lacks empirical consequences for successful wine tourism, or systematic testing (e.g., Cummings, culinary tourism with the introduction of 2006; Fried, 1993; Oliver, 2003; Pearce, value-added features such as history, story- 2007). Therefore, a key purpose of this ex- telling and authenticity as well as creating ploratory study is to determine which style a synergistic relationship that maximizes of beer creates a perception of the best the level of “gastronomic satisfaction” for match when consumed with a spicy or non- consumers (Harrington, 2008). As with spicy food. Specifically, we consider three many types of beverage, popularity varies beer categories (lager, ale and stout) and substantially by culture whether the beve- the perceived level of match with a spicy rage of choice is coffee, tea, wine, beer or and non-spicy pizza. some other beverage. Traditions of beve- The topic of beer and pizza pairings is rage consumption with specific foods have important because restaurateurs that have become a key factor of distinguishing cui- a business focus on selling beer and pizza sine and tourist products. can use the knowledge to cater to custom- As suggested by American brewmaster er’s needs and wants (Popp, 2006; Stin- and author Garrett Oliver (Rubin, 2007), chfield, 2004). This information will help beer can be a good accompaniment to clas- restaurateurs and service staff in menu sic North American pairings such as pizza planning, inventory control and time man- and chicken wings, but it also provides agement through greater understanding the beer and pizza pairing combination. opportunities for matching a lot of other Thus, greater knowledge in this area will foods usually thought of as wine territory. aid in suggesting pairings for customers, The different varieties of beer on the mar- increasing revenues through up selling, ket have an abundance of flavors and aro- and ultimately enhancing the overall cus- mas that if experimented with can be tomer experience. paired many types of foods and styles of The primary research question is to de- cuisine. Much like wine styles, beer can termine: Does spiciness in pizza impact provide a vast array of tastes (Beaumont, taste preferences when tasted with three 2006; Kochak, 1999; Pearce, 2007). For different categories of beer? Secondary example, a combination of sweetness, sour- questions include: Does an individual’s ness, bitterness or smokiness to name only preference towards a particular beer affect a few tastes can all be identified in any their choice of a best match when consumed particular beer (e.g., Guinard, et al., 1998). with pizza? Does it change with the addi- Oliver suggests that beer has a much wider tion of spiciness? And, is there a difference range of flavors than wine and he is leading in these findings across genders? the charge in North America towards giv- ing beer what he calls its rightful place at Literature Review the dinner table (Oliver, 2003; Rubin, 2007). One of the hottest niches to emerge in Although the literature on beer and food the travel industry in the past five years is pairings is immature, there is a good deal culinary tourism. A part of a successful of information printed in newspaper ar- PASOS. Revista de Turismo y Patrimonio Cultural, 6(2). 2008 ISSN 1695-7121 Número Especial. Turismo Gastronómico y Enoturismo
R.J. Harrington; D.C. Miszczak and M.C. Ottenbacher 175 ticles, periodicals, industry magazines and chicken and the traditional hot dog (Stin- internet blogs where writers have provided chfield, 2004). Medium to moderately, full- valuable personal opinions about the pair- bodied beer are brands such as Heineken or ings of food with beer. For instance, Zac ales such as Alex Keith’s or Bass Ale. These (2007) suggests that beer can offer sweet- medium-bodied beers have been proposed ness, sourness, and tannin as well as in- as a terrific pairing with lobster, meat tense hop bitterness, smokiness and a roasts or stews as well as with Indian food range of additional flavors. Moreover, he (Stinchfield, 2004). suggests beer offers a wider range of body The last general category of beer is full- than that offered by wine ranging from bodied and includes brands such as Guin- viscous and flat to bone dry and efferves- ness or Bavaria Dark Reserve. These full- cent tastes when paired with food (Zac, bodied beers have been proposed to com- 2007). In February 2007, Ontario (Canada) pliment full flavored foods such as those Beer Stores began a promotion by provid- involving spicy tastes and sausages (Fried, ing food and beer suggestions to their cus- 1991). Contrary to Fried’s pairing sugges- tomers via a pamphlet describing food and tion of these full-bodied beers with spicy beer pairings. The marketing campaign foods, American brewmaster Garret Oliver uses a slogan stating “Did you know the suggests a hoppier Indian Pale Ale as a Beer Store is the only place in Ontario wonderful pairing for spicy foods such as you’ll find the largest selection of beer to Thai, Mexican or Indian foods (Oliver, match any dish?”(The Beer Store, 2007) In 2003). the United Kingdom, matching beer with With an increase in restaurateurs brew- fine food is already in vogue and moving ing their own beer on site, expectations of quickly into the North American market food pairings suggested by restaurateurs is (Cummings, 2006). Hence, it would behoove becoming evident (Kochak, 1999). The in- foodservice establishments of all types to crease in brew pubs has increased rapidly utilize beer and food pairings as a customer from only a few operations in the early enhancement proposition and take part in 1980’s to over 1200 in the US by 1997 and this global trend. the numbers of operations continue to While there is a lack of empirical re- climb. These micro brewers are a gateway search specifically about the pairing of beer to a new market of beer drinkers, creating and pizza, recent models of wine and food an opportunity to capitalize on this trend pairing decision-making provide indicators by catering beer to food. of important elements for creating an ideal Some literature exists to help under- match for beer and food. Important ele- stand perceptions that consumers have ments for pairing wine and food include about how wine and beer relate to food. As sourness, sweetness, body level, high levels suggested by Pettigrew (2005), the history of salt, bitterness, spiciness and efferves- of food and alcohol pairing is considered cence (Harrington, 2008). Many of these symbolic. Also, Pettigrew’s (2005) study factors apply to beer and pizza pairing de- explored a sample from Australia’s popula- cisions with body, bitterness and spiciness tion to identify perceptions about wine and being particularly relevant. beer in relation to food. The findings indi- While there is a wide array of beer cated that the perception of beer as a viable styles available, basic categories of beers pairing with food is weak. A key implica- that are clearly distinguishable by regular tion is that producers or servers of food beer consumers include lagers, ales and should align the consumption of their food stout beers (e.g., Rande & Luciani, 2001; products with the consumption of specific Schmid, 2004). These categories can be beer selections. And, with this identified divided up by the strengths and body of the importance, restaurateurs can cater to beer. In North America, the lighter beer their customer’s needs and wants through category (lagers) consists of beers such as effective pairing options and menu design if Coors and Budweiser. Writers have pro- they understand their customer’s percep- posed that lighter beers are excellent with tions and wants in terms of food and beer North American favorites such as fried pairing (Pettigrew, 2005). PASOS. Journal of Tourism and Cultural Heritage, 6(2). 