General image and attribute perceptions of traditional food in six European countries

Food Quality and Preference 22 (2011) 129–138

Contents lists available at ScienceDirect

Food Quality and Preference journal homepage: www.elsevier.com/locate/foodqual

General image and attribute perceptions of traditional food in six European countries Valérie Lengard Almli a,b,⇑, Wim Verbeke c, Filiep Vanhonacker c, Tormod Næs a, Margrethe Hersleth a a

Nofima Mat AS, Osloveien 1, N-1430 Ås, Norway The Norwegian University of Life Sciences, Department of Chemistry, Biotechnology and Food Science, Box 5003, N-1432 Ås, Norway c Ghent University, Department of Agricultural Economics, Coupure Links 653, 9000 Gent, Belgium b

a r t i c l e

i n f o

Article history: Received 22 February 2010 Received in revised form 3 June 2010 Accepted 10 August 2010 Available online 13 August 2010 Keywords: Traditional food Image Attribute perception Survey Cross-cultural Partial Least Squares Regression (PLSR) Principal Component Analysis (PCA)

a b s t r a c t This paper studies the image of traditional food at pan-European and national levels in six countries: Belgium, France, Italy, Norway, Poland and Spain. A consumer survey about traditional food products (TFP) was conducted. The respondents (n = 4828) indicated their personal general opinion and feelings about TFP and characterised TFP according to 15 intrinsic and extrinsic product attributes. Traditional food is found to have a unanimously positive general image across Europe. The main patterns of product attribute perceptions are coherent in the six countries. The results show that European consumers trade-off the relative expensiveness and time-consuming preparation of traditional food for the specific taste, quality, appearance, nutritional value, healthiness and safety they find in TFP. Further, the general image of TFP relates to attributes typical of a festive consumption of traditional foods rather than a consumption of daily character. The implications of our findings for the future market of traditional food are discussed. Ó 2010 Elsevier Ltd. All rights reserved.

1. Introduction Product image has a powerful influence on consumer behaviour. Studies on brand image have shown that brand exposure impacts on our product choice behaviour (Guinard, Uotani, & Schlich, 2001; Pohjanheimo & Sandell, 2009) and even on our overall behaviour (Fitzsimons, Chartrand, & Fitzsimons, 2008). Marketers have long exploited the power of brand image to sell their products, and there is an emergence of marketing efforts focusing on whole product categories, such as campaigns inviting to a daily consumption of fruits and vegetables, for example in Norway and France. A positive image creates positive consumer expectations, which in turn may lead to product purchase. According to Deliza and MacFie (1996), it is necessary that consumers’ expectations are met upon product consumption to yield satisfaction and ensure re-purchase of the product. If consumers’ expectations fail to be met, a negative disconfirmation occurs and the product might never be purchased again. Consequently, in order to be successful on the market a product or product category needs to both benefit from a positive general image and offer product qualities that match or surpass consumers’ expectations. These two pieces of information, i.e., image and ⇑ Corresponding author at: Nofima Mat AS, Osloveien 1, N-1430 Ås, Norway. Tel.: +47 64 97 03 05; fax: +47 64 97 03 33. E-mail address: valerie.almli@nofima.no (V.L. Almli). 0950-3293/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.foodqual.2010.08.008

perception of product attributes, are therefore important to collect and study. This paper investigates the general image of traditional food products (TFP). TFP have per definition a long history on the food market, and accordingly every consumer has a relationship to this product category. Published definitions of traditional foods include temporal, territorial and cultural dimensions (Bertozzi, 1998; EU, 2006; Jordana, 2000), the idea of a transmission from generation to generation (EU, 2006; Trichopoulou, Soukara, & Vasilopoulou, 2007) and, more recently, elaborative statements about traditional ingredients, traditional composition and traditional production and/or processing (Weichselbaum, Benelam, & Soares Costa, 2009). These definitions take source in food professionals’ perspectives, but do not necessarily correspond to consumers’ envision of traditional food. Consumer studies on traditional food have recently been conducted in order to define and characterise the concept of TFP seen from a European consumers’ perspective (Guerrero et al., 2009, 2010; Vanhonacker, Verbeke, et al., 2010). According to Vanhonacker, Verbeke, et al. (2010), European consumers define traditional foods as ‘‘frequently consumed or associated to specific celebrations and/or seasons, transmitted from one generation to another, made in a specific way according to the gastronomic heritage, naturally processed, distinguished and known because of their sensory properties and associated to a certain local area, region or country”. This definition reflects the broadness of TFP as

