Wine Economics and Policy 10(1): 57-71, 2021 Firenze University Press www.fupress.com/wep ISSN 2212-9774 (online) | ISSN 2213-3968 (print) | DOI: 10.36253/wep-9418 Wine Economics and Policy Citation: Carla Ferreira, Lina Louren- ço-Gomes, Lígia M.C. Pinto (2021) Region of origin and perceived quality of wine: an assimilation – contrast approach. Wine Economics and Policy 10(1): 57-71. doi: 10.36253/wep-9418 Copyright: © 2021 Carla Ferreira, Lina Lourenço-Gomes, Lígia M.C. Pinto. This is an open access, peer-reviewed article published by Firenze University Press (http://www.fupress.com/wep) and distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distri- bution, and reproduction in any medi- um, provided the original author and source are credited. Data Availability Statement: All rel- evant data are within the paper and its Supporting Information fi les. Competing Interests: The Author(s) declare(s) no confl ict of interest. Region of origin and perceived quality of wine: an assimilation-contrast approach Carla Ferreira1,*, Lina Lourenço-Gomes2, Lígia M.C. Pinto3 1 University of Minho, 4710-057 Braga, Portugal. E-mail: carlacrisfe@gmail.com 2 CETRAD and DESG, University of Trás-os-Montes and Alto Douro, 5000-801 Vila Real, Portugal. E-mail: lsofi a@utad.pt 3 EEG and NIPE, University of Minho, 4710-057 Braga, Portugal. E-mail: pintol@eeg. uminho.pt *Corresponding author. Abstract. Wine quality perception involves both intrinsic and extrinsic attributes and is related to consumer liking and acceptability of a product. Th e main purpose of this paper is to evaluate the actual role of the region of origin cue on the experienced, expected, and perceived quality of wine, as well as on the discrepancies between them. Using an experimental design set up, real tasting sessions were applied to elicit con- sumer quality perception in three diff erent information conditions: (1) blind tasting (2) labelled tasting (region informed evaluation); and (3) wine tasting under full informa- tion. In total, 136 wine consumers stated their preferences through liking score. Th e results from the assimilation-contrast framework show that region of origin aff ects the experienced, expected, and perceived quality, as well as the agreement between them. Th us, the region of origin may off er a good predictive value of the product, increasing the consumer expectations. Th ese results have important implications for producers as they demonstrate that the region of origin may be used as a brand. Keywords: Assimilation-Contrast approach, product quality, region of origin, wine. 1. INTRODUCTION Increased competition between food suppliers, especially in terms of price and product diff erentiation [1], [2] has enhanced the complexity of the consumers’ choice task. Th e concepts of expected, experienced and perceived quality have been widely reported in the literature pertaining to food quality [3,4]. Cohen and Basu [5] defi ned expected quality as the expectation or belief regarding the anticipated performance of a product. It can then be compared with true evaluation of quality obtained through blind tasting, designated by experi- enced quality [6]. Perceived quality can be defi ned as the subjective response to several explicit features of a product and should be seen in relation to the perceptions and expectations of consumers [7]. In sum, consumer liking and acceptability of the product can be infl uenced by the available information which in turn aff ects expectations. 58 Carla Ferreira, Lina Lourenço-Gomes, Lígia M.C. Pinto It is widely agreed that wine is one of the most dif- ferentiated products on the food market, where consum- ers have to choose from an extended product line with varying objective and subjective characteristics [8,9]. Wine perceived quality is influenced, simultaneously or successively, by non-sensory cues, and sensory cues when the product is tasted [3,4,10,11]. However, in a pur- chasing context, the intrinsic cues, such as sensory prop- erties, are seldom available [12,13] and thus non-sensory cues tend to dominate the choice [15]. Many extrinsic cues, i.e. price, medals, ratings, region of origin, pack- aging, can affect consumers’ choices by creating quality expectations. Perrouty, et al. [16] showed how the region of ori- gin is an extrinsic cue with added value to the consum- ers. In particular, existing literature supports that the expected quality of wine is strongly associated with the region of origin, which is the main extrinsic cue underlying choice (see for example [17-19]). Further- more, the region of origin can play a direct effect in determining consumer behaviour, through the effec- tive linkage between trust and authenticity [20–22]. For Madureira and Nunes [23] and Pettigrew and Charters [24] the influence of information on the region of ori- gin depends on consumer’s knowledge level, gender, and economic status. Empirical studies have revealed that expected quality and experienced quality may not match, showing differences between blind evaluations and extrinsic cue evaluations [3,6,25]. Also, the mis- match found between expected and perceived quality is generally understood as “disconfirmation of expecta- tion” which meaning can depend on the sensory evalu- ation of wine, but also on its extrinsic cues. In this vein, the present research intends to measure the role of the region of origin cue on the experienced, expected, and perceived quality of wine, as well as on the discrepan- cies between them. Applying the conceptual frame- work of expectancy disconfirmation [26,27] this study empirically investigates whether there is a dissonance between perceived, expected and experienced quality among three Portuguese wine regions of origin (Douro, Dão and Alentejo) with different levels of notoriety and image content [28]. Furthermore, the influence of the consumer’s knowledge level of wine in both experienced and perceived quality is analyzed. The novelty of the approach developed derives from the elicitation of the perceived quality obtained through real tasting sessions applied in 5 Portuguese geographical locations, using a specific experimental design based on hedonic evalua- tions under different information conditions. The next section presents the theoretical background and the research hypotheses. 