2008 ISSN 1695-7121 Special Issue. Gastronomic and wine tourism
176 The impact of beer type, pizza spiciness... To stay profitable and to compete in the can cater their toppings to the food buying restaurant industry, every competitive edge habits of beer drinkers. can be beneficial. To assist in capitalizing In summary, the topic of beer and food on alcohol and food pairings, it has been is gaining momentum. The importance of shown that providing wine pairing recom- understanding beer and pizza pairings can mendations as well as wine tasting can aid in menu planning and design. With increase revenues for a restaurant (Blair, trends indicating that beer is gaining more et al., 2006). This field study demonstrates respect for its vast array of flavors, beer the potential for success by making restau- might very well earn its rightful place at rateurs aware of the relationship among the dinner table, particularly, in many restaurant staff alcohol and food recom- parts of North America. While beer and mendation training, pairing recommenda- food pairing appears to be a fruitful area tions to guests, and increased revenue. In for creating gastronomic satisfaction for the study it was demonstrated that con- many guests, empirical research in this sumer knowledge about wine and food is area is conspicuously absent from the lite- lacking and thus it is up to the servers and rature. The suggested pairings in the popu- bartenders to practice the idea of the food lar press have many contradictions. pairings. While the study by Blair et al. (2006) looked at wine and food recommen- Hypotheses dations, it stands to reason that the same relationship could be shown for beer and Most of the literature in beer and food food recommendations. Although North pairing is anecdotal in nature and has con- American consumers may recognize that flicting suggestions for beer styles with beer pairs nicely with wings, nachos and spicy foods. Based on a synthesis of the burgers, they may lack the initiative to available literature, we derive four hypo- explore beyond these North American tra- theses. ditional pairings without additional prod- For Hypothesis 1, we suggest that an ding. individual’s preference for a particular type The importance of this study relates to of beer when consumed without the addi- opportunities for pizza and other restaura- tion of food will be a strong indicator of teurs to differentiate themselves from the their best match choice when consumed competition. For instance, understanding with a non-spicy pizza sample. Our sup- the buying habits of consumers that pur- port for this relationship is derived from chase beer can help restaurateurs identify general rules of other food and beverage characteristics of their customers. Specifi- pairing suggestions (i.e. wine). For this cally, understanding these buying habits study, the non-spicy pizza sample will in- can help restaurateurs design an appropri- clude crust, an herbed tomato sauce and ate beer and pizza product mix. A study Italian-style cheeses (mozzarella and conducted by Friis and colleagues (2006) parmesan). These ingredients do not con- shows evidence that consumers who buy tain an excessive level of tastes and flavors wine at grocery stores are likely to buy that are likely to limit pairing choices (i.e. Mediterranean items. Other items bought sourness, sweetness, saltiness, bitterness by the wine buyer included fruits and vege- and spiciness) (e.g., Baldy, 2003; Harring- tables. The study revealed that wine buy- ton, 2008; Immer, 2002). Therefore, the ers bought healthier foods than beer buyers non-spicy pizza becomes a beer-friendly, (Friis, et al., 2006). Conversely, consumers blank canvas and participants will primari- that purchased beer bought food items that ly select the best match based on percep- were more basic or non-complex. Items tions of beer likeability. Formally stated: such as pork, sausages and cold cuts were H1: An individual’s preference for a partic- foods of preference for the beer buyer. This ular type of beer will heavily impact information can be useful in menu plan- their best match ning. Restaurateurs who focus on pizza and choice when tasted with a non-spicy piz- beer and who make the assumption that za sample. the customer base is mostly beer drinkers In contrast, the spicy pizza sample (with PASOS. Revista de Turismo y Patrimonio Cultural, 6(2). 2008 ISSN 1695-7121 Número Especial. Turismo Gastronómico y Enoturismo
R.J. Harrington; D.C. Miszczak and M.C. Ottenbacher 177 the addition of crushed red pepper) creates to balance flavors) (Websters Online, 2007), a food item that limits beer choices due to the number of IBUs is an important factor spicy characteristics and a greater sense of in the beer and food pairing process. intensity and persistency of flavors. This Bitterness has been proposed as an hypothesized relationship is supported in important issue in food and beverage the wine literature providing evidence of pairing in general; and, earlier studies in the limiting effects of food spiciness on be- the pairing process have suggested high verage choice and the desire to create and levels of bitterness create pairing problems. equal footing (i.e. flavor intensity and per- A cultural factor in this regard is a lack of sistency) between the food and beverage cultural affinity for bitterness in general in selection (e.g., Harrington, 2008; Jackson, the North American culture (Harrington, 2000; The Beer Store, 2007). Therefore, 2008). Hypothesis 2 states: While bitterness is an obvious issue to H2: The impact of an individual’s prefe- consider in beer and food pairing as a rence for a particular type of beer as the whole, earlier empirical studies have shown best match that females as a group have a larger choice will be reduced when tasted with percentage of ‘super-tasters’ and are more a spicy pizza sample. sensitive