Product-specific regional image as a moderator of regional identity and consumer regiocentrism effects on wine purchasing behaviour Sandra Pestar Bizjak, Hristo Hristov, Aleš Kuhar Abstract Consumer ethnocentrism and national identity are constructs deriving from social identity theory which address the importance of place of origin in consumer behavior where consumers prefer products from their own country. The objective of this study was to examine the effects of consumer regiocentrism and regional identity on wine purchasing behavior. Further, we aimed to understand the influence of wine-specific regional image on the perception and wine purchasing behavior. A sample of 221 residents from two Slovenian wine regions who at least occasionally buy and drink wine filled in an online questionaire. To measure consumer regiocentrism and regional identity adapted versions of CETSCALE and regional identity scales were used. Once realiability and validity of constructs were established, hypotheses were tested. Consumer regiocentrism and regional identity scores in two studied regions were similar, however only in wine region Podravje it was shown that those consumers who had higher scores of consumer regiocentrism and regional identity were more likely to buy wine from their own wine region. Contrary, respondents from region Primorska purchased majority of wine in the last year from their home wine region independently of levels of consumer regiocentrism and regional identity, which could be a consequence of very positive wine-specific regional image. Our study confirms that product perception is influenced by product-specific regional image and we propose that when product-specific regional image is very positive, consumer regiocentrism and regional identity scores seem to be less relevant in wine purchasing decisions. Introduction Regionality of food products has become important criterion for European consumers in their product evaluation and purchasing decisions (Lorenz et al., 2015). Aim of this study was to identify opportunities to tailor marketing approach to a region-of-origin level considering the effects of regional identity and consumer regiocentrism on wine purchasing behavior on the background of product-specific (in this study wine-specific) regional image from the perspective of regional consumers. In addition to quality cue, country-of-origin has symbolic and emotional meaning for consumers (Verlegh and Steenkamp, 1999). However, while country-of-origin strategies have been succesful in helping a firm or an industry to attain brand equity for a given product; these strategies do not necessarily derive their uniqueness from a specific geographic location (Thode and Maskulka, 1998). Region-of-origin effects are expected to operate through similar processes as country-of-origin, however regions have some unique aspects, which make useful to investigate them (van Ittersum, 2002; van Ittersum et al., 2003). Strategy of marketing agricultural products based on region-of-origin enables the perception of superior quality and development of long-term competitive advantage (Thode and Maskulka, 1998). Moreover, regions in general are much more homogenous in terms of cultural, social, emotional and environmental factors as countries. According to van Ittersum (2002) and van Ittersum et al. (2007) a region is defined as an area situated in one or more countries which forms an entity based on local and regional characteristics, such as traditions, culture and scenery. They defined regional product as a product whose quality and/or fame can be attibuted to its region of origin and which it is marketed using the name of the region-of-origin. Van Ittersum et al. (2003) demonstrated that product perception is significantly influenced by the perceived product-specific regional image. With our study we aimed to understand if product-specific regional image also affects consumers purchasing behaviour. Normative preference for regional products is based on the concept of ethnocentrism (Lorenz et al., 2015). The consumer ethnocentrism evidence and CETSCALE importantly contribute to the growing body of county-of-origin studies (Shimp and Sharma, 1987). General concept of ethnocentrism was introduced more than 100 years ago by Sumner (Sumner, 1908) but conceptualised later on the backbone of social identity theory which examines the relationship of the individual with the group (Sharma et al., 1995; Zeugner-Roth et al., 2015). Shimp and Sharma (1987) used term consumer ethnocentrism to describe beliefs of consumers about the appropriateness and morality of purchasing foreign-made products, which would (from the perspective of ethnocentric consumers) be wrong as it hurts domestic economy, causes loss of jobs and is unpatriotic. In functional terms they said that consumer ethnocentrism gives an individual a sense of identity, feelings of belongingness and an understanding of what kind of purchasing behavior is acceptable or unacceptable for the ingroup. They also formulated and validated CETSCALE, an instrument to measure consumer ethnocentric tendencies related to purchasing foreign versus American made products (Shimp and Sharma, 1987). Later, consumer ethnocentrism as a construct to study consumer behaviour has been studied extensively (Sharma et al., 1995; Vida and Fairhurst, 1999; Lindquist et al., 2001; Vida and Reardon, 2008; Vida et al., 2008; and others), also in relation to agricultural products (Orth and Firbasova, 2003; Bianchi and Mortimer, 2015), and also on the