2. THEORETICAL BACKGROUND AND RESEARCH HYPOTHESES Experienced quality of food product depends on sensory characteristics, while perceived quality is also influenced by extrinsic cues, on the other hand expect- ed quality depends crucially on extrinsic cues. When a product is consumed, expectation and sensory experi- ences are combined into a global product evaluation, designated as perceived quality [3,6]. Anderson [26] seminal work, proposed four psy- chological theories to explain the effect of the difference between the expected quality and the overall perceived product quality: (1) cognitive dissonance (assimila- tion); (2) contrast; (3) generalized negativity; and (4) assimilation-contrast. Dissonance or assimilation theory assumes that any discrepancy between expected quality and the perceived quality will be minimized or assimi- lated by a consumer adjustment of the evaluation of the product to be more consistent (less dissonant) with his expectations. This theory argues that an unconfirmed expectancy generates a state of dissonance or “psy- chological discomfort” given that the outcome contra- dicts the consumers’ original hypothesis. Based on this proposition, the extrinsic attributes of a product should substantially lead to expected quality above perceived quality. In this case, the consumer receives two percep- tions that are psychologically dissonant and attempts to reduce this mental discomfort by changing or dis- torting one or the two perceptions to make them more consonant [6,29]. Several criticisms emerged, especially because this theory assumes that the consumer instead of learning from his purchasing mistakes, increases the probability of making them again as he tries to reduce post-purchase dissonance by justification and ration- alization of his decisions [26,30]. Contrast theory (2), argues that if the perceived quality of the product fails to meet the expected quality, the consumer will assess the product less favorably than if he had no prior expecta- tions for it. In this sense, contrast theory assumes that the surprise effect or the contrast among expectations and evaluation will lead to exaggerate or magnify the disparity. Thus, contrast and assimilation theories pre- dict opposing effects [26,30]. The third theory argues that any discrepancy between expected and perceived quality leads to a generalized negative hedonic state, in which the product will receive a more unfavorable rat- ing than if it had coincided with expectations. Following this theory, even if perceived quality exceeds the expe- rienced quality, the product will be perceived as less satisfying than its perceived quality would justify [26]. Finally, the assimilation-contrast theory (4), as the name 59Region of origin and perceived quality of wine: an assimilation-contrast approach implies, combines the theories of assimilation (1) and contrast (2). Th is theory suggests that there are zones of acceptance, rejection, and neutrality in consumer per- ception. Th erefore, if the disparity between expected quality and perceived quality is suffi ciently small to fall into the zone of acceptance, consumers tend to assimi- late the diff erence, rating the product more in line with expected quality than with perceived quality (assimi- lation eff ect). On the other hand, if the discrepancy between expected quality and perceived quality is too large that it falls into the zone of rejection, the consum- er will tend to increase the perceived disparity between expected and perceived quality (contrast eff ect). Th us, an assimilation or contrast eff ect arises as a function of the relative disparity among expected and perceived quality [6,26, 29–31]. This conceptual framework is widely applied by marketing managers to study consumer satisfaction and the likelihood of purchase [6, 29]. Most empirical stud- ies have shown that matching between expected, expe- rienced and perceived quality is not a rule, and that the size of the discrepancy among expected and perceived quality may determine consumers’ fi nal behavior. Sev- eral authors call these discrepancies as “disconfi rmation of expectations” [31–33]. Th e analysis of the compet- ing theories requires the elicitation of consumers’ per- ception of quality and acceptance, for which diff erent approaches have been used: hedonic scores [25,34,35] incentive compatible mechanisms such as auctions [36– 38] and a combination of hedonic scores and auctions [11,18, 39–41]. Th e application of the assimilation-contrast theory to analyze the eff ect of a region of origin on expected quality and therefore its strength [3,6,18], lead to the for- mulation the following research hypotheses: a. Th e sensory perception of a wine is infl uenced by the knowledge of the region of origin; b. Th e region of origin signifi cantly aff ects the experi- enced quality; c. Th e region of origin signifi cantly aff ects the expect- ed quality; d. Th e region of origin signifi cantly aff ects the per- ceived quality; e. Th e region of origin signifi cantly aff ects the diff er- ences between expected and experienced quality; f. Th e region of origin signifi cantly aff ects diff erences between perceived and experienced quality; g. Th e consumers’ wine knowledge type signifi cantly aff ects experienced and perceived quality. To test the research hypotheses, hedonic scores under diff erent information conditions were gathered: (1) blind tasting (evaluates the intrinsic features of wine and provides a measure of experienced quality); (2) the evaluation of region of origin information (a measure of expected quality based on a wine region); and (3) wine tasting under full information (a measure of perceived quality). Moreover, specifi c indicators to test the assim- ilation-contrast theory were computed (see appendix for a detailed description). 3. MATERIALS AND METHODS 3.1 Experimental design and procedure Following the approach adopted by D’Hauteville et al. [3], Kokthi and Kruja [6], and Stefani et al. [18], the hedonic scores were collected through real tasting apply- ing an experimental design replicated over six sessions in fi ve Portuguese regions (Figure 1). Th e tasting session asked participants to evaluate red wines from three Portuguese wine regions (Douro, Dão and Alentejo) under diff erent information condi- tions (blind evaluation; region informed or labelled evaluation; full information). In each session, two Sce- narios were carried (A and B). Each participant took part in only one Scenario. Th e procedure started with a brief explanation of the research goals and tasks to be performed. In case of agreement, the participant signed an informed consent form and was endowed with a gift Figure 1. Summary of experimental protocol. 60 Carla Ferreira, Lina Lourenço-Gomes, Lígia M.C. Pinto card as an incentive. To minimize session effects, the instructions were read aloud by the same experimenter in all sessions. Each red wine sample (30 mL) was served in standard glasses and identified with a three-digit code randomly assigned. The presentation order of wines was randomized across sessions according to a Wil- liams’ Latin square design, balanced for order and first carry-over effects [6]. The full set of six possible combi- nations was used. In Scenario A – blind Scenario, par- ticipants were asked to evaluate the wines on a hedonic scale using a 1-9 Likert scale (1= dislike extremely to 9= like extremely) and to evaluate the intrinsic attributes for each wine sample(colour, aromatic intensity, sour fla- vour, and structure). In Scenario B- informed Scenario, participants received information about the region of origin before the expectation test liking score was obtained. Then, participants were invited to taste each wine and evaluate it using a 1-9 Likert scale (1= dislike extremely to 9= like extremely). Participants were also asked to assess intrin- sic attributes as in Scenario A (Figure 1). Finally, both Scenarios included a questionnaire to collect information regarding: i) socio-demographics; ii) wine consumption and purchasing habits; ii) objective wine knowledge; iii) subjective wine knowledge, follow- ing previous studies on wine consumer behaviour [42]. To identify objective knowledge, Forbes, Cohen, & Dean (2008) test was used (Table 1 reports the specific ques- tions posed and