to bitterness compared to their Similarly to the arguments for Hypothe- male counterparts. A study of 400 women sis 2 above, Hypothesis 3 suggests partici- indicated about 25% fell into the ‘super- pant beer selections with the spicy pizza taster’ category. In the study, women who sample will change to more flavorful and were sensitive to sharp and bitter foods fuller-bodied styles than those selected limited their exposure to food items known with the non-spicy pizza. Here again, an to reduce cancer risk. The foods in this equal footing of flavor intensity, persistency category included many bitter vegetables and spiciness will generally create a sensa- (i.e. broccoli and Brussels sprouts), citrus tion of equal partnership between the food fruits (i.e. grapefruit), and other bitter item and beverage, raising the perception berries and roots (Gilbert, 2005). of match for the majority of individuals Because fuller-bodied beers will (Harrington, 2008). While it is unclear in generally be perceived as having bitterer the literature if the best relationship will flavor and because females on average be with a full-bodied beer (i.e. Fried, 1991) appear more sensitive to bitter flavor, we or a medium-bodied and fuller flavored hypothesize that (on average) women in beer (i.e. Oliver, 2003; Rubin, 2007), there this study will select lighter-bodied beers appears to be at least a general consensus than men when consumed with food. that spicier foods are a better match with Specifically - beers that are more flavorful (hoppier) and H4: Compared to males, females will prefer fuller-bodied than lighter beers such as lighter beers regardless of spice level in standard lagers. Formally stated: pizza samples. H3: The addition of moderate spiciness to the pizza sample will change the match Methodology & Research Design relationship to more flavorful, fuller- bodied beers. Extensive research has been performed An important issue in beer flavor is the using a quantitative approach of deviation- level of International Bitterness Units from-match or ‘just right’ scales in assess- (IBUs). The International Bitterness Units ing food, beverage or food and beverage scale provides a measure of the bitterness pairing characteristics (Harrington & of beer which is provided by the hops used Hammond, 2006; King & Cliff, 2004; during brewing. An IBU is one part per Shepherd, 1989). This study utilized a tast- million of Isohumulone; the higher the ing panel method to assess ‘just right’ number, the greater the bitterness. match levels for pizza and beer. A ‘just Although the bitter effect is generally less right’ scale provides a graphic description noticeable in beers with a high quantity of that can also be described as deviation- malt (a common ingredient in heavier beers from-match where participants rate the PASOS. Journal of Tourism and Cultural Heritage, 6(2). 2008 ISSN 1695-7121 Special Issue. Gastronomic and wine tourism
178 The impact of beer type, pizza spiciness... combination of food and drink for too little, of hot spice based on typical North Ameri- too much or just right sensations of match can standards for cuisine. in defined characteristics (e.g., overall pair- The 3 beers for this study were selected ing match, spiciness, etc.). to allow differentiation among three gener- To minimize threats to the validity of al styles: lager, ale and stout. The beers this study, we chose to focus on key pizza range in an ascending order in both body ingredients (crust, cheese, tomato sauce, style and hoppiness (bitterness). For the spicy and non-spicy versions), key beer tasting the IBUs for each beer were as fol- categories (lager, ale and stout), and key lows: Trailhead lager at 18, Wellington elements impacting level of match in the S.P.A (ale) rated at 23, and Imperial Stout food and wine literature (body, bitterness at 50. All beers in the study were from Wel- and spice). Therefore, this study adapts lington Brewery (located in Guelph, Ontar- King and Cliff’s (2004) methodology to per- io, Canada)1. form a similar taste experiment by gather- ing non-experts and have them sample 3 Tasting procedures different categories of beers to determine which beers are the best match with a spicy The tasting process of this research used and non-spicy pizza. sequential and mixed approaches to sam- pling the beer and pizza used in our expe- Profile of participants riment (Nygren, Gustafsson & Johansson, 2003). As described by Nygren et al., a se- The participants consisted of graduate quential approach is defined as an assess- students, faculty and undergraduate stu- ment when the beer or food is tasted sepa- dents from a large university in North rately. A mixed approach is defined as an America. The age of individuals ranged assessment with the beer and food tasted from 20 to 70 years and included 10 men simultaneously (one after another). and 24 women. A preliminary assessment The study was conducted in one day and of the participants ensured all had con- consisted of three stages. The first two sumed a beer of some type within the last stages of the experiment used a sequential two years, enjoys eating pizza, could identi- approach while the last stage used a mixed fy bitterness levels, and could differentiate approach. During each tasting, dry bread between the two spice levels. The partici- was provided to clear palates and to elimi- pants ranged from novice to experienced nate a carryover effect of taste as suggested beer drinkers and pizza consumers. by Fried (1993). Before the tasting oc- curred, the participants were not told the Pizza, spiciness and beer type definitions. types of beer used or if the pizza samples The study used two pizza samples: one were spicy or not. They were instructed to non-spicy and one spicy. To minimize ef- refrain from speaking to one another, and fects of other potentially competing food to use the water and soda crackers along elements, both pizzas utilized the same with five minute breaks to minimize a car- ingredients, measured amounts, and cook- ryover effect between tasting1. ing method (with the exception of spice Stage 1. Four small 30 ml samples of added to the spicy version). Specifically, each beer in red plastic cups were given to the non-spicy sample was prepared with a the participants. They were then asked to pre-made thin crust, herb tomato-based sample each beer individually in a pre- pizza sauce, and shredded pizza cheeses. determined order and answer four ques- The spicy version used the same ingre- tions. The first two questions were to rank dients with the addition of 5 grams of the beers according to their bitterness and crushed red pepper evenly distributed un- body. The third question asked was to iden- der the cheese layer to ensure an equal tify preference of the four beers and the last amount of hot spice in each sample and to question was designed to identify if the avoid cueing the participants of the hot participants could identify beer sample four spice addition. Thus, for this study, spice as being identical to beer sample one. The or spiciness is defined as a moderate level participants did not know prior to the expe- PASOS. Revista de Turismo y Patrimonio Cultural, 6(2). 2008 ISSN 1695-7121 Número Especial. Turismo Gastronómico y Enoturismo