Slovene population (Vida and Maher Pirc, 2006; Vida and Reardon, 2008; Zeugner-Roth et al., 2015) where it was shown that consumer´s attachment to and concern for his or her country, its people and national symbols directly transcend into consumers´preference for domestic vs. international products, brands and institutions (Vida and Reardon, 2008) and also that the relationship between consumer ethnocentric tendencies is moderated by product necessity and the extent of threat these products are perceived as having on the consumer personally and on the domestic economy more generally (Sharma et al., 1995; Lee et al., 2016). Shimp and Sharma (1987) noted, that regional marketing is an especially exciting application of the CETSCALE. Etnocentrism relates to the social identity theory which is based on a set of concepts which address intergroup relations and was proposed by Tajfel and Turner (Tajfel, 1981; Tajfel and Turner, 1986). According to the social identity theory a person´s social identity is seen as a part of the self-concept, which tipically contains characteristics that represent the social groups or categories to which one belongs (Halldorson, 2009). Transformation of the self-concept to the collective level of identity is presumed to underlie many important group processes, including conformity to group norms, group polarization, crowd behavior, and in-group favoritism in relation to out-groups (Brewer et al., 1993). Later it was shown that social identity approach can also be employed to social contexts comprising ingroups at different levels of inclusiveness and that social groups with which people identify vary, among other dimensions, on the dimension of inclusiveness, where regional identity is an example where smaller, in this case regional group of people consider themselves as a rather disctinct entity and not just a sample of the more inclusive national group (Simon et al., 1995). Regional identity is explaining pro-in group while consumer regiocentrism desribes anti-out group tendencies meaning they are conceptually different constructs therefor it is important to study the effects of both independently (Zeugner-Roth et al., 2015). One of the aims of our study was to determine how both constructs independently affect actual wine purchasing behavior. Contrary to consumer ethnocentrism and national identity, which were intensively analysed, the constructs of consumer regiocentrism and regional identity have been somewhat neglected from academic point of view. The first study measuring both constructs was done in 1998 by Lantz and Loeb, where they measured what they called community identification (regional identity) and community consumer ethnocentric tendencies (consumer regiocentrism) and showed on a sample of students that greater levels of regional identity lead to greater levels of consumer regiocentrism and that those consumers who demonstrated greater levels of consumer regiocentrism had a tendency to express preference for locally manufactured products (Lantz and Loeb, 1998). Similarly, van Ittersum (2002, p. 93-94) found that consumer regiocentrism negatively influenced consumers´ intention to purchase products from other regions and positively affected their intention to purchase from their own region of residence. In 2011 a qualitative study from Poland on a small sample of students was published, where they showed importance of regional identity and indicated that it should play a much more important role in the future research on consumer ethnocentrism (Siemieniako et al., 2011). There are two studies from Spain, both published in 2013 (Bernabeu et al., 2013; Fernandez-Ferrin and Bande-Vilela, 2013). In paper from Bernabeu et al. (2013) they identified low levels of consumer regiocentrism and did not link it to wine purchasing behavior directly, however the most regiocentric consumers from Barcelona tended to prefer regional wines. The second study (Fernandez-Ferrin and Bande-Vilela, 2013) demonstrated that consumer regiocentrism significantly influenced preference for regional products. Results showed that as consumer regiocentrism increases, so do tendencies to avoid products from outside the region and to prefer regional products, and also that regional identity is an important antecedent of consumer regiocentrism. Final study was done in Australia in 2016 (Lee et al., 2016) confirming similar findings as other studies but also indicating that for some of the low necessity products (golf clubs, jewellery and chocholate; but not for wine), consumers with high levels of regiocentric tendencies were more likely to buy products from their own region. Our study was done with an attempt to understand the impact that the two different socio-psychological factors have directly on the consumers purchasing behavior. In a recent review on consumer perceptions and preferences for local food (Feldmann and Hamm, 2015) it was poited out that in the context of local food research there is an evidence of a gap between attitudes and behavior: while attitudes determine consumers´intentions to purchase local food, these intentions might deviate from actual behavior, as there is a difference in consumers´stated behavior and true behavior. In previous studies on effect of consumer regiocentrism and/or regional identity (Lantz and Loeb, 1998; Fernandez-Ferrin and Bande-Vilela, 2013; Lee et al., 2016) it was shown that consumer regiocentrism is positively related to preference or willingnesss to buy products from own region and that regional identity is an important antecedent of consumer regiocentrism. None of the studies looked at the effect of both constructs independently on actual purchasing behavior. Study from Zeugner Roth et al. (2015) done on a national level showed that national identity is a stronger predictor of product judgement and willingness to buy than consumer ethnocentrism however the influence of both constructs directly on purchasing behavior in a regional setting has not been explored yet. To add additional level of understanding into how the two constructs independently affect regional agricultural product purchasing behavior we chose a product which is very typical for the two regions which we were studying: wine. Wine-making is one of agricultural segments with biggest development potential in Slovenia