the alternative answers, identifying in italics the correct option). Moreover, to assess subjective knowledge, Flynn and Goldsmith [43] eight-item meas- ure was used. In addition, the two six-item measures proposed by Flynn et al. [44] were applied to measure opinion leadership and opinion seeking. Selection of region of origin and wine Portugal is typically associated with wine produc- tion and consumption. In 2019, it was the 2nd largest wine consuming country among European countries [45]. Historically, wine production in Portugal is struc- tured in 13 demarcated mainland wine regions, where wine can be sold as a certified product (see map of Portugal’s Wine Region in Silva et al. [46]). This certi- fication represents a signal of perceived quality for the consumer, although there are differences as to how the wines connect to the winemaker and contribute to the local economy [47]. In 2018, 62% of still wine consumed in Portugal was red wine [48]. Comparing the market share (in volume and value) of still wines by the thirteen Portuguese wine regions, in 2018 (Figure 2), Alentejo and Douro regions were the most important contribu- tors for total sales in value. However, the Douro region contributed significantly less for total sales in volume. For each wine-producing region under evaluation (Dão, Douro, Alentejo), the wine was selected according to the following criteria: to have an average price in the middle range of the Portuguese off-trade channel (5€ - 12€), the same vintage (2017), and to possess a similar alcohol content. Furthermore, a specialist wine consult- ant firm was recruited to select a wine from each wine region that fulfilled these criteria. Table 2 shows the main characteristics of the three wines selected to taste. Participants One hundred and thirty-six red wine consumers liv- ing in different Portuguese wine regions of origin partic- ipated in this study. A consulting firm recruited the par- ticipants, based on the following criteria: (1) Portuguese native speakers; (2) to have a good general state of health (self-reported); (3) to have some experience in choosing wine; (4) regular still wine consumers; and (5) to have 35 or more years old (according to Bruwer et al. [49], and Wolf et al. [50], older consumers have more experience choosing and consuming red wine). Table 1. Objective wine knowledge test. Question Answer choice (correct choice in italics) Which of the following is a grape of red wine? Alvarinho Chardonnay Touriga Nacional Loureiro Don’t know A peppery character is most associated with which wine? Merlot Shiraz/Syrah Semillion Pinot Noir Don’t know Which is not a famous French wine region? Bordeaux Champagne Rheingau Alsace Don’t know Which is the most appropriate designation for port wine? Still wine Fortified wine Sparkling wine Lat Harvest wine Do not know In 2017, which was the largest producer (in quantity) of wine at European level? Spain Portugal Italy France Do not know 61Region of origin and perceived quality of wine: an assimilation-contrast approach 3.2 Data analysis Participants’ characterization Participants’ characteristics were analysed using univariate descriptive statistics for socio-demographics, wine consumption and purchasing habits, self-reported knowledge, subjective knowledge, opinion leadership, and opinion-seeking behaviour. For objective knowl- edge, a single score of individuals was determined depending on whether participants answered correctly or not the fi ve multiple-choice items that make up the scale. To investigate the psychometric properties of these measures, a principal component factor analysis with a varimax rotation was performed [43,51]. To identify the wine knowledge types, median splits for objective and subjective knowledge measures were determined: par- ticipants with scores above the median on each meas- ure were classifi ed as “high” while the other participants were classifi ed as “low” [51]. Th e resulting four consumer wine knowledge types were identifi ed and labelled as show in Figure 3. Hedonic evaluation To explore the direct impact of the region of ori- gin, we analyzed the diff erence between the evaluation of intrinsic cues (colour, aromatic intensity, acid taste, and structure) and the hedonic scores for each wine. To interpret how hedonic score was aff ected by region of origin information several indicators were calculated, according to Table A2 in the Appendix. Furthermore, we investigate the impact of consum- ers’ knowledge level on experienced and perceived qual- ity. For this, a Kolmogorov-Smirnov test was performed to test the statistical signifi cance of positive and negative diff erences between the blind test liking score (experi- enced quality) and the full information test liking score Figure 2. Market share (in volume and value) of still wines by thirteen Portuguese wine regions, 2018, Source: IVV [48]. Table 2. Main characteristics of the three wines selected to taste. Region of origin Douro Dão Alentejo Grape variety Touriga Nacional, Tinta Roriz and Touriga Franca Touriga Nacional, Tinta Roriz, Alfrocheiro and Jaen Trincadeira and Aragonez Alcohol Content 13,5% 13% 14% Year 2017 2017 2017 Type of bottle Bordeaux Burgundy Bordeaux Colour of bottle Black Black Black Geographical indication PDO PDO PDO Price (€/bottle) * 7 € 6 € 9 € *Mean price off -trade; PDO: Protected Designation of Origin. 62 Carla Ferreira, Lina Lourenço-Gomes, Lígia M.C. Pinto (perceived quality). Statistically signifi cant diff erences were signalled at the confi dence level of 95%. 4. RESULTS AND DISCUSSION Sample description Participants’ profi le is reported in Table 3. Concern- ing the socio-demographic characteristics, participants’ mean age was 44,3 years (SD=8,63 years), 52% of partici- pants were women, household average size (over 18 years old) was 2,27 individuals (SD=1,13), 87% stated to have a higher education level and 43% earn a monthly house- hold income between 581€ and 1 500€. Regarding the purchasing and consumption behaviour, 49% of partici- pants drink wine several times per week, 77 % stated to buy mainly wine from the Douro region, and 50% stated to spend 4,99€ per week on wine. Th e majority (74%) prefer to buy wine in the supermarket. Comparing par- ticipants’ profi les between Scenario A and B, at a signifi - cance level of 5%, there are no signifi cant statistical dif- ferences for all variables, except for monthly purchasing of wine. It is thus possible to compare Scenario eff ects between the two groups [52]. To classify participants into the four types of wine knowledge proposed by Ellis and Coruana [51] we fi rst investigated the validity of the measures of the 20 items making up the three constructs in study (subjec- tive knowledge, opinion leadership and opinion seek- ing) through a principal components factor analysis by applying a varimax rotation. Table 4 shows as each item is loaded separately and distinctively onto four fac- tors. Two items for the opinion leadership measures and one item for the subjective knowledge were excluded to improve model robustness, increasing the explained var- iance to 68%. Figure 3. Wine knowledge types. Source: Adapted from Ellis and Caruana [51]. Table 3. Participants’ profi le description. Relative Frequency Total p-valueScenario A (N= 71) Scenario B (N=65) Gender 0,128 Women 57,7 44,6 51,5 Men 42,3 55,4 48,5 Education level 0,407 5-9 years 2,8 1,5 2,2 10-12 years 12,7 9,2 11 Higher Education 84,5 89,2 86,8 Household monthly income 0,100* < 580 € 0 3,1 1,5 581 €- 1 500 € 42,3 44,6 43,4 1501 € - 2 500 € 33,8 27,7 30,9 2501 € - 3 500 € 18,3 16,9 17,6 3501 € - 4 500 € 1,4 7,7 4,4 > 4 501 € 4,2 0 2,2 Wine consumption frequency 0,075* Never 4,2 4,6 4,4 Once 28,2 38,5 33,1 Several times 47,9 49,2 48,5 Every day 19,7 7,7 14 Wine region of origin that most buys 0,696 Verdes 1,4 3,1 2,2 Douro 78,9 75,4 77,2 Dão 8,5 7,7 8,1 Lisboa 2,8 1,5 2,2 Alentejo 8,5 12,3 10,3 Monthly purchasing of wine (bottle) 0,047** 1 or less 36,6 49,2 42,6 2 to 3 33,8 35,4 34,6 4 or more 29,6 15,4 22,8 Weekly spending of wine 0,161 ≤ 4,99 € 45,1 55,4 50 5,00 € - 9,99 € 39,4 33,8 36,8 10,00 € -14,99 € 5,6 6,2 5,9 15,00 € -49,99 € 8,5 4,6 6,6 ≥ 50,00 € 1,4 0 0,7 Place of purchase 0,097* Hypermarket 71,8 75,4 73,5 Wine Store 11,3 13,8 12,5 Producer 16,9 10,8 14 Notes: *** p<0,001; **p<0,05; *p<0,1. 