130

V.L. Almli et al. / Food Quality and Preference 22 (2011) 129–138

a concept, and a large amount of subjectivity is shown in European consumers’ definition of TFP (Vanhonacker, Verbeke, et al., 2010). Further, when investigating the association between traditional food consumption and motives for food choice in Europe, Pieniak, Verbeke, Vanhonacker, Guerrero, and Hersleth (2009) found healthiness and convenience constructs to have significant negative associations with traditional food consumption, while familiarity and natural content constructs showed positive significant associations. These results corroborate Vanhonacker, Lengard, Hersleth, and Verbeke (2010) whom in their work of profiling European traditional food consumers found that the attitudes and behaviours of traditional food consumers significantly differed from those of non-traditional food consumers in diverse domains including health, cooking, food shopping, convenience and ethnocentrism. These publications give information about European consumers’ definition of, associations to and consumption attitudes for TFP, but they neither give information about the standing and the reputation of TFP in the food market, nor about the consumers’ perceptions of TFP attributes. The purpose of this paper is to reveal the general image of traditional food in six geographically spread European countries, and to explore consumers’ perceptions of specific product attributes of TFP covering a selection of sensory, health, ethics, purchasing and convenience factors. Correlations between the overall image of TFP and the attributes perceptions will be examined in a multivariate approach. By mapping consumers’ perception and overall evaluation of TFP, insight will be gained on consumers’ expectations and intention to purchase this product category, and relevant information will be gathered with respect to future product positioning and marketing communications. To achieve our goal, cross-cultural data collected from Belgium, France, Italy, Norway, Poland and Spain were studied. The selection of countries was done with the aim to cover the geographical North–South and East–West axes of Europe. As these data contain a high amount of collinearity, and as our purpose invites for an exploratory approach, Principal Component Analysis (PCA) and Partial Least Squares Regression (PLSR) multivariate modelling were used. Because these models do not require any hard-modelling inputs from the researcher, such that no assumptions on data relationships are required, they have been applied extensively in sensory and consumer hedonic studies. PLSR in particular is said to be a versatile and cognitive method (Martens, 2001), yet it has remained largely out of the scene when it comes to the analysis of consumer attitudinal and behavioural data (for an exception, see Vanhonacker, Lengard, et al. (2010)), where the stepwise linear regression and PLS Path Modelling (PLS-PM) methods are more of-