R.J. Harrington; D.C. Miszczak and M.C. Ottenbacher 179 riment that a duplicate beer was the fourth ticipants’ perceptions of match and general sample. beer and pizza relationships. These ana- Stage 2. A 2.5 cm by 2.5 cm square piece lyses are integrated into the results discus- of non-spicy sample was distributed to each sions for each area of interest. of the participants. After the tasting of the sample, they were asked to rank the sam- Tasting Ability and Preferences ple according to the level of spice. Next, a 2.5 cm by 2.5 cm square piece spicy pizza Beer bitterness perception. A first as- was distributed after participants cleansed sessment was to determine participants’ their palates with water and crackers and ability to identify bitterness level. All par- the same question was asked. After the two ticipants used in this study could identify samples were tasted, participants were the progression of bitterness level from the asked to record their preference between least to most bitter beer. On a 5-point scale the two pizza samples. Information from (1 = not bitter, 5 = extremely bitter), the these answers provided data on whether or mean bitterness level for each beer type not the sample population could identify was as follows: 2.16 lager, 2.99 ale, and spice and also, whether there was a prefe- 4.22 stout. While the ability to distinguish rence for the spicy or non-spicy pizza sam- ascending bitterness levels was true for ple. both men and women, the female tasters Stage 3. This last stage involved a displayed the highest mean at 4.49 for the mixed tasting procedure. All three beers stout suggesting that they had a stronger were poured and then the non-spicy pizza sensation than their male counterparts sample was distributed. The participants (mean = 3.57). tasted beer sample one with the non-spicy Beer body perception. Similar to the pizza, then used a ranking style question to means described for the bitterness variable, rate their perception of match. The same the mean body level perception identified procedure was done for the remaining two by the participants increased from the least sequential beers and non-spicy pizza tast- to most full bodied beer (mean = 2.01 lager, ing. After all three beer samples were 2.33 ale, and 3.80 stout [1 = light bodied, 5 tasted with the non-spicy pizza, the partic- = full bodied]). This indicated participants ipants were asked to circle their best match felt the lager was the least full bodied beer of the sample beers with the pizza sample. followed by the ale then stout as being ful- After this tasting was complete, the beer ler bodied. No differences were evident samples were re-poured and the spicy pizza when the mean was analyzed specifically was tasted using the same tasting sequence by gender. Lastly, the fourth beer which as with the non-spicy pizza. was the same as the first sample proved to be identifiable as being of the same body Analysis and Results intensity among the participants (mean = 2.01 for sample 1 and 1.99 for sample 4) The analysis of tasting results utilizes t- and no differences in the mean were no- tests to compare any significant differences ticed when analyzed specifically by gender. between groups and graphic ‘just right’ Participant beer preference. The partic- plots based on mean values of participants’ ipant response indicated a 64.7 % prefe- perceptions. Specifically, multiple t–tests rence for the lager followed by the ale at were conducted to assess significant differ- 23.5 %. Not surprisingly, the least pre- ences among beer choices with non-spicy ferred beer was the stout with only 11.8 % and spicy pizza samples and to determine if selecting it as the preferred beer. Females significant differences exist by gender. overwhelming (70.8 %) preferred the lager Two-tailed p–values are reported for the compared to about 50% of males. Converse- analysis conducted by SPSS v15 software. ly, 20% of males preferred the stout com- Details on the ‘just right’ plotting process pared to only 8.3 % of females. are described in the sections that follow. Participant ability to identify identical Thus, this study used ‘just right’ graphic beer samples. To test participants’ ability plots and 5-point Likert scales assess par- to distinguish identical beer samples in a PASOS. Journal of Tourism and Cultural Heritage, 6(2). 2008 ISSN 1695-7121 Special Issue. Gastronomic and wine tourism
180 The impact of beer type, pizza spiciness... blind tasting, two of the four beer samples pared to the non-spicy pizza sample and (both lagers) were the same and partici- provides initial support for H2 and H3 indi- pants were asked to determine the two cating a shift to more flavorful and fuller samples that matched. Overall, 82.4 % of bodied beers to match the fuller-flavored, the participants were able to identify to two spicy pizza. An increase in the preference lager samples. The remaining 17.6 % chose for stout was evident (32.4% with the spicy ale as match for the lager sample. pizza vs. 14.7% with the non-spicy pizza). Perception of non-spicy sample & spicy The mean response by male participants sample. Results indicate participants demonstrated the ale was liked more when could clearly distinguish between the non- tasted with a spicy sample (70% like spicy and spicy pizza samples. On the 5- slightly/like extremely vs. 30% for non- point scale (1 = not spicy and 5 = extremely spicy) followed by a liking of stout beer spicy), the participants’ ratings averaged tasted with a spicy pizza (20% like ex- 1.20 for the non-spicy pizza sample (for tremely vs. 0% for non-spicy). level of spiciness). The spicy sample’s mean The standard deviations of the partici- rating was 3.84 indicating an overall mod- pants indicated less of a spread in response erately-high level of spiciness. to pairing the lager and ale with the non- Participant pizza preference. The partic- spicy pizza. This tighter dispersion indi- ipants indicated that 58.8 % preferred the cated a closer range for the lager and ale spicy pizza sample, while 41.2 % preferred beer categories with no participant select- the non-spicy sample. What is interesting ing the category of “disliking extremely” is that 80% of males preferred the spicy the lager or ale with the non-spicy pizza sample as opposed to a 50/50 split between sample. In contrast, the standard devia- female preferences for the two samples. tion and range increased for the stout paired with the