63Region of origin and perceived quality of wine: an assimilation-contrast approach The findings indicate a cross loading for item six of Flynn and Goldsmith [43] proposed measure. In other words, the item related to the opinion leadership is placed on the subjective knowledge measure. This result can be explained by the relationship between the two measures, as subjective knowledge involves opinion seekers. Vigar-Ellis et al. [42] also found cross loading among factors and items with poor loading. The results show a division of the opinion leadership measure into two constructs, with a leading opinion relationship, the negative opinion leader and the positive opinion lead- er. However, the computation of Cronbach alpha sup- ports the convergent and discriminant validity of the constructs (the Cronbach alpha score for all measures exceed 0,7, providing support for internal consistency, as stated by Nunnally [53]. Regarding the measurement of objective wine knowledge, each question was evaluated as either cor- rect (1 mark) or incorrect (0 mark). The scores for the objective knowledge ranged from 0 to 5, with an average value of 2,60 (SD=1,06). Based on the marks, the sample was split into four segments using subjective and objec- tive knowledge results of participants, according to Fig- ure 3. This resulted in 93 of the participants being classi- fied as “Neophytes” (low subjective-low objective), 25 as “Modest” (low subjective-high objective), 14 as “Snobs” (high subjective- low objective), and only 4 as “Experts” (high subjective-high objective). Table 5 reports the results by consumers’ knowl- edge type, regarding the importance of information on consumers’ choice [11,36]. For all consumer segments, the most important wine cue is the region of origin. Environmental certification appears as indifferent for all knowledge types. Neophytes give more importance to front label design and medals/awards, while Experts ascribe more importance to information as grape vari- ety, winemaker, expected quality price ratio, recommen- dation, previous experience and brand. Comparing the Modest with the Snobs, Snobs give more attention to the quality-price ratio, alcohol content, wine history, brand, and front label design. Moreover, the distribution of the importance of information across knowledge types is statistically different (p-value <0,05) for bottle shape, wine history, winemaker, brand, and medals/awards. In general, these results corroborate those in the literature for the four wine knowledge types [42,54]. Impact of origin region on Hedonic score To assess the impact of the region of origin on the scores ascribed by participants to the features colour, aromatic intensity, acid taste, structure, and overall hedonic scores in two information conditions (blind tasting and full information) a between means unpaired test (Z- Wilcoxon test) was performed (Table 6). Results show that, in general, participants value more the wine attributes when they have previous knowledge about the region of origin (Scenario B) than in the blind informa- tion condition (Scenario A). For the four intrinsic attributes under evaluation, statistically significant differences were found for colour and acid taste (Alentejo wine) as well as aromatic inten- sity (Douro wine). Thus, intrinsic attributes such as col- our, acidity, and aromatic intensity were perceived dif- Table 4. Results of principal components factor analysis followed by varimax rotation. Components 1 2 3 4 (1) I feel quite knowledgeable about wine 0,848 (2) Among my friends, I am one of the ‘experts’ on wine 0,790 (4) I know pretty much about wine 0,724 (5) I do not feel very knowledgeable about wine (R) 0,720 (7) When it comes to wine, I really do not know a lot (R) 0,714 Cronbach’s α 0,99 (16) I do not need to talk to others before I buy a wine 0,820 (17) I rarely ask other people what wine to buy 0,809 (15) When I consider buying wine I ask other people for advice (R 0,753 (18) I like to get others’ opinions before I buy a wine (R 0,704 (20) When choosing wine, other people’s opinions are not important to me 0,659 Cronbach’s α 0,89 (9) My opinion on wine seems not to count with other people 0,885 (10) When they choose a wine, people do not turn to me for advice 0,760 (11) Other people rarely come to me for advice about choosing wine 0,667 (6) Compared to most other people, I know less about wine 0,560 Cronbach’s α 0,86 (13) I often persuade other people to buy the wine that I like 0,874 (14) I often influence other people’s opinions about wine 0,870 (12) People that I know pick wine based upon what I have told them 0,717 Cronbach’s α 0,84 64 Carla Ferreira, Lina Lourenço-Gomes, Lígia M.C. Pinto ferently, depending on the region of origin information (Table 6). Comparing the means of hedonic scores by Sce- nario and by region of origin, there is a valorization of all regions of origin (Table 6), i.e, the information on the region of origin increases the hedonic scores. In blind tasting (Scenario A), consumers assign the high- est mean hedonic score to Douro wine. However, in the full information condition, the Dão wine achieved the highest mean hedonic score. Differences between information Scenarios are statistically significant for Alentejo and Dão wine at p value < 0,05. These results can be explained by the general idea among wine Por- tuguese consumers of an overvalued Alentejo wine region, as well as Dão wine region. According to IVV [48], in volume, the Alentejo wines were the most con- sumed in Portugal, representing 37,4 % of total sales, 73,1% through the retail channel. On the other hand, for Douro wine, the differences were not statistically significant between both scenarios (at a significance level of 5%). Consumers follow the same hedonic assess- ment with or without information about the region of origin. In 2018, Douro wine represented 12,4 % of total sales, in volume, mainly (68%) in restaurants [48]. The hypothesis that sensory perception of the wine is influ- enced by the knowledge of the region of origin was sup- ported by the results, reinforced by the need of tasting in hedonic evaluation to avoid individuals’ assumptions about the perceived quality of the products [55, 56]. Ste- fani et al. [18]and D’Hauteville et al. [3] found a similar behaviour