ten used (Guinot, Latreille, & Tenenhaus, 2001; Tenenhaus, Vinzi, Chatelin, & Lauro, 2005). A side-objective of this paper will be to illustrate the possibilities of using PCA and PLSR on attitudinal consumer survey data, and possibly to raise an interest in these methods among food consumer researchers. 2. Materials and methods 2.1. Research approach and sampling Quantitative descriptive data were collected in October– November 2007 through a cross-sectional self-administered consumer survey with samples representative for age, gender and region in Norway, Belgium, France, Spain, Italy and Poland. The age range of the population was defined as 20–70 years. The total sample size was 4828 respondents, i.e., around 800 respondents in each of the six considered European countries. Participants were randomly selected from the representative TNS European Online Access Panel (Malhotra & Peterson, 2006) in line with the national population distributions with respect to age, gender and region. Detailed socio-demographic characteristics of the national and pooled samples are provided in Table 1. Gender is equally distributed, which reflects that the population was intentionally not restricted to the main responsible person for food purchasing. Age distributions, mean age and mean household sizes match closely with the national census data of the respective countries. Table 1 also presents an indicator of socio-economic class, which was a subjective assessment of the household’s financial situation (Chittleborough, Baum, Taylor, & Hiller, 2008). The measurement was given on a seven-point interval scale ranging from ‘‘difficult” to ‘‘well off”. Subdivisions were made between respondents who perceived their financial situation as rather unsatisfactory (answering 1, 2 or 3 on the seven-point scale), moderate (answering 4) or satisfactory (answering 5, 6 or 7). Although this subjective measure may not necessarily reflect the actual financial situation or socioeconomic class of the households, its distribution suggests that different socio-economic classes were represented in the sample. The sample is slightly biased towards higher education, which may be attributed to the use of an electronic data collection method. 2.2. Measurement and scaling The questionnaire consisted of five sections: (1) behavioural and attitudinal items relating to food purchasing in general, (2) items probing for consumers’ definition, attitude and perceptions

Table 1 Selected socio-demographic characteristics of the samples. Pooled sample (n = 4828)

Norway (n = 798)

Belgium (n = 826)

France (n = 801)

Spain (n = 800)

Italy (n = 800)

Poland (n = 803)

Gender (%) Female Male

49.2 50.8

49.1 50.9

49.4 50.6

51.9 48.1

47.4 52.6

47.3 52.7

50.2 49.8

Age (years) <35 35–55 >55 Mean S.D.

34.1 46.4 19.5 41.5 12.8

34.1 47.5 18.4 41.4 12.5

28.5 46.4 25.1 43.7 13.3

33.7 46.4 19.9 41.4 12.8

35.5 47.4 17.1 40.7 12.3

35.0 45.8 19.2 41.2 12.8

37.9 44.8 17.3 40.6 12.8

Household size (number) Mean 2.9 S.D. 1.3

2.6 1.3

2.7 1.3

2.7 1.2

3.1 1.3

3.2 1.3

3.0 1.4

24.8 31.5 43.7

17.8 28.6 53.6

35.5 32.5 32.0

18.9 36.2 44.9

29.8 32.8 37.4

21.3 31.0 47.7

Financial situation (%) Difficult – moderate Moderate Moderate – well off

24.6 32.1 43.3

131

V.L. Almli et al. / Food Quality and Preference 22 (2011) 129–138

of traditional food, (3) consumer awareness of and associations with quality and origin labels, (4) items probing for consumer’s definition and acceptance of innovations in traditional food products, and (5) personal data relating to lifestyle and general interests, and socio-demographics. A description of the selection of items and pre-testing of the questionnaire is given in Vanhonacker, Verbeke, et al. (2010). In this paper, attention is brought on two questions from section (2) of the survey, focusing on attitudes toward and perception of traditional food. The first selected question is a direct measurement of the image of traditional food: ‘‘When you think about the image you have of traditional food in general, how would you describe your personal opinion/feelings about it?”. For this measurement, a seven-point Likert scale anchored with ‘‘very negative” to the left, ‘‘neither positive nor negative” at mid-point and ‘‘very positive” to the right was used. The second question of interest relates to the respondents’ perceptions of a series of intrinsic and extrinsic product attributes of traditional food: ‘‘Please indicate to what extent traditional food has the following characteristics according to you”. A set of 15 items with seven-point semantic differential scales was given, with a negative anchor to the left (e.g. Low in quality) and a positive anchor to the right (e.g. High in quality). The full list of attributes is presented in Table 2. The selection of these attributes was based on the results from a qualitative study involving focus groups discussions and word association tests (Guerrero et al., 2009, 2010).