(Kuhar, 2011). Annual production is between 80 to 100 million litres of wine produced in little over 16.000 hectares of vineyards. 63 % of produced wine is white and 27 % is red, the rest is rose and sparkling wine. 83 % of wineyards are located in Podravje and Primorska wine regions, the rest is in the smallest Slovenian wine region Posavje (Zagorc in sod., 2016). Besides being in diferent wine zones (“Council regulation (EC) No 479/2008 on the common organisation of the market in wine”) and consequently producing diferent wines by variety and style these two regions are historically and culturally very different. Therefor their selection as the study regions was appropriate in order to achieve the research objectives: to understand how regional identity and consumer regiocentrism independently predict wine purchasing behavior in two main Slovenian wine regions. Due to the fact, that there are many differences between the two studied wine regions and their wines we also wanted to understand, what is the wine-specific regional image from the perspective of regional consumers; and to understand how it affects wine purchasing behavior. Van Ittersum et al. (2003) demonstrated that region-of-origin has product-specific influence on product preference and that product preference is a function of product attribute perception. They also found that product perception is significantly influenced by the perceived product-specific regional image, which was confirmed by Lorenz et al. (2015). With our study we wanted to understand if product-specific (in this case wine-specific) regional image has effect not only on product preference but also on consumers purchasing behavior. In light of the above theoretical background, our research hypotheses were the following: H1: In the perception of Slovenian wine consumers their home wine region has significantly higher wine-specific regional image than the other studied wine region. H2: Wine-specific regional image significantly affects wine purchasing behaviour. H3: More regiocentric consumers buy statistically significant more wine from their home wine region. H4: Consumers with higher degree of regional identity will buy significantly more wine from their wine region. H5: Regional identity has a stronger positive impact on consumers wine purchasing behaviour than consumer regiocentrism. The paper is structured as follows: first we explain the methodology in detail, including the development of the scale to measure wine-specific regional image. Results including validation of scales to measure consumer regiocentrism and regional identity and discussion with regard to other published studies are described in the last part of the paper, which is ending with limitations of the study. Material and Methods Data collection First part of the reseach was of qualitative nature done by means of focus groups. Test focus group with 7 participants was done in June 2011 in order to prepare questions for the other focus groups in the language understandable to average wine drinker. In July 2011 three focus groups were done with in total 20 Slovenian wine consumers. Each focus group consisted of participants from different adult socio-economic status groups. Main purpose was to get a set of typical attributes describing perception of Slovenian wine regions with the purpose to prepare measurement tool to assess wine-specific regional image. Through non-structured leaded discussion participants were asked to describe with as many words as possible characteristics of Slovenian wine regions and wines from different regions. Most common attributes were then selected and included in the quantitative questionnaire. The data for the quantitative part of the study were gathered in July and August 2016 from 221 wine consumers from two Slovenian wine regions, namely 109 from Primorska and 112 from Podravje wine region. After pre-testing of the questionnaire data were gathered through an online questionnaire using 1KA online software (Univeristy of Ljubljana, Faculty of social sciences) with convenience sampling, strictly following equal distribution by age, gender and place of residence. There were four inclusion questions in the beginning of the questionaire: respondents had to be born and resident in one of the two studied wine regions, had to be above 18 years of age, had to drink wine at least once per month and had to buy wine at least twice per year. Proffesional winemakers and sommeliers were excluded with one exclusion question that followed inclusion questions part. After initial set of above mentioned inclusion and exclusion questions, the questionnaire was divided into 4 sections: wine purchasing and consumption habits, wine-specific regional image, consumer regiocentrism and regional identity measures and socio-economic part. Response rate of contacted respondents was 48 %. Variable measurement Closed questions with one possible answer were used to assess wine purchasing, consumption habits and socio-demographics. Focus groups were valuable instrument for the formulation of wine purchasing and consumption habits questions. For the assessment of socio-demographics we used standard questions. To measure wine-specific regional image a list of eight wine-specific image attributes was used, where respondents needed to choose one wine region (from the two studied) for which each of the listed attributes is more typical. We tested observed frequencies from respondents of two wine regions above the expected 50:50 percent ratio to see how the frequencies accross wine regions differ from the neutral perception. To evaluate

wine-specific regional image from the perspective of regional consumers we checked if proportion of answers for one or another wine region is statistically significant which was performed with one-way chi-square test. We analysed differences in perception from consumers from the two wine regions and compared it to likelihood ratio between place of residence and region of most commonly purchased wine. Of particular interest in our study was to understand the relevance of consumer regioncentrism and regional identity in wine purchasing decisions. Seven point Likert scale was used to measure intensity of both constructs. To measure consumer regiocentrism a shortened version of CETSCALE was used, which was previously used in a regional context by Fernandez-Ferrin and Bande-Vilela (2013) as a five-item scale, adapted from Klein et al. (2006), where it was used and validated as a six-item scale in a national setting. Regional identity was initially measured in a sub-national context with a four-item version by Lantz and Loeb (1998). They named it community or local identity and derived and adapted the measurement instument from Luthanen and Crocker’s collective self-esteem scale (1992) and Hawes and Lumpkin (1984) paper on consumers shopping outside of their place of residence (outshoppers) purchasing behavior. Regional identity scale was adapted to 4-item scale by Fernandez-Ferrin and Bande-Vilela (2013) and this was the scale we used in our study. Pearson correlation coefficient was used as a tool to evaluate the connection between consumer regiocentrism and regional identity scores and proportion of wine bought from different wine regions in the last year. Results and Discussion Descriptive analysis Total sample consisted of 53.4 % male and 46.6 % female respondents. Average age of the respondent was 40.6 years, with most of them having at least secondary school education. 78.9 % of respondents were coming from small town or village. 21.1 % had below, 17 % above and the rest average monthly income relative to the national average. 24,4 % of respondents drank wine two to three times per month, one third once per week and one third more than once per week. Majority of respondents (81,5 %) purchased wine two to three times per month or more often. More than half of repondents (52,9 %) usually drank white wine, 44,3 % red and a small percentage sparkling or rose wine. 47,7 % of respondents most commonly purchased wine at the supermarket store and 35 % at winemakers. Usual price of purchased wine was eight euros or below, highest number of respondents (35,3 %) usually bought wine for the price between three and five euros. A comparison of socio-demographic characteristics of the repondents from the two wine regions indicates that samples were similar in relation to gender distribution, employment status and level of education. Some socio-demographic differences occured due to demographic differences between the compared wine regions (table 1). Table 1: Socio-demographic characteristics of the respondents:

Wine region of residence Primorska (n=109) Podravje (n=112) Total (n=221)

Count Column

N % Count Column

N % Count Column

N % Gender male 62 56.9 56 50.0 118 53.4

female 47 43.1 56 50.0 103 46.6 Employment status student 8 7.3 13 11.6 21 9.5

unemployed 8 7.3 11 9.8 19 8.6 employed 81 74.3 78 69.6 159 71.9 retired 12 11.0 10 8.9 22 10.0

Partnership status single 17 15.6 13 11.6 30 13.6 with partner or relatives* 32 29.4 52 46.4 84 38.0 alone with child/ren 4 3.7 8 7.1 12 5.4 with partner and child/ren* 56 51.4 39 34.8 95 43.0

Size of place of residence town (>100.000 inhabitants)* 9 8.4 37 33.3 46 21.1 small town (10.000-100.000 inhabitants)* 34 31.8 20 18.0 54 24.8

village (< 10.000) 64 59.8 54 48.6 118 54.1 Education High school and below 41 37.6 48 42.9 89 40.3

Graduate degree 30 27.5 26 23.2 56 25.3 Post-graduate degree 38 34.9 38 33.9 76 34.4

Income relative to national average

below average* 14 13.1 32 28.8 46 21.1 average 73 68.2 62 55.9 135 61.9

above average 20 18.7 17 15.3 37 17.0 Note: *= significantly different at p< .05 in the two-sided test of equality for column proportions. Tests assume equal variances.1 1. Tests are adjusted for all pairwise comparisons within a row of each innermost subtable using the Bonferroni correction.

Validity and reliability of consumer regiocentrism and regional identity scales Validity and reliability of both scales was confirmed with factor analysis. Reliability of the two constructs was assesed using cronbach´s alpha and composite reliability measures. The recommended threshold for a sufficient composite reliability (CR) is 0.7 or above (Nunnally and Bernstein, 1994). Cronbach alpha and CR values on both factors were well above 0.7 (Table 2). After establishing construct reliability, convergent validity and discriminant validity were assessed to test the relationship between the constructs (Vida and Reardon, 2008). Chi square was significant, but since it is sensitive to sample size, we also checked other model fit indices: RMSEA and SRMR, which should be below 0.08 cut off; and NFI, NNFI, CFI, IFI indices which should be above 0.90 limit. The measurement did not fit the model well so we had to modify it by removing two indicators: »I am proud to be a member of my region« and »Consumers from our region that purchase products made in other regions are responsible for the losses of jobs of our regional fellows« and add one error covariance between two similar worded indicators. Then the model conformed well to the data (χ2 (12) =23.53; p<0.024; RMSEA=0.07; NFI=0.98; NNFI=0.98; CFI=0.99; IFI=0.99; SRMR=0.05). Convergent validity was assesed through average variance extracted (AVE) which should be above 0.5 treshold (Fornell and Larcker, 1981) and was satisfactory in our study. Discriminant validity was demonstrated by correlations between factors, which should according to Ping (2004) not exceed 0.7. The average value of indicators that measure the same factor was calculated and these two variables were used for futher analysis. The measured items and their means with standard deviation, Crombach´s alpha, CR and percent of total variance are shown in table 2. Standardized path coefficients and AVE are presented in Table 3. Confirmatory factor analysis was done in AMOS IBM program by robust maximum likelihood method. Table 2: Measurement scale characteristics of constructs (n=221)

Construct and items Overal mean SD Cronbach´s α CR

Percent of total

variance explained

Regional identity 0,95 0,97 38,3 I am happy to form a part of my region. 6,02 1,40 I strongly identify with my region. 5,71 1,54 I am very commited to my region. 5,68 1,53 I am proud to be a member of my region. 6,13 1,32 Regiocentrism

0,86 0,90 31,4 Products from outside our region should only be bought when regional equivalents are not available 3,16 1,85 Our regional products come first and foremost 4,97 1,80 A true member of our region should always buy products made in our region 4,01 1,95 Member of our region should not buy products from outside the region since it harms our companies and causes unemployment

3,46 1,95

Consumers from our region that purchase products made in other regions are responsible for the losses of jobs of our regional fellows

2,62 1,76

SD = standard deviation; CR = composite reliability Table 3: Scale items with standardized path coefficients and AVE (n=221)