when investigating the impact of region of origin on hedonic score. The hedonic score expressed in the full information scenario is higher than the hedonic score obtained under blind test condition. Furthermore, Masson et al. [12] and Vecchio et al. [57] demonstrate the influence of extrinsic cues (i.e. low-alcohol wien and process impacts) on the sensory perception. In same line, these authors show that the sensory perception of a wine is influenced by the knowledge of the extrinsic cue. Assimilation and Contrast effects To test the assimilation and contrast effects six indi- cators were computed: Expected quality – Experienced quality; Perceived quality – Experienced quality; Per- ceived quality – Expected quality; Assimilation effect (α); Moderating effect of information (MI); and Dis- sonance effect (DI). According to the results reported in Table 7, a statistically significant difference between expected quality and experienced quality was found for Table 5. Mean importance score of information seek by consumers’ knowledge type. Mean score Consumers knowledge type Kruskal-Wallis test p-values Neophytes Modest Snobs Experts Region of origin 6 6 6 6 0,406 Sensory profile 5 5 5 5 0,426 Food pairing 5 5 5 5 0,446 Environmental certification 4 4 4 4 0,051* Grape variety 3 5 5 6 0,444 Front label design 6 3 4 3 0,132 Bottle form 5 4 4 3 0,024** Wine history 4 4 5 5 0,000*** Winemaker 3 5 5 6 0,000*** Brand 4 5 6 6 0,005** Medals/awards 6 5 5 4 0,038** Expected quality-price ratio 5 5 6 6 0,703 Recommendation 5 5 5 6 0,445 Alcohol content 4 4 5 5 0,271 Qr code 3 4 4 4 0,051* Previous experience 5 5 5 6 0,659 Importance level on a scale of one to seven with one equal to No at all important and seven equal to Extremely important; *** p<0,001; **p<0,05; *p<0,1 65Region of origin and perceived quality of wine: an assimilation-contrast approach the three regions of origin. In other words, the score of expected quality was slightly above the experienced quality in blind tasting, indicating the non-confirmation of expectations for each wine tested and the region of origin effect on consumers’ preferences. The mean of disagreement between the expected quality and experienced quality was higher for Alentejo wine, with a dissonance (DI) value of 24%. On the other hand, for Douro wine the DI value is only 7%, suggest- ing that the effect of region of origin is not homogene- ous. These findings are in line with the results reported in Stefani et al. [18], D’Hauteville et al. [3] and Masson et al. [12]. The effect of assimilation or contrast is significant and positive for the three wines under study (Table 7). The region of origin information affects the overall wine evaluation increasing the mean of liking ratings. Especially, for Alentejo wine, the information about the region of origin leads to a 16% increase in experienced quality. Thus, the findings suggest that there is an assimi- lation effect for the three regions of origin under analysis. The results reveal statistically significant differ- ences between full information conditions and expec- tated evaluation (Table 7). For the three wines, the lik- ing scores decreased in full information conditions, showing that the product did not meet the expectations. This effect is greatest for Alentejo wine, the least appre- ciated in sensorial terms. In other words, there is a posi- tive partial assimilation or negative disconfirmation of expectations for the three regions of origin. These find- ings suggest that the wines are less tasty than the aver- age participants’ expectancy, probably because partici- pants expected better, given some recognized regions of origin, as explained by Lange et al. [40]. Regarding the assimilation coefficients (α), the three wines reported a coefficient higher than 0,5, indicating the predominant effect of region of origin on the overall eval- uation of the wine. Overall results confirm that perceived quality depends on the expectation of the region of ori- gin, as reported by Kokthi and Kruja [6] and Vecchio et al. [57]. Furthermore, these results confirm the empirical evi- dence found in previous research that sensory cue by itself is not a discriminative of consumers’ evaluation [18]. Assimilation-contrast theory helps to understand the differences that may exist in terms of the strength Table 6. Mean values of hedonic scores with blind tasting (Scenario A) and with full information (Scenario B) for the three wines. Attributes Region of origin Douro Alentejo Dão Colour A 3,68 3,18 3,65 Colour B 3,57 3,54 3,74 Colour B-Colour A -0,11 0,36** 0,09 Aromatic intensity A 3,21 3,18 3,35 Aromatic intensity B 3,49 3,43 3,48 Aromatic intensity B- Aromatic intensity A 0,28** 0,25 0,13 Acid taste A 3,18 3,54 3,28 Acid taste B 3,40 3,25 3,3 Acid taste B-Acid teste A 0,22 -0,29* 0,02 Structure A 3,27 3,18 3,38 Structure B 3,35 3,28 3,31 Structure B-Structure A 0,08 0,1 -0,07 Hedonic score A 6,55 5,96 6,18 Hedonic score B 6,82 6,89 7 Hedonic score B- Hedonic A 0,27* 0,93** 0,82** Nº Obs. Scenario A 71 71 71 Nº Obs. . Scenario B 65 65 65 Attribute A = score attribute mean with blind tasting; Attribute B= score attribute mean with full information. ***Statistically significant at p-value<0,01; **Statistically significant at p-value<0,05; *Statistically significant at p-value<0,1 Table 7. Computed indicators by region of origin. Indicators Region of origin Douro Alentejo Dão Expected quality – Experienced quality 0,45*** 1,44*** 1,22 *** Perceived quality – Experienced quality 0,27* 0,93** 0,82*** Perceived quality – Expected quality -0,18*** -0,51*** -0,40 ** Assimilation coefficients (α) 0,60 >0,5 0,65>0,5 0,67 >0,5 Moderating effect of information (%) 4 16 13 Dissonance effect (%) 7 24 20 Assimilation/Contrast effect Partial Positive Assimilation Partial Positive Assimilation Partial Positive Assimilation ***Statistically significant at p-value<0,01; (z-Wilcoxon test). 66 Carla Ferreira, Lina Lourenço-Gomes, Lígia M.C. Pinto of the region of origin on the wine [57]. Based on this theory, the results suggest that if the disparity between expected quality and perceived quality is sufficiently small to fall into the zone of acceptance, the consum- ers tend to partly assimilate the difference. Therefore, the hypotheses that the region of origin significantly affects experienced, expected, and perceived quality are supported. Also, these results confirm that the region of origin significantly affects differences between expected and experienced quality; and the differences between perceived quality and experienced quality. In sum, these results highlight the effect of region of origin information on wine consumers’ preferences. Previously, several authors have shown that the wine evaluation is influenced by both intrinsic cues (as taste) and extrinsic cues (as region of origin or brand), which affect the perceived quality of the wine [34,58–60]. On the other hand, Masson et al. [12] and Vecchio et al. [57] applied the assimilation-contrast theory to study the effect of other extrinsic cues, such as low-alchol and process impact, respectively, on wine perceived qual- ity. The results of this study are in line with previous research findings, however, few studies have applied the assimilation-contrast theory to investigate the effect of region of origin