dummy variables (Martens & Martens, 2001), allowing to project the six countries onto the PCA loadings plot. In addition to the pan-European model, similar PCA models were run for each of the six countries separately. Partial Least Squares Regression (PLSR) (Esbensen, 2001; Martens & Martens, 2001) was used to relate the respondents’ attribute perceptions of traditional food to their personal general image of traditional food. Seven models were built in total: one per country and one for the pan-European sample (n = 4765 after removal of cases with missing values). The data were mean-centred along respondents for reasons mentioned above. In each model, the dependent variable is the measurement of general image and the set of independent variables is composed of the 15 product attribute perceptions of TFP given in Table 2, where each attribute is labelled by its right scale anchor. All data were standardised. Country-specific models were cross-validated with 20 random segments (Stone, 1974); the pan-European model was cross-validated with six segments corresponding to the six countries involved. Jack-knife uncertainty testing (Martens & Martens, 2000) was performed on the regression coefficients to detect the attributes significantly related to general image with a 95% confidence interval. In the case of the pan-European model, the cross-validation on six countries combined with uncertainty testing revealed the consensus attributes, that is to say attributes that are applicable across countries. All multivariate models were run in The Unscrambler 9.8 (CAMO Software AS).

2.3. Multivariate analysis

3. Results and discussion

Principal Component Analysis (PCA) (Esbensen, 2001; Martens & Martens, 2001) was used to study the main patterns in attribute perceptions of traditional food. The model was computed over 15 variables corresponding to the 15 product attributes presented to the respondents, labelled with their right anchors (see Table 2). The full sample of European respondents except for five cases with missing values was included in the model (n = 4823). The data were mean-centred along respondents so as to highlight differences in perceptions rather than differences in scale usage (Næs, Lengard, Bølling Johansen, & Hersleth, 2010). Mean-centring at respondent level was done by subtracting the mean attribute score of a respondent to each of his/her 15 attribute scores. The respondents’ nationalities were included in the model as downweighted

3.1. General image of traditional food

Table 2 Perception measurement of 15 attributes of traditional foods (seven-point scales). Left anchor (score 1)

Right anchor (score 7)

Low in quality Inconsistent quality Bad taste Ordinary taste Bad appearance Unhealthy Difficult to prepare Low availability Time-consuming to prepare Unsafe Low nutritional value Expensive Narrow assortment Environmental unfriendly Not supportive for our local economy

High in quality Consistent quality Good taste Special taste Good appearance Healthy Easy to prepare High availability Not time-consuming to prepare Safe High nutritional value Inexpensive Wide assortment Environmental friendly Supportive for our local economy

The general image of TFP receives a unanimous positive score across the six European countries, scoring above 5.5 on average on the seven-point scale (Table 3). The two highest mean scores are seen in Spain (6.04) and Poland (6.01), while the lowest mean scores are observed for Belgium (5.51) and France (5.62). The mean score for the pan-European sample is 5.80, with a standard deviation of 1.06. The frequency distributions of the general image scores corroborate the positive image of TFP across Europe (Fig. 1). A strikingly low number of 81 respondents (1.7%) out of the 4765 pan-European valid answers collected for this question utilised the «negative» side of the scale, i.e., scores from 1 to 3 on the seven-point scale, when reporting their personal image of traditional food. An analysis of scores distributions by gender and age groups showed no effect of gender, but a tendency to differences in age groups: the older the consumers, the higher their positivity towards TFP (results not shown). This corroborates prior results where traditional food consumers were profiled to be middle-aged to elderly (Vanhonacker, Lengard, et al., 2010). Further segmentation and profiling of the respondents other than country-wise is outside the scope of this paper; we refer to Vanhonacker, Lengard, et al. (2010) for a consumer profiling in relation to TFP based on all sections of the questionnaire. 3.2. Attribute perceptions of traditional food Fig. 2 presents the PCA correlation loadings (PC1 vs. PC2) on the perceptions of TFP attributes. The first two model dimensions reproduce 35% of the total variation and give insights in the main

Table 3 Mean and standard deviation of the general image of traditional food (scale: 1–7).