non-spicy pizza sample, Non-spicy/spicy sample perception with indicating responses fell across the spec- beer samples 1, 2 & 3 trum of possible pairing perceptions (rang- ing from “disliking extremely” to “liking The mean value for level of match across extremely). This interpretation was the all participants was the highest when the same when each gender was analyzed. non-spicy sample was tasted with the lager (mean = 4.03 on a 5-point scale) with a Beer preference and correlation with pizza score range of 2 to 5. This finding indicates match selections the lager was the preferred sample with most respondents (79.5%). The lager sam- The beer preference of the participants ple ranged from “liked slightly” (58.8%) to was positively correlated with the non-spicy “liked extremely” (23.5%) when paired with best match (r = .522, p < .01). This indi- the non-spicy pizza. cates that there is a high correlation of a The ale had the second highest mean participant’s preference towards a beer and value with the non-spicy sample (mean = their decision of a best match with the non- 3.68) but most participants (82.4%) selected spicy pizza sample. While this correlation match levels ranging from “dislike slightly” does not prove cause and effect, it does (11.8%), “neither liked nor disliked” support the idea that initial beer prefe- (35.3%), to “like slightly” (35.3%) catego- rences will impact best match selections for ries. The stout did not receive a positive the non-spicy pizza stated in H1. Further, perception by most participants when the correlation between the initial beer paired with the non-spicy pizza sample. For preference dropped substantially and to a this pairing, 64.7% selected the “dislike (statistically) non-significant level when a extremely” and 20.6% selected “dislike spiciness element was added to the pizza slightly. Only 14.7% selected “like slightly” sample (r = .321). or higher. To further evaluate this relationship, we The results of the spicy pizza sample coded participant selections on a 5-point pairing with the 3 beer samples demon- scale ranging from -2 to +2. These scores strated quite different results when com- were based on whether the participant se- PASOS. Revista de Turismo y Patrimonio Cultural, 6(2). 2008 ISSN 1695-7121 Número Especial. Turismo Gastronómico y Enoturismo
R.J. Harrington; D.C. Miszczak and M.C. Ottenbacher 181 lected a beer that was lighter in body than their original preferred beer (i.e. – 1 or – 2), The findings of this process indicated the same as their original preferred beer the majority of participants (67.6%) chose (i.e. 0), or selected a beer that was heavier the beer they preferred the most at the than their original beer (i.e. +1 or +2). For beginning of the experiment when tasted example, if the participant selected a lager with the non-spicy pizza. For instance, if a as their preferred beer choice prior to eat- person chose a lager as their favorite beer, ing pizza and selected a stout when con- they chose it as their best match with the sumed with pizza, their score would be non-spicy sample. Also, 14.7% selected a coded as +2 as stout is two steps up in both lighter beer and 14.7% selected a heavier body and bitterness. These frequencies are beer than their initial preferred beer with summarized in Tables 1 and 2. the non-spicy pizza sample. Cumulative Per- FrequencyPercent Valid Percent cent Valid -1.00 5 14.7 14.7 14.7 .00 23 67.6 67.6 82.4 1.00 5 14.7 14.7 97.1 2.00 1 2.9 2.9 100.0 Total 34 100.0 100.0 Table 1. Preferred Beer & Non-Spicy Best Match Cumulative Per- Frequency Percent Valid Percent cent Valid -2.00 1 2.9 2.9 2.9 -1.00 12 35.3 35.3 38.2 .00 15 44.1 44.1 82.4 1.00 6 17.6 17.6 100.0 Total 34 100.0 100.0 Table 2. Preferred Beer & Spicy Best Match 0 = match with preferred beer -1 or -2 = preferred beer lighter than preferred beer +1or +2 = preferred heavier beer than preferred beer Pairings Mean N Std. Std. t Sig. Dev. Error (2- tailed) Pair 1 NSPB1 4.03 34 .717 .123 .183 .856 SPB1 4.00 34 .816 .140 Pair 2 NSPB2 3.59 34 .925 .159 - .008 SPB2 4.00 34 1.044 .179 2.802 Pair 3 NSPB3 1.68 34 1.147 .197 - .004 SPB3 2.44 34 1.481 .254 3.059 Code Note: NSPB1 = non-spicy pizza with lager, SPB1 = spicy pizza with lager NSPB2 = non-spicy pizza with ale, SPB2 = spicy pizza with ale NSPB3 = non-spicy pizza with stout, SPB3 = spicy pizza with stout Judgments (shown as means) were made on 5-point scales (1 = dislike extremely, 3 = neither like nor dislike, 5 = like extremely). Table 3. Paired Samples Statistics and Test PASOS. Journal of Tourism and Cultural Heritage, 6(2). 2008 ISSN 1695-7121 Special Issue. Gastronomic and wine tourism
182 The impact of beer type, pizza spiciness... In contrast, participants’ original beer provide a refreshing sensation in both spicy preference changed when tasted with the and non-spicy situations, the sensation of spicy pizza. Only 44.1% participants picked match is improved for beers with higher their original beer choice as a best match bitterness and fuller body when consumed with the spicy sample. With the spicy pizza, with moderately spicy foods in general. 2.9 % preferred the lager if stout was the Second, these positive sensations, for many original beer choice, 35.3% selected ale or individuals, transcend their preferred ini- lager (depending on whether the original tial beer selections (when food was not part beer choice was stout or ale), and 17.6 % of the equation) supporting H2. preferred the stout or ale if they had picked a lighter beer as their original favorite beer Just right triangles (i.e. ale or lager). Thus, this finding pro- vides evidence of the impact of spice on In product testing, evaluations may be beer and food match perceptions, support- gathered on numerous product characteris- ing H2 where many participants moved tics, e.g. flavors, ingredients, aromas etc., away from their initial beer preference and using ‘just right’ scales (Market Facts, moved toward more flavorful, fuller-bodied 2007). In this research, the ‘just right’ beers when spiciness was included (sup- evaluation is about participants’ perception porting H3) . of the match of the sample pizza with each of the three beers. This process is used frequently in prod- Tests for significant differences in per- uct testing and allows evaluations to be ceived beer match with non-spicy and spicy gathered on a variety of product characte- pizza ristics. Data gathered in this process typi- cally divides responses into three catego- To consider whether changes existed in ries, for instance, ranging from ‘too little’, perceived match levels with beer samples ‘just right’ and ‘too much’ (Market Facts, and non-spicy or spicy pizza, we conducted 2006). For our purposes, we divided the three t-tests comparing changes in the responses into categories by the pizza sam- mean value of match for each beer type (B1 ple (non-spicy = Fig. 1 or spicy = Fig 2) and = lager, B2 = ale, B3 = stout) with non- whether consumption with the beer sample spicy pizza (NSP) and spicy pizza (SP). created a “negative” impact (when tasted Results descriptive statistics are shown in together), had “little” impact, or a “positive” Table 3 for all participants in the study (N (liked extremely) effect. = 34). To graphically display the results of this Based on these three tests, we deter- study, we decided to use a triangle plot mined two significant differences. When (Figures 1 & 2) with each side representing lager was tasted with the non-spicy pizza an axis relating to the three response or the spicy pizza, there was no significant groupings (negative, little impact, and posi- difference in participants’ perceived level of tive effects). Each triangle side (axis) has a match. For the pairing of ale, a significant 0 to 100% scale for plotting the percentage increase was shown for level of match with of participants concurring with the catego- the spicy pizza compared to the non-spicy ry. In this process, we coded responses perception (mean = 4.00 vs. 3.59, p = .008). from our 5-point scale as follows: dislike Similarly, for the stout pairing, a signifi- extremely and dislike slightly = negative cant increase was shown for level of match effect, neither like or dislike and like with the spicy pizza compared to the non- slightly = little impact, and like extremely spicy (mean = 2.44 vs. 1.68, p = .008). = positive effect. The findings shown in the change in After the location is determined for all beer preference section and the t-test re- three triangle sides, the scales can be sults indicate an important impact of spici- summarized (and plotted) to one point in ness when paired with pizza and support the triangle that depicts the impact rela- H2 and H3. Specifically, while lighter and tionship of pizza type with the beer type less bitter beers such as lagers are likely to across all respondents. This approach PASOS. Revista de Turismo y Patrimonio Cultural, 6(2). 2008 ISSN 1695-7121 Número Especial. Turismo Gastronómico y Enoturismo
R.J. Harrington; D.C. Miszczak and M.C. Ottenbacher 183 creates a balanced and graphic depiction of the impact across the sample of partici- 100 0 pants and seems particularly valuable giv- Positive en substantial individual differences in Impact 25 perceptions of match when food and drink 75 is consumed together. For both figures, the left side of the 50 triangle is the “positive” impact percentage, 50 the right side the “little impact” percen- tage, and the bottom side the “negative impact” percentage. The point of intersec- 75 Little tion of all three points displays whether or 25 6 Impact not the effects of pizza type and beer cha- racteristics has the most positive effect 100 when consumed together. For instance, the 0 point of intersection at the highest point of the triangle (towards the top) indicates the 100 75 50 25 0 participants perceived the beer sample as Negative Impact with Non-Spicy having the greatest positive combination Sample (pizza and beer) in this study. If the point of intersection is shifted further downward 6 = Lager = Ale = Stout in the triangle, it indicates (overall) the Figure 1. Just right plot based on match of lager, participants believed the effect to be too ale and stout paired with the non-spicy pizza sam- little than that desired to create a great ple (N = 34 All participants) relationship. If the point of intersection is shifted to the left side of the triangle, a larger percentage perceived the interaction 100 0 between the beer and pizza to have a nega- tive sensory effect. Positive Thus, as shown in Fig. 1 and Fig. 2, the 25 single point in the triangle plot depicts all Impact 75 three ‘just right’ match category percentag- 50 es for the non-spicy (Fig. 1) and spicy (Fig. 50 2) pizza tasted with the three different beers. This summarizes very simply in pic- ture form all the evaluations gathered 75 Little (Market Facts, 2007). 25 6 Impact Non-spicy pizza sample and 3 beer samples. As shown in Figure 1, the lager sample is highest toward the top of the 0 triangle but with the greatest percentage of the participants indicating the lager had a 100 75 50 25 0 relatively neutral to slight positive Negative Impact with relationship to the non-spicy pizza sample. Spicy Sample The ale had little impact when tasted with the non-spicy sample but fewer participants indicated it had positive 6 = Lager = Ale = Stout effects and a larger number perceived negative impacts. For the vast majority, the Figure 2. Just right plot based on match of lager, stout caused a negative impact when tried ale and stout paired with the spicy pizza sample with the non-spicy sample and a minority (N = 34 All participants) indicated little or positive effects. PASOS. Journal of Tourism and Cultural Heritage, 6(2). 2008 ISSN 1695-7121 Special Issue. Gastronomic and wine tourism
184 The impact of beer type, pizza spiciness... In this study, participants were asked to Second was the ale at 29.4 %, while the select the beer sample that they believed stout was preferred the least at 5.9 %. No- provided the best match with each pizza ticeably, males perceived their best match sample. For the non-spicy sample, best of lager paired with the non-spicy pizza at match selections mirrored the results of our 90 %, ale at 0% and stout at 10%. The fe- just right plots with 64.7% selecting the males were split between the lager (54.2%) lager, 29.4% selecting the ale, and 5.9% and ale (41.7%) paired as a best match with selecting the stout as the best match with the non-spicy pizza; stout at 4.2%. the non-spicy pizza. The participant overall best match of Spicy pizza sample and 3 beer samples. the three beers paired with the spicy pizza As shown in Figure 2, in contrast to the sample was the ale at 52 %. Secondly, the non-spicy pizza sample, the ale sample is lager was preferred by 38.2 % of partici- highest toward the top of the triangle and was rated higher overall than the lager pants while the stout once again was pre- paired with the non-spicy pizza. This find- ferred the least at 8.8 %. The female best ing is the result of more participants indi- match moved in the direction of ale as the cating positive effects and less indicating preferred choice, preferring the lager little impact of the ale. Also, results show a (37.5%) and ale (58.3%) with the spicy similar percentage of participants indicat- sample while just 4.2% enjoyed the stout as ing the ale (14.7%) and lager (8.8%) had a the best pairing. Conversely, the males negative sensory relationship to the spicy preferred the lager and ale equally at 40% pizza sample. Overall, the lager main- each followed by the stout at 20%. Thus, tained the same relative positive in the when spice was introduced, males taste triangle whether consumed with the non- perceptions changed (90% [lager], 0% [ale], spicy or spicy pizza. 