Constructs and items Standardized

path coefficients

AVE

Regional identity 0,91

I strongly identify with my region. 0,92

I am very commited to my region. 0,95

I am proud to be a member of my region. 0,87

Regiocentrism 0,70

Products from outside our region should only be bought when regional equivalents are not available. 0,65

Our regional products come first and foremost. 0,72

A true member of our region should always buy products made in our region. 0,95

Member of our region should not buy products from outside the region since it harms our companies and causes unemployment. 0,69 AVE=Average variance extracted Wine-specific regional image Independently of region of residence, in minds of Slovenian wine consumers from both wine regions (total sample) all attributes to measure wine-specific regional image, except for »In this wine region wine is present everywhere« (where there was no difference between the two regions) were related more to Primorska wine region with statistical significance. According to listed statements we can say that wine region Primorska has a much higher overall wine-specific regional image than Podravje. Looking at each of the samples independently, more than 75 % of Primorska respondents perceived all the wine-specific image indicators were more typical for their home wine region (figure 1). On the other hand Podravje respondents did not perceive that there are any differences in five out of eight wine-specific regional image descriptors between the two wine regions. At the same time they perceived that two wine-specific image attributes are significantly more related to the other (Primorska) wine region (“in this wine region wine is closely connected with tourism” and “in this wine region there are many well known and recongized winemakers”) and the only image descriptor they believed was significantly more related to their home wine region was “in this wine region you feel that winemakers cooperate well” (figure 2). With H1 we proposed that in the perception of Slovenian wine consumers their home wine region has significantly higher wine-specific regional image than the other studied wine region, which can be partly accepted as it was confirmed only for one (Primorska), but not for the other wine region (Podravje). Figure 1: Wine-specific regional image from the perspective of Primorska respondents

Figure 2: Wine-specific regional image from the perspective of Podravje respondents

With H2 we proposed that wine-specific regional image significantly affects wine purchasing behavior. Looking at the likelihood ratio between region of residence and wine region from which consumers most often purchased wines, we saw that in the last year Primorska respondents bought 92.7 % of all wine from their home wine region and less than 5 % from region Podravje. Podravje respondents bought 61.6 % of wine from their home wine region, 33 % from Primorska and less than 5 % from elsewhere. Association is statistically significant (p < 0.001), meaning that consumers from both wine regions prefer to buy wine from their home wine region, despite the fact that wine-specific regional image is higher for Primorska for both samples (table 4). H2 can be only partly accepted (only for respondents from wine region Primorska). Table 4: Association between region of residence and wine region from where majority of wine was bought

Consumer regiocentrism and regional identity scores In the total sample in our study the average score for consumer regiocentrism was 3,9 and average regional identity score was 5,8 on a seven point Likert scale. The first study measuring consumer regiocentrism and regional identity on a sample of Canadian students from 1998 (Lantz and Loeb) demonstrated average consumer regiocentrism score 5.2 and average regional identity score 4. In study by Barnabeu et al. (2013) average scores for consumer regiocentrism were 2.3 for Madrid and 2.6 for Barcelona samples on a seven point Likert scale. There are two studies published from Slovenia on a national level where they measured consumer ethnocentrism. Average score in the first study was 2.8 on a five point Likert scale (Vida and Maher Pirc, 2006) and 3.5 on a seven point Likert scale in the second study (Zeugner-Roth et al., 2015). In the latter study they also measured national identity and average score was 6.1. Interestingly, average national identity score is higher than the one we measured for regional identity, which could be due to sample differences, where in the first study 58 % of people were from urban environment and only 21 % in our study. This is somehow unexpected result considering abovementioned theory on the subject, and also according to Kucan (1996) who wrote that each Slovene person is more or less attached to their specific region, which they like more than any other region and

Primorska PodravjePrimorska f 101 37

f% 92,7% 33,0%Podravje f 5 69

f% 4,6% 61,6%Posavje f 2 2

f% 1,8% 1,8%Foreign wine f 1 4

f% 0,9% 3,6%

* f = freqvency; f% = percentage; LR = testn statistics, likelihood ratio; Df = degrees of freedom

LR Df P - value

Wine region from which majority of wine was

purchased in the last year98,72 3 < 0,001

REGION OF RESIDENCE

which is a part of their personal identity. Furthermore, Slovenian regions are continuously present in people´s minds, on a cognitive level as well as on a symbolic identity level (Polic et al., 2005). Siemieniako et al. (2011) found high importance of regional versus national identity as image of Polish brand was heavily relying on respondents´ regional connections and associations, however quantitative research was not performed. Comparing average scores for consumer regiocentrism and regional identity between the respondents of two Slovenian wine regions it was found that average scores for regional identity are higher than those for consumer regiocentrism. Both constructs had higher scores in Primorska region, but not significantly different between the two studied samples (table 5). National identity is primarily measuring the predictive power of one pro-in-group while consumer ethnocentrism is also looking at one anti-out-group (Zeugner-Roth et al., 2015). Same principal can be applied to a regional level and by our results we could imply that Slovene people are more attached to their own region as they are positioned against other regions, which was an expected outcome. Table 5: RI and CR scores in two studied wine region samples (n=221).