on wine’s perceived quality [3,18], as developed here. Impact of wine consumers’ knowledge type on experienced and perceived quality To investigate the difference of experienced and per- ceived quality across consumers’ wine knowledge type, a Kolmogorov-Smirnov test was performed (results for Experts are not reported as only one subject belongs to this category). Table 8 shows that only Neophytes pre- sent statistically significant differences between expe- rienced and perceived quality. Comparing the hedonic score distribution for the three wines, statistically sig- nificant differences were found only for the Alentejo wine. The results indicate that this group ascribes higher hedonic scores for Alentejo wine in blind tasting (expe- rienced quality). Following the distinctions discussed by Ellis & Caruana [51] for the different consumer knowl- edge types, Neophytes recognize that they know very lit- tle about wine, but like to consume wine. A basic prod- uct with low prices and intensively distributed will likely be the most sought by this segment of consumers. Thus, a feasible reason for the results obtained is the familiari- ty of the consumers to certain sensorial profile, respond- ing more to brands than to the region of origin. In this context, the hypothesis that wine consumers’ knowledge type has significant effects on experienced and perceived quality was partially verified. This result is in line with those reported in previous literature [3,12,57]. A summary comparison table of our results and those from previous literature is presented in the appen- dix (Table A3). 5. CONCLUSION The region of origin cues influence the consum- er evaluation of food products as far as it can act as a quality cue to other features of the good and/or it can affect the liking of food through its symbolic or affective meaning. This is especially important for wine as it is an information-intensive product offering multidimensional decision challenges for consumers. Understanding the strength of region of origin on perceived quality of wine, and how it varies across mar- ket segments is essential for the design of successful marketing strategies. Considering three Portuguese wine regions of ori- gin, the present study provides empirical evidence that Table 8. Distributions of hedonic scores by consumer knowledge type between two informational Scenarios (blind tasting and com- plete information). Consumer knowledge type1 Region of origin Hypotheses2 Kolmogorov- Smirnov Z (p-values) Neophytes Douro hs(EQ)hs(PQ) 0,644 Alentejo hs(EQ)hs(PQ) 1,000 Dão hs(EQ)hs(PQ) 1,00 Modest Douro hs(EQ)hs(PQ) 0,877 Alentejo hs(EQ)hs(PQ) 1,000 Dão hs(EQ)hs(PQ) 1,000 Snobs Douro hs(EQ)hs(PQ) 0,953 Alentejo hs(EQ)hs(PQ) 0,953 Dão hs(EQ)hs(PQ) 0,497 **Statistically significant at p-value<0,05; *Statistically significant at p-value<0,1. 1The expert knowledge consumer group is composed of only one individual, thus the group is absent from the table. 2EQ=Experienced quality; PQ=Perceived quality. 67Region of origin and perceived quality of wine: an assimilation-contrast approach attest the impact of the region of origin on consum- ers’ preferences, namely that it affects the expected, the experienced and the perceived quality of the wine. It also shows that consumers’ knowledge provides a use- ful basis for segmenting the wine market, which rein- forces the bet on the characterization of consumers by wine marketers. The Neophytes segment shows hedonic sensitivity to positively evaluate a known sensory pro- file. However, further research is required to test the responses of the segments to other marketing mix vari- ables. Additionally, a predominant effect of region of origin on the overall evaluation of the three wines was found. This paper supports important findings with respect to the relationships between expected quality of region of origin and its market strength. In the full informa- tion condition, participants decreased hedonic rating of all regions of origin, especially for Alentejo, which pre- sented the highest percentage of dissonance. This sug- gests that the Alentejo region has a brand in the market that leads to higher consumer expectations. On the oth- er hand, for other regions, Dão and Douro, investments should go to brand construction. Moreover, the paper sheds light on the role of the region of origin in moderating the impact of experi- enced quality on consumers’ preferences. In particular, it emerged that each region of origin is perceived different- ly according to its strength in the wine market. In light of this, intensive advertising and communication strate- gies can help to enhance the region of origin as a brand in the market thus improving the perceived quality of its wine. The results reported in this study need to be con- sidered in light of its limitations. Part of our results may depend on the choice of wines, although we controlled the selection criterion to obtain a representative sample. In this line, further research needs to be carried using authentic consumption situations, including other mar- keting mix variables and other wine regions. Several practical implications derive from these findings. Wine producers should carefully transmit the information and the specific product features, both in terms of sensory profile and in terms of market reputa- tion. Moreover, wineries could run information cam- paigns to communicate differences in sensory profile between regions of origin. In future research, it is cru- cial to investigate more deeply specific sensory attributes that influence wine consumer preferences, affect the per- ceived wine quality with a special focus on specific con- sumer segments. 6. REFERENCES [1] C. Altomonte, I. Colantone, and E. Pennings, “Het- erogeneous Firms and Asymmetric Product Differ- entiation,” J. Ind. Econ., vol. 64, no. 835–874, 2016. [2] P. Combris, P. Bazoche, E. Giraud-Héraud, and S. Issanchou, “Food choices: What do we learn from combining sensory and economic experiments?,” Food Qual. Prefer., vol. 20, no. 8, pp. 550–557, 2009. [3] F. D’Hauteville, M. Fornerino, and P. Perrouty, ““Disconfirmation of taste as a measure of region of origin equity: An experimental study on five French wine regions,” Int. J. Wine Bus. Res., vol. 19, no. 1, pp. 33–48, 2007. [4] S. Charters and S. Pettigrew, “The dimensions of wine quality,” Food Qual. Prefer., vol. 18, no. 7, pp. 997–1007, 2007. [5] J. Cohen and K. Basu, “Alternative models of cate- gorization: towards a contingent processing frame- work,” J. Consum. Res., vol. 13, no. 4, pp. 455–72, 1987. [6] E. Kokthi and D. Kruja, “Consumer Expectations for Geographical Origin: Eliciting Willingness to Pay (WTP) Using the Disconfirmation of Expecta- tion Theory (EDT),” J. Food Prod. Mark., vol. 23, no. 8, pp. 873–889, 2017. [7] D. Calvo, “Analysis of brand equity supplied by appellations of origin: An empirical application for beef,” J. Int. Food Agribus. Mark., vol. 14, no. 3, pp. 21–34, 2002. [8] A. Bouet, C. Emlinger, and V. Lamani, “No Title,” J. Wine Econ., vol. 12, no. 1, pp. 37–58, 2017. [9] L. Lockshin and A. Corsi, “Consumer behaviour for wine 2 . 