10% [stout] for non-spicy vs. 40% [lager], The stout’s relationship improved sub- 40% [ale], 20% [stout] for spicy). Female stantially but was still the least favorite tastes were more consistent for preferred overall. With a spicy pizza, participants beers with both non-spicy and spicy but found the stout to have less negative effects with similar trends (54.2% [lager], 41.7% and more neutral to positive effects com- [ale], and 4.2% [stout] for non-spicy vs. pared to perceptions with a non-spicy pizza. 37.5% [lager], 58.3% [ale], 4.2% [stout] for As with the non-spicy sample, partici- spicy). pants were asked to select the beer sample Results of t-tests for significant differ- that they believed the best match with spi- cy pizza sample. For the spicy sample, best ences across genders are shown in Table 4. match selections reflected the just right Statistically, only one difference is present; plots with 38.2% selecting the lager, 52.9% the mean level of match between the non- selecting the ale, and 8.8% selecting the spicy pizza and ale was much lower for stout as the best match with the spicy piz- males than that perceived by females. za. The means in Table 4 and the lack of The graphic summaries shown in Fig- significant differences provide marginal to ures 1 and 2 provide additional support for no support for Hypothesis 4. In general, Hypothesis 3. It clearly displays the gen- females rated the level of match higher for eral choice of a lighter beer selection to medium and full-bodied beers (ale and balance the less flavorful, non-spicy pizza stout) than did their male counterparts. sample and the general choice of fuller- And, while the earlier tests for bitterness flavored choices to balance the greater in- levels supported the idea that females may tensity of the spicy pizza sample. be more sensitive to bitterness, the means in this study do not provide evidence that Tests for differences by gender this sensitivity has a negative impact on female perceptions match between food The participant overall best match of items and beer style. the three beers paired with the non-spicy pizza sample was the lager at 64.7 %. PASOS. Revista de Turismo y Patrimonio Cultural, 6(2). 2008 ISSN 1695-7121 Número Especial. Turismo Gastronómico y Enoturismo
R.J. Harrington; D.C. Miszczak and M.C. Ottenbacher 185 Mean N Std. Std. t Sig. (2- In this exploratory study, three Dev. Error (df = of our hypotheses received support. tailed)a 32) An individual’s preference for a par- NSPB1 4.30 10 .675 .213 ticular type of beer had a large im- 3.92 (M) .717 .146 1.48 .156 pact on their beer selection with 24 beer-friendly food such as non-spicy (F) pizza. While beer preference was NSPB2 3.00 10 .816 .258 also an important factor for the selec- 3.83 (M) .868 .177 -2.66 .016 tion with spicy pizza, the relationship 24 was greatly reduced with individuals (F) selecting fuller-flavored, fuller-bodied NSPB3 1.90 10 1.197 .379 beers to match flavor intensity and 1.58 (M) 1.139 .232 .713 .486 persistency of the pizza. 24 Therefore, spiciness in food has an (F) important impact on perceived level SPB1 3.70 10 1.337 .423 of match and ultimately customer 4.13 (M) .448 .092 -.982 .350 satisfaction. 24 While a greater percentage of (F) females selected a lighter beer as SPB2 3.60 10 1.174 .371 their initial preference compared to 4.17 (M) .963 .197 - .198 the males, we saw the greatest 24 1.349 (F) change in beer selection for females SPB3 2.40 10 1.713 .542 based on perceived level of match 2.46 (M) 1.414 .289 -.095 .926 when spiciness was introduced. 24 Therefore, an interesting finding is (F) that, while females seem to prefer to Code Note: consume lighter beers as a beverage NSPB1 = non-spicy pizza with lager, SPB1 = spicy pizza by itself, they seemed to perceive the with lager value of matching intensity levels NSPB2 = non-spicy pizza with ale, SPB2 = spicy pizza based on food selection. This pro- with ale vides an interesting conundrum for NSPB3 = non-spicy pizza with stout, SPB3 = spicy pizza practitioners in the field. with stout Research implications. Although Judgments (shown as means) were made on 5-point scales exploratory in nature, the study (1 = dislike extremely, 3 = neither like nor dislike, 5 = yields valuable information for the like extremely). general public and restaurateurs that M = male participants, F = female participants have a beverage focus of beer on their a. Equal variances not assumed. menus. A key implication for restau- Table 4. Best Match Selections for Male and Female rants is the potential impact on guest Participants satisfaction and return business based on superior food and drink Discussion & Conclusion experiences. To achieve these superior experiences, restaurateurs need server While many of us in the Western culture training programs and communication me- immediately think of wine as the key pair- thods that save time and increase sugges- ing beverage with food, a global trend has tions of pairings for customers. Training been for increased pairing with other beve- methods should provide a thought process rages and beer in particular. For foodser- for servers to suggestive sell and communi- vice professionals to take advantage of this cation tools should be developed recom- trend, they need to clearly understand the mend appropriate and interesting beer and elements in beer and impact level of food choices. For instance, an implication matches with food as well as consumer from this study for the restaurateur is that behaviors in selecting and consuming beer if the customer orders a lager and then with food. orders a spicy type pizza the server could PASOS. Journal of Tourism and Cultural Heritage, 6(2). 2008 ISSN 1695-7121 Special Issue. Gastronomic and wine tourism