n M SD t df P-value

Regional identity Primorska 109 5.9 1.3 0.8 219 0.4

Podravje 112 5.8 1.4

Consumer regiocentrism Primorska 109 4.0 1.5 1.0 219 0.3

Podravje 112 3.8 1.6

M = median; SD = standard deviation; n = number of respondents; df = degrees of freedom

Further we analysed how rates of regional identity and consumer regiocentrism affect wine purchasing behavior. Pearson´s correlation coefficient was used as a tool to evaluate the connection between consumer regiocentrism and regional identity scores and proportion of wine bought in the last year from different wine regions. Very interestingly we observed no statistically significant correlation between regional identity and consumer regiocentrism and proportion of purchased wine with respondents from Primorska wine region. With respondents from Podravje wine region we observed weak positive statistically significant correlation between regional identity, consumer regiocentrism and proportion of wine purchased from their home wine region (table 6). This means that consumers from Podravje, who are more regiocentric and more attached to their region, buy more wine from their own wine region. With H3 we anticipated that more regiocentric consumers buy statistically significant more wine from their home wine region and with H4 we proposed that consumers with higher degree of regional identity will buy significantly more wine from their wine region. Both hypotheses are only partly accepted as they are only confirmed for sample from wine region Podravje and not for the sample from wine region Primorska. At the same time, we observed that people from Podravje region who have higher scores of regional identity (but not consumer regiocentrism) buy significantly less wine from wine region Posavje (r = -0.19; p = 0.045), which implies that regional identity might be a better predictor of purchasing behavior than consumer regiocentrism, therefore H5 can be accepted. No effect of consumer regiocentrism and regional identity on wine purchasing decisions in the sample from wine region Primorska is an interesting finding, as in most before mentioned studies measuring consumer ethno- or regiocentrism it was demonstrated that higher rates relate to preference/willingness to buy domestic vs. foreign (regional vs. non-regional) products. People from Primorska wine region in the last year purchased 92.7 % of wine from their region independently of lelevs of consumer regiocentrism and regional identity, but most probably due to the fact, that wine-specific regional image of their home wine region is so positive. Table 6: Pearsons correlation coefficient between percentage of purchased wine and RI and CR scores (n=221).

Region of origin of purchased wine

Living in Primorska wine region

Living in Podravje wine region

RI CR RI CR

Primorska wine region

r 0.03 -0.04 -0.04 -0.12 P-

value 0.740 0.689 0.651 0.212

n 109 109 112 112

Podravje wine region

r 0.02 -0.14 0.19 0.23 P-

value 0.842 0.165 0.041* 0.014*

n 96 96 112 112

Posavje wine region r -0.04 0.04 -0.19 -0.06

P-value 0.673 0.723 0.045* 0.520

n 94 94 108 108

Foreign wine

r -0.05 0.05 -0.13 -0.16 P-

value 0.649 0.599 0.188 0.091

n 96 96 107 107 Note: *= significantly different at p< .05 r = Pearsons correlation coefficient; n = number of respondents, RI=regional identity, CR=consumer regiocentrism

Conslusions and implications The main objective of this study was to gain more insights into importance of consumer regiocentrism and regional identity and to understand the role of wine-specific regional image in wine purchasing decisions from the perspective of regional wine consumers. Our findings confirm strong attachment of Slovenian people to their region, which was confirmed with high regional identity scores. Kucan (1996) is describing that each of Slovenians is especially attached to a specific region or even smaller entity inside the region which is a part of a person´s identity. Closely connected with regional identity, initially a sociological concept, which is distinguishing between in-groups (the groups with which an individual identifies) and out-groups (the others), consumer regiocentrism is a construct demontrating how strongly an individual feels morally obliged to purchase products from his region and avoid products from other regions in order for his region to prosper. Results of our study show that Slovenian people from two different wine regions differ in terms of how consumer regiocentrism and regional identity affect their wine purchasing behavior. Despite similar levels of consumer regiocentrism and regional identity in both samples, we demonstrated that only in one wine region (Podravje) both constructs affect wine purchasing behavior; which is not the case for the other (Primorska) wine region. Importantly we also found that regional identity is a better predictor of wine purchasing behavior than consumer regiocentrism. Our findings are somehow in contrary to findings of Lantz and Loeb (1998), van Ittersum (2002), Fernandez-Ferrin and Bande-Vilela (2013) and Lee et al. (2016) which were studying consumer attitudes in a regional setting and found that greater levels of consumer regiocentrism result in tendency to express preference for locally manufactured products. The difference could be due to the fact that we were looking specifically at purchasing behavior and not only at preference or purchase intention. According to Feldmann and Hamm (2015) there is an evidence of a gap between attitudes and behavior in the context of local food research. Our study confirms their findings: while attitudes determine consumers´intentions to purchase local food, these intentions might deviate from actual behavior, as there is a difference in consumers´stated behavior and true behavior. Evaluating the wine-specific regional image in our study showed that it is closely linked to preference, but not necesarilly to purchasing bahavior. Despite the fact that in the Podravje sample wine region Primorska has a very high wine-specific regional image, they still purchase majority of the wine from their home wine region. Van Ittersum et al. (2003) found that region-of-origin has product-specific influence on product preference and that product preference is a function of product attribute perception. Further studies would be necessary to demonstrate effect of consumer regiocentrism, regional identity and product-specific regional image on consumers purchasing behavior. From marketing perspective the studies on consumer regiocentrism and regional identity are relevant for application of marketing strategies to consumers living inside the studied regions and therefor largely depend on the size of the region. On the other hand product-specific regional image is more general, not related only to residents of the region, but also to other consumers which are familiar with the specific region. According to van Ittersum (2003) when a product and a region match, the regional image seems to influence the evaluation of the product by consumers positively, and regional indication works similarly to a brand name. At the same time, using a regional indication has advantages over introducing a new brand, as creation of the value of a brand name is based on a set of associations which is a long and difficult process which might take years; while in case of region-of-origin, these associacions do not have to be created because most of the consumers already have some associations with the region (Aaker, 1991). Our study proposes that even in a small country like Slovenia there are important differences between consumers of two different wine regions therefor marketing approach should be adapted to a regional level. Only consumers from Podravje, but not Primorska wine region who are more regiocentric and attached to their region will buy more regional wine. Importantly and independently of wine-specific regional image and levels of consumer regiocentrism and regional identity, consumers from both wine regions purchase majority of wine from their home wine region. This could be due to fact that residents of both regions in general have high consumer regicentrism and regional identity scores. It should not be neglected that consumers from Podravje purchase a substantial percentage of wine (more than one third) from Primorska wine region which seem to be the consequence of very positive perception of wine-specific regional image of Primorska. Considering other Slovenian population where measurements of consumer regiocentrism and