0 : A review since 2003 and future directions,” Wine Econ. Policy, vol. 1, no. 1, pp. 2–23, 2012. [10] S. Bowen and T. Mutersbaugh, “Local or localized? Exploring the contributions of Franco-Mediterra- nean agri-food theory to alternative food research,” Agric. Human Values, vol. 31, pp. 201–213, 2014. [11] M. Costanigro, S. Kroll, D. Thilmany, and M. Bun- ning, “Is it love for local/organic or hate for conven- tional? Asymmetric effects of information and taste on label preferences in an experimental auction,” Food Qual. Prefer., vol. 31, no. 1, pp. 94–105, 2014. [12] J. Masson, P. Aurier, and F. D’hauteville, “Effects of non-sensory cues on perceived quality: the case of low-alcohol wine,” Int. J. Wine Bus. Res., vol. 20, no. 3, pp. 215–229, 2008. [13] P. O. Williamson, L. Lockshin, I. L. Francis, and S. Mueller Loose, “Influencing consumer choice: 68 Carla Ferreira, Lina Lourenço-Gomes, Lígia M.C. Pinto Short and medium term effect of country of origin information on wine choice,” FOOD Qual. Prefer., vol. 51, pp. 89–99, 2016. [14] P. Tait, C. Saunders, M. Guenther, and P. Ruther- ford, “Emerging versus developed economy con- sumer willingness to pay for environmentally sus- tainable food production: A choice experiment approach comparing Indian, Chinese and United Kingdom lamb consumers,” J. Clean. Prod., vol. 124, pp. 65–72, 2016. [15] R. Cross, A. Plantinga, and R. Stavins, “What Is the Value of Terroir?,” Am. Econ. Rev., vol. 101, no. 3, 2011. [16] J. Perrouty, F. D’Hauteville, and L. Lockshin, “The influence of wine attributes on region of origin equity : An analysis of the moderating effect of consumer’s perceived expertise,” Agribusiness, vol. 22, no. 3, pp. 323–341, 2006. [17] R. Johnson and J. Bruwer, “Regional brand image and perceived wine quality: the consumer perspec- tive,” Int. J. Wine Bus. Res., vol. 19, no. 4, pp. 276– 97, 2007. [18] G. Stefani, D. Romano, and A. Cavicchi, “Consum- er expectations, liking and willingness to pay for specialty foods: Do sensory characteristics tell the whole story?,” Food Qual. Prefer., vol. 17, no. 1–2, pp. 53–62, 2006. [19] C. Vaquero-Piñeiro, “A voyage in the role of terri- tory: are territories capable of instilling their pecu- liarities in local production systems?,” Collana del Dip. di Econ. Univ. degli Stud. Roma Tre Work. Pap., no. 251, 2020. [20] P. Verlegh and K. Van Ittersum, “The origin of spices: the impact of geographic product origin on consumer decision making,” in Food, people and society, L. Frewer, E. Risvik, and H. Schifferstein, Eds. Berlin: Springer, 2001, pp. 267–280. [21] S. Charters and N. Spielmann, “Characteristics of strong territorial brands: the case of champagne,” J. Bus. Res., vol. 67, pp. 461–467, 2014. [22] L. Lourenço-Gomes, L. M. C. Pinto, and J. Rebelo, “Wine and cultural heritage. The experience of the Alto Douro Wine Region,” Wine Econ. Policy, vol. 4, no. 2, pp. 78–87, 2015. [23] T. Madureira and F. Nunes, “Relevant attributes of Portuguese wines: matching regions and consum- er’s involvement level,” Int. J. Wine Bus. Res., vol. 25, no. 1, pp. 75–86, 2013. [24] S. Charters and S. Pettigrew, “Product involvement and the evaluation of wine quality,” Qual. Mark. Res. An Int. J., vol. 9, no. 2, pp. 181–193, 2006. [25] E. Lick, B. König, M. Kpossa, and V. Buller, “Sen- sory expectations generated by colours of red wine labels,” J. Retail. Consum. Serv., vol. 37, no. Octo- ber 2016, pp. 146–158, 2017. [26] R. Anderson, “Consumer dissatisfaction: The effect of disconfirmed expectancy on perceived product performance,” J. Mark. Res., vol. 10, 1973. [27] H. N. Schifferstein and J. Mojet, “Asymmetry in the Disconfirmation of expectations for natural yogurt,” Appetite, vol. 32, 1999. [28] C. Ferreira, L. Lourenço-Gomes, L. Pinto, and P. Silva, “Is there a gender effect on wine choice in Portugal? – A qualitative approach,” Int. J. Wine Bus. Res., pp. 1751–1062, 2019. [29] K. R. Teas and K. Palan, “Disconfirmed Expecta- tions Theory of Consumer Satisfaction: An Exami- nation of Representational and Response Language Effect,” J. Consum. Satisf. Complain. Behav., vol. 16, no. 81, pp. 81–105, 2003. [30] L. Festinger, A Theory of Cognitive Dissonance. Evanston: Row and Peterson, 1957. [31] H. Schifferstein, “Effects of products beliefs on product perception and liking,” in Food, People and Society a European Perspective of Consumer Choices, 2001, pp. 73–96. [32] E. Anderson and M. Sullivan, “The Antecedents and Consequences of Customer Satisfaction for Firm2,” Mark. Sci., vol. 2, no. 12, pp. 125–143, 1993. [33] R. Deliza and H. J. H. MacFie, “The generation of sensory expectation by external cues and its effect on sensory perception and hedonic ratings: a review,” J. Sens. Stud., vol. 11, pp. 103–128, 1996. [34] J. Bruwer, A. Saliba, and B. Miller, “Consumer behaviour and sensory preference differences: implications for wine product marketing,” J. Con- sum. Mark., vol. 28, no. 1, pp. 5–18, 2011. [35] I. L. Francis and P. O. Williamson, “Application of consumer sensory science in wine research,” Aust. J. Grape Wine Res., vol. 21, pp. 554–567, 2015. [36] C. R. Gustafson, “The role of knowledge in choice, valuation, and outcomes for multi-attribute goods,” J. Agric. Food Ind. Organ., vol. 13, no. 1, pp. 33–43, 2015. [37] S. Lecocq, T. Magnac, M. Pichery, and M. Vis- ser, “The Impact of Information on Wine Auction Prices: Results of an Experiment,” Ann. Econ. Stat., 2004. [38] E. Pomarici, D. Asioli, R. Vecchio, and T. Næs, “Young consumers’ preferences for water-saving wines: An experimental study,” Wine Econ. Policy, vol. 7, no. 1, pp. 65–76, 2018. [39] R. K. Gallardo, Y. A. Hong, M. Silva Jaimes, and J. Flores Orozco, “Investigating consumer food 69Region of origin and perceived quality of wine: an assimilation-contrast approach choice behavior: An application combining senso- ry evaluation and experimental auctions,” Cienc. e Investig. Agrar., vol. 45, no. 1, pp. 1–10, 2018. [40] C. Lange, C. Martin, C. Chabanet, P. Combris, and S. Issanchou, “Impact of the information pro- vided to consumers on their willingness to pay for Champagne: Comparison with hedonic scores,” Food Qual. Prefer., vol. 13, no. 7–8, pp. 597–608, 2002. [41] S. Muller, L. Lockshin, and J. Louviere, “What you see may not be what you get: Asking consumers what metters may not reflect what they choose,” Mark. Lett., vol. 21, pp. 335–350, 2010. [42] D. Vigar-Ellis, L. Pitt, and A. Caruana, “Does objective and subjective knowledge vary between opinion leaders and opinion seekers? Implications for wine marketing,” J. Wine Res., vol. 26, no. 4, pp. 304–318, 2015. [43] L. R. Flynn and R. E. Goldsmith, “A short, reliable measure of subjective knowledge,” J. Bus. Res., vol. 46, no. 1, pp. 57–66, 1999. [44] L. Flynn, R. Goldsmith, and J. Eastman, “Opinion Leaders and Opinion Seekers: Two New Measure- ment Raju, P. S., Lonial, Subhash C., and Mangold, W. Glynn: Differential Scales,” J. Acad. Mark. Sci., pp. 137–147, 1996. [45] Nationmaster,2020. [Online]. Available: https:// www.nationmaster.com/nmx/timeseries/portugal- wine-consumption-per-capita. [46] A. Silva, M. Fernão-Pires, and F. Bianchi-de-Agu- iar, “Portuguese vines and wines: Heritage, quality symbol, tourism asset,” Ciência e Técnica Vitiviní- cola, vol. 33, no. 1, pp. 31–46, 2018. [47] J. Rebelo, L. Lourenço-Gomes, T. Gonçalves, and J. Caldas, “A hedonic price analysis for the portu- guese wine market: Does the distribution channel matter?,” J. Appl. Econ., vol. 22, no. 1, pp. 40–59, 2019. [48] IVV, “Mercado Nacional Vinhos Tranquilos.” 