186 The impact of beer type, pizza spiciness... suggest a fuller bodied beer to compliment tomer orders a lager beer then pairings can that pizza or have it described on the be suggested using non-spicy type pizzas. menu; thus, enhancing the customers ‘just However, when spice is introduced, other right’ beverage and food pairing. appropriate selections should be consi- While the findings provided a limited dered. number of significant differences across In conclusion, when it comes down to genders, the practical implications provide identifying a ‘just right’ match between an understanding of different tastes exhi- different categories of beer and food, best bited by gender, which can also add to in- match selections are driven by both indi- ternal marketing campaigns. Knowing that vidual preferences and food and drink cha- females enjoy a lighter bodied beer such as racteristics. Only through experimentation lager and ale with non-spicy pizza, staff and trying new combinations can a ‘just and restaurateurs can exploit this combina- right’ match be personally accomplished. In tion. But, staff and restaurateurs should the North America foodservice culture, not be shy about coaxing female customers many restaurateurs operate under the as- to try new food and beer combinations that sumption that the customers with ‘order will increase a match sensations and ulti- what they like’. While this is probably mately customer satisfaction with the din- true, restaurateurs who create operational ing experience. The same can be applied to methods that simultaneously allow cus- males with their preference of ales with a tomer to stick with what they know (with- spicy pizza. Promotions in this regard can out making them feel unsophisticated) and take the form of direct marketing to female recommend pairings that enhance the din- or male customers advertising appropriate ing experience are those who are more like- combinations based on food elements and ly to prosper in this increasingly competi- gender preferences. This approach can as- tive and food-experience conscious envi- sist in differentiating the foodservice opera- ronment. The information in this study tion from competitors and allow targeting a can be used as a vehicle for servers and certain product to a target market to ex- restaurateurs in up-selling, inventory con- ploit this marketing niche. trol, promotions, menu design, and (most The results from this study can also aid importantly) to suggest and engage the restaurateurs in menu design. If spice is an customer in a dialogue of determining a element of pizzas (or other food items) on a possible ‘just right’ match beer type and menu, then an inventory of different ales food. and stouts is recommended. If non-spicy Study limitations. The goal of this study foods are part of the menu mix, a diverse was to determine if spice had an impact assortment of lagers should be made avail- when tasted with different beers and to able. Knowing this information can help provide empirical information to the global the restaurateur in controlling and design- trend of food and drink pairing. While this ing their beverage and product mix. Of study provides some interesting results, the course care should be taken in regard to study has several limitations. First, the spiciness; in this study, the level of per- study was conducted in a North American ceived spice was at a moderate to mod- setting and any generalization of percep- erately-high level. Highly spicy foods may tions of match between food and drink are have other effects with beer style. Similar- likely to be limited to similar settings in ly, to the limiting effect of hot and spicy North America. Second, the study was of a foods with wine, these foods may require a limited size (N = 34) and future research refreshing and lighter beer style to cleanse should replicate and extend this study in the palate and prepare the diner for the new situations, locations, food types and next bite (Harrington, 2008). beer types. Third, level of spiciness was Finally, as shown through the use of the controlled at a moderate level. Any genera- ‘just right’ triangles, participants’ initial lizability to high levels of hot spice or other preference to a beer changed when spice types of spiciness (i.e. black pepper or was introduced. For the restaurateur, this sweet spices [cinnamon, nutmeg, carda- is valuable information because if the cus- mom, etc.]) is limited. Finally, the partici- PASOS. Revista de Turismo y Patrimonio Cultural, 6(2). 2008 ISSN 1695-7121 Número Especial. Turismo Gastronómico y Enoturismo
R.J. Harrington; D.C. Miszczak and M.C. Ottenbacher 187 pants in this study were volunteers and 22(1): 117. thus are a convenience sample. While we Friedrick, J. preformed several tests on tasting abilities 2006. “Get ideas brewing for cheese & beer (bitter, body, spice, etc.), issues of internal pairings”. Gourmet News Periodical, validity could be raised due to the non- June: 12-13. random nature of the sample used in this Friis, K., Gronbaek, M., Johansen, D., Sko- study. venborg, E. 2006. “Food buying habits of people who References buy wine or beer: Cross sectional study”. BMJ Online: pp. 1-4. Baldy, M.W. Gilbert, D. 2003. The university wine course, 3rd ed. 2005. “ 'Super-tasters' may avoid tart vege- The Wine Appreciation Guild, San tables, fruits that contain cancer pre- Francisco, CA. ventive compounds, says U-M research- Beaumont, S. er”. Online document at 2006. “Cocktailians & restaurateurs raise http://www.umich.edu/~urecord/9697/Fe the bar with beer & food pairings”. Na- b18_97/artcl03.htm tion’s Guinard, J-X., Yip, D., Cubero, E., Mazzuc- Restaurant News, 40(37): 1- 3. chelli, R. Bellamy, G. 1998. “Quality ratings by experts and rela- 2005. “Beer-food pairings brew up interest”. tion with descriptive analysis ratings: A Restaurant Hospitality Journal, 89(7): case study with beer”. Food Quality and 98. Preference, 10(1): 59-67. Billups, A. Harrington, R.J. 2007, November 8. “Have food, will travel”. 2008. Food and wine pairing: A sensory Washington Times, On-line document: experience. John Wiley & Sons, New http://www.washingtontimes.com/apps/p York. bcs.dll/article?AID=/20071108/NATION/ Harrington, R.J. & Hammond, R. 11108 0101/1002 2006. “Body deviation-from-match: The yin Blair, E., Cordua, G., Geiger, S., Payne, C., and yang of wine and food pairing?” Wansink, B. Journal of Culinary Science & Technolo- 2006. “Wine promotions in restaurants”. gy, 5(1): 51-69 . Cornell Hotel & Restaurant Quarterly, Immer, A. 47: 327-336. 2002. Great tastes made simple: Extraordi- Charters, S., Pettigrew, S. nary food and wine pairing for every pa- 2006. “Consumers expectations of food and late. Broadway Books, New York. alcohol pairing”. British Food Journal, Jackson, R.S. 108(3): 169-180. 2000. Wine science: Principles, practice, Cliff, M., King, M. perception, 2nd ed. Academic Press, 2004. “Evaluation of ideal wine and cheese New York. pairs using a deviation-from-ideal scale Kochak, J. with food and wine experts”. Journal of 1999. “Brew plate special”. Restaurant Food Quality, 28: 245-56. Business Periodical, 98(15): 41-48. Cummings, M. Market Facts. 2006, May 19. “Matching international 2007. “Triangle plots: Graphic display of beers with meals: A trend already in vo- ‘just right’ scale data”. Research on Re- gue in the U.K.” The Post Newspaper, B: search, 56, 1-6. 1-3. Market Facts of Canada, Toronto, Can- Fried, E. ada. 1993. “Dinner’s perfect companion”. Black Oliver, G. Enterprise Journal, 24(1): 95. 2003. The brewmaster’s table: Discovering Fried, E. the pleasures of real beer with real food. 1991. “Wines & spirits. How to hold a beer Reed. Elsevier, New York. tasting”. Black Enterprise Journal, PASOS. Journal of Tourism and Cultural Heritage, 6(2). 2008 ISSN 1695-7121 Special Issue. Gastronomic and wine tourism
You can also read