regional identity is not applicable (non-residents), building a high wine-specific regional image would be important. Studying connection between product-specific regional image and purchasing bahavior of other products would give more insight into importance of product-specific regional image in the purchasing decisions, where it is important to note that the product and the region should match (van Ittersum, 2003). Most importantly we demonstrated that there is a clear gap between perception and actual purchasing behavior therefore future studies should further look into effect of different constructs on actual purchasing bahavior, not only on preference, product judgment or purchase intention. Limitations Limitation of the study is that it was done only in two wine regions, which was necessary to assess consumer regiocentrism and regional identity; and therefor did not capture the view of the national population, especially the consumers from the capitol, where socio-demographics, wine purchasing and consumption habits might be different. To measure wine-specific regional image respondents needed to choose one of the two studied wine regions, for which given wine-specific image attributes were more typical, which was done due to the length of the questionnaire. In future research, we would suggest to use Likert scale and evaluate intensity for the studied regions which would allow more comprehensive statistical analyses. With use of structural modelling we could also look at the relationships between consumer regiocentrism, regional identity, product-specific regional image and other related constructs. We would also suggest to use more indicators in the scales to measure consumer regiocentrism and regional identity because in the process of factor analysis some of the indicators might need to be removed as in our case. References Aaker D.A. (1991): Managing brand equity: capitalizing on the value of a brand name. The Free Press, New York. Bernabeu R., Prieto A., Diaz M. (2013): Preference patterns for wine consumption in Spain depending on the degree of consumer ethnocentrism. Food Quality and Preference, 28: 77-84. Bianchi C., Mortimer G. (2015): Drivers of local food consumption: a comparative study. British Food Journal, 117, 9: 2282-2299. Brewer M.B., Manzi J.M., Shaw J.S. (1993): In-group identification as a function of depersonalization, distinctiveness, and status. Psychological science, 4, 2: 88-92. Feldmann C., Hamm U. (2015): Consumers´perceptions and preferences for local food: A review. Food Quality and Preference, 40: 152-164. Fernandez-Ferrin P., Bande-Vilela B. (2013): Regional ethnocentrism: Antecedents, consequences, and moderating effects. Food Quality and Preference, 30: 299-308. Fornell, C., Larcker, D. F. (1981): Structural equation models with unobservable variables and measurement error: Algebra and statistics. Journal of marketing research, 18, 3: 382-388. Halldorson J.D. (2009): An Exploration of Tajfel´s Social Identity Theory and its Application to Understanding Metis as a Social Identity. Master of social work thesis. Faculty of social work, university of Manitoba. Hawes J.M., Lumpkin J.R. (1984): Understanding the Outshopper. Academy of Marketing Science, 12, 4: 200-218.

Kucan A. (1996): Slovenian landscapes: Aspects of their perception in relation to national identity. Disertation thesis. Univerza v Ljubljani, Biotehniška fakulteta, Inštitut za krajinsko arhitekturo, Ljubljana (In Slovenian). Kuhar A. (2011): Slovenija na svetovnem vinskem zemljevidu. Revija Vino, Letnik IX, 1: 20-21 (In Slovenian).