2018. [49] J. Bruwer, E. Li, and M. Reid, “Segmentation of the Australian wine market using a wine-related lifestyle approach,” J. Wine Res., vol. 13, no. 3, pp. 217–242, 2002. [50] M. Wolf, S. Carpenter, and E. Qenani-Petrela, “A comparison of X, Y, and boomer generation wine consumers in California,” J. Food Distrib. Res., vol. 36, no. 1, pp. 86–91, 2005. [51] D. Ellis and A. Caruana, “Consumer wine knowl- edge: components and segments,” Int. J. Wine Bus. Res., vol. 30, no. 3, pp. 277–291, 2018. [52] A. Botelho, I. Dinis, L. Lourenço-Gomes, J. Morei- ra, L. Costa Pinto, and O. Simões, “The effect of sequential information on consumers’ willingness to pay for credence food attributes,” Appetite, vol. 118, pp. 17–25, 2017. [53] J. Nunnally, Psychometric Theory. New York: McGraw Hill, 1978. [54] D. Ellis and F. Thompson, “The effect of wine knowledge type on variety seeking behavior in wine purchasing,” J. Wine Res., vol. 29, no. 2, pp. 71–86, 2018. [55] G. Caporale, S. Policastro, A. Carlucci, and E. Monteleone, “Consumer expectations for sensory properties in virgin olive oils,” Food Qual. Prefer., vol. 17, pp. 116–115, 2006. [56] S. Stolzenbach, W. Bredie, R. Christensen, and D. Byrne, “Impact of product information and repeat- ed exposure on consumer liking, sensory percep- tion, and concept associations of local apple juice,” Food Res. Int., vol. 52, pp. 91–98, 2013. [57] R. Vecchio et al., “The role of production process and information on quality expectations and per- ceptions of sparkling wines,” J. Sci. Food Agric., vol. 99, no. 1, pp. 124–135, 2019. [58] L. Lockshin and A. Corsi, “Consumer behaviour for wine 2.0: A review since 2003 and future direc- tions,” Wine Econ. Policy, vol. 1, no. 1, pp. 2–23, 2012. [59] M. Sáenz-Navajas, E. Campo, A. Sutan, J. Ball- ester, and D. Valentin, “Perception of wine quality according to extrinsic cues: The case of Burgundy wine consumers,” Food Qual. Prefer., vol. 27, pp. 44–53, 2013. [60] I. Schäufele and U. Hamm, “Consumers’ percep- tions, preferences and willingness-to-pay for wine with sustainability characteristics: A review,” J. Clean. Prod., vol. 147, pp. 379–394, 2017. [61] M. Fornerino and F. D’Hauteville, “How good does it taste? Is it the product or the brand?,” J. Prod. Brand Manag., vol. 19, pp. 34–43, 2010. 70 Carla Ferreira, Lina Lourenço-Gomes, Lígia M.C. Pinto APPENDIX According to Schifferstein [31] there are three ways to elicit sensory and non-sensory quality preferences depending on the information set available: blind test liking score (B – experienced quality: no information); expectation test liking score (E- expected quality: pro- vision of non-sensory information) and full informa- tion test liking score (F-perceived quality: provision of non-sensory and sensory information). The difference between perceived quality and expected quality is des- ignated as degree of disconfirmation; if expected qual- ity is compared to experienced quality the degree of incongruence can be computed. Finally, comparing the perceived quality with experienced quality, the degree of response shift is computed. Schifferstein [31] proposes the analysis of ratio α, equal to the degree of response shift over the degree of incongruence, translating the assimilation effect. The assimilation-contrast theory can be interpreted as a mechanism by which the individu- als try to adapt psychologically to their environment [61]. Table A1 summarises the different assimilation and contrast effects. Assimilation is absent (α equal to zero) when there is no discrepancy between expected quality and perceived quality. On the other hand, there is an assimilation effect (positive or negative) whenever that change of perceived quality is in the same direction of expected quality; while contrast effect (positive or nega- tive) occurs when the change of perceived quality moves in the opposite direction of expected quality [6]. Table A1. Assimilation and Contrast effects. Perception (Information conditions) Assimilation Contrast Partial Positive Partial Negative Complete Assimilation Positive Negative Expected quality – Experienced quality (E-B) >0 <0 >0 >0 <0 Perceived quality – Experienced quality (F-B) >0 <0 >0 <0 >0 Perceived quality – Experienced quality (F-E) <0 >0 0 >0 <0 Notes: B -Blind test liking score; E -Expectation test liking score; F -Full information test liking score. Table A2. Hedonic score differences tested. Indicators Application Data analysis Expected quality – Experienced quality Expectation test liking score (E) - Blind test liking score (B) – It is calculated to identify the effect of region of origin information on consumers preferences. – There are effects of region of origin on consumers preferences if E -B >0 Perceived quality – Experienced quality Full information test liking score (F) - Blind test liking score (B) – It is calculated to identify if there is assimilation or contrast effect – It shows to what extent product information (region of origin + sensory test) affects hedonic scores. Perceived quality – Expected quality Full information test liking score (F) - Expectation test liking score (E) – It is calculated to identify if assimilation is partial or full; – There is complete assimilation if F-E=0. Assimilation coefficients (α) α = Perceived quality – Experienced quality (F-B), Expected quality – Experienced quality (E-B) 0≤ α≥1 – if α< 0,5, then sensory features are the most important in the product evaluation; – if α> 0,5 region of origin is preferable to sensory features. Dissonance effect (DI) DI(%) = Expected quality – Experienced quality (E-B) * 100 Experienced quality (B) – It measures the distance among expected quality and experienced quality as a percentage from the baseline outcome experienced quality Moderating effect of information (MI) α = Perceived quality – Experienced quality (F-B) * 100 Experienced quality (B) – It measures the average effect of information, as a percentage from the experienced quality on the perceived quality 71Region of origin and perceived quality of wine: an assimilation-contrast approach Table A3. Assimilation-Contrast theory findings: comparison by wine evaluation’ studies. Present paper Stefani et al. [18] D’Hauteville et al. [3] Masson et al. [12] Vecchio et al. [57] Characteristics of study Extrinsic cues under evaluation Region of origin Region of origin Region of origin Low-alcohol wine Process impacts Main Results The sensory perception of a wine is influenced by the knowledge of the extrinsic cue (i.e., region of origin) ✓ ✓ ✓ ✓ ✓ The extrinsic cue under evaluation (i.e., region of origin) significantly affects the experienced quality ✓ ✓ ✓ ✓ ✓ The extrinsic cue under evaluation (i.e., region of origin) significantly affects the expected quality ✓ ✓ ✓ ✓ ✓ The extrinsic cue (i.e., region of origin) significantly affects the perceived quality ✓ ✓ ✓ ✓ ✓ The extrinsic cue (i.e., region of origin) significantly affects the differences between expected quality and experienced quality ✓ ✓ ✓ ✓ ✓ The extrinsic cue (i.e., region of origin) significantly affects differences between perceived quality and experienced quality ✓ ✓ ✓ ✓ ✓ The consumers’ wine knowledge type significantly affects experienced and perceived quality ✓ n.a. ✓ ✓ ✓ n.a.: not application; ✓: Supported.