Lantz G., Loeb S. (1998): An Examination of the Community Identity and Purchase Preferences Using the Social Identity Approach. Advances in Consumer Research, 25: 486-491. Lee W.J.(T.), Cheah I., Phau I., Teah M., Elenein B.A. (2016): Conceptualising consumer regiocentrism: Examining consumers´ willingness to buy products from their own region. Journal of Retailing and Consumer Services, 33: 78-85. Lindquist J.D., Vida I., Plank R.E., Fairhurst A. (2001): The modified CETSCALE: validity tests in the Czech Republic, Hungary, and Poland. International Business Review, 10: 505-516. Lorenz B.A., Hartmann M., Simons J. (2015): Impacts from region-of-origin labelling on consumer product perception and purchasing intention – Causal relationships in a TPB based model. Food Quality and Preference, 45: 149-157. Luthanen R., Crocker J. (1992): A Collective Self-Esteem Scale: Self-Evaluation of One´s Social Identity, 18, 3: 302-318. Nunnally, J. C., Bernstein, I. H. (1994): The assessment of reliability. Psychometric theory, 3, 1: 248-292. Orth U.R., Firbasova Z. (2003): The Role of Consumer Ethnocentrism in Food Product Evaluation. Agribusiness, 19, 2: 137-153. Ping, R.A. Jr (2004): On assuring valid measures for theoretical models using survey data. Journal of Business Research, 57: 125-141. Polic M., Repovs G., Natek K., Klemencic M., Kos D., Ule M., Marusic I., Kucan A. (2005): A cognitive map of Slovenia: Perceptions of the regions. International Journal of Psychology, 40 (1): 27-35. Sharma S., Shimp T.S., Shin J. (1995): Consumer Ethnocentrism: A test of Antecedents and Moderators. Journal of the Academy of Marketing Science, 23, 1: 26-37. Shimp T.A., Sharma S. (1987): Consumer Ethnocentrism: Construction and Validation of the CETSCALE. Journal of Marketing Research, 24, 3: 280-289. Siemieniako D., Kubacki K., Glinska E., Krot K. (2011): National and regional ethnocentrism: a case study of beer consumers in Poland. British Food Journal, 113, 3: 404-418. Simon B., Kulla C., Zobel M. (1995): On being more than just a part of the whole: regional identity and social distinctiveness. European Journal of Social Psychology, 25: 325-340. Sumner G.A. (1906): Folkways: The Sociological Importance of Usages, Manners, Customs, Mores, and Morals. Ginn Custom publishing, New York. Tajfel H. (1981): Human groups and social categories. Cambridge University Press, Cambridge, England. Tajfel H., Turner C. (1986): The social identity theory of inter-group behaviour, Psychology of intergroup relations. Nelson-Hall, Chicago. Thode, S.F., Maskulka, J.M. (1998): Place-based marketing strategies, brand equity and vineyard evaluation. Journal of Product and Brand Management, 7, 5: 379–399. Van Ittersum K. (2002): The role of region of origin in consumer decision-making and choice. Doctoral dissertation, Mansholt Graduate School, Wageningen, The Netherlands. Van Ittersum K., Candel M.J.J.M., Meulenberg T.G. (2003): The influence of the image of a product´s region of origin on product evaluation. Journal of Business Research, 56: 215-226. Van Ittersum K., Meulenberg M.T.G., Van Trijp H.C.M., Candel M.J.J.M. (2007): Consumers´Appreciation of Regional Certification Labels: A Pan-European Study. Journal of Agricultural Economics, 58, 1: 1-23.

Verlegh P.W.J., Steenkamp J.-B.E.M. (1999): A review and meta-analysis of country-of-origin research. Journal of Economic Psychology, 20: 521-546. Vida I., Fairhurst A. (1999): Factors underlying the phenomenon of consumer ethnocentricity: evidence from four central European countries. The International Review of Retail, Distribution and Consumer Research, 9, 4: 321-337. Vida I., Maher Pirc M. (2006): Nakupno vedenje slovenskih porabnikov: vloga nacionalne identitete. Management, 1, 1: 49-63 (in Slovenian). Vida I., Reardon J. (2008): Domestic consumption: rational, affective or normative choice? Journal of Consumer Marketing, 25, 1: 34-44. Vida I., Dmitrovic T., Obadia C. (2008): The role of ethnic affiliation in consumer ethnocentrism. European Journal of Marketing, 42, 3: 327-343. Zagorc B., Moljk B., Pintar M., Brečko J. 2016. Poročilo o stanju kmetijstva, živilstva, gozdarstva in ribištva. Pregled po kmetijskih trgih. Ljubljana, Kmetijski inštitut Slovenije, Ministrstvo za kmetijstvo, gozdarstvo in prehrano: 132 str. http://www.kis.si/f/docs/Porocila_o_stanju_v_kmetijstvu_OEK/ZP_2016_trgi.pdf (11. jan. 2018) Zeugner-Roth K.P., Zabkar V., Diamantopoulos A. (2015): Consumer ethnocentrism, national identity, and consumer cosmopolitanism as drivers of consumer behaviour: A social identity theory perspective. Journal of International Marketing, 23, 2: 25-54.