Mean (S.D.)

Belgium (n = 801)

France (n = 783)

Italy (n = 793)

Norway (n = 796)

Poland (n = 796)

Spain (n = 796)

Total sample (n = 4765)

5.51 (1.02)

5.62 (1.07)

5.93 (0.96)

5.71 (1.15)

6.01 (1.11)

6.04 (0.94)

5.80 (1.06)

132

V.L. Almli et al. / Food Quality and Preference 22 (2011) 129–138

Fig. 1. Frequencies of respondents’ scores on the general image scale per country. 1 = Very negative; 4 = neither positive nor negative; 7 = very positive.

1 HEALTH & ETHICS 0.8

0.6

Supportive economy Environ. friendly

Component 2 (12%)

0.4

High nutrition

PURCHASE & CONVENIENCE

Safe 0.2

FRANCE 0

−0.2

−0.4

Inexpensive Wide assortment Not time−consuming to prepare

POLAND

Healthy ITALY SPAIN BELGIUM NORWAY High quality Special taste

Easy to prepare High availability

Good taste Consistent quality Good appearance

−0.6

−0.8

−1 −1

SENSORY

−0.8

−0.6

−0.4

−0.2

0

0.2

0.4

0.6

0.8

1

Component 1 (23%) Fig. 2. Attribute perceptions of traditional food in six European countries. Three categories of attributes are found.

structured patterns within the attributes. Only the right anchors of the attributes are shown, while the left anchors are implicit as their negative correlations. None of the attributes dominate the map nor are fully explained with only two principal components. This indicates that these 15 attributes cover multidimensional aspects of traditional foods, as was the intention of the survey. Further, it indicates diverging attribute perceptions of TFP across the panEuropean sample and population, which relates to the broad character of these products’ definition.

The model reveals a splitting of intrinsic and extrinsic attributes (Ophuis & van Trijp, 1995) into three categories sharing common concepts: (1) Sensory, composed of intrinsic attributes (taste, appearance) and experience attributes (taste specialty, quality, quality consistency). (2) Health & Ethics, including intrinsic product attributes (nutritional value, safety), and credence attributes related to health

V.L. Almli et al. / Food Quality and Preference 22 (2011) 129–138

(healthy) and ethics (environmental friendliness, support of local economy). It must be noted that healthy is projected in-between the first two categories of attributes on the loadings map, indicating that healthiness varies somewhat independently of the sensory, ethical or convenience attributes. We chose to include this item with category (2) because it is conceptually more closely related to safety and nutritional value than to sensory characteristics (Grunert, 2005; Verbeke, Frewer, Scholderer, & De Brabander, 2007). (3) Purchase & Convenience, consisting of extrinsic experience attributes characterising commercial (availability, assortment, price) and convenience aspects (ease of preparation, preparation time). Fig. 4 presents the mean perceptions per country according to the three attribute categories observed in PCA. Category (3) of properties scored poorly in all countries but less so in Belgium (Fig. 4), explaining the projection of Belgium in the direction of Purchase & Convenience attributes on Fig. 2. It is interesting to note that the intrinsic product properties in the sensory category are positively correlated to extrinsic health and ethical attributes on PC1. The fact that TFP can both offer a good sensory experience and meet ethical concerns may partly explain the positive image of traditional food. Further, these attributes are negatively correlated to the extrinsic properties for purchase and convenience, which score relatively low (Fig. 4). This finding is both in accordance with Chambers, Lobb, Butler, Harvey, & Traill (2007) who found that people focusing on convenience in food choice had a more negative attitude and indicated a lower consumption of local foods, and with Pieniak et al. (2009) who concluded that convenience acted as a barrier to traditional food consumption. 3.3. Attribute perceptions of traditional food: national results in six countries 3.3.1. Average perception scores per country per attribute In order to highlight national similarities and differences in the attribute perceptions of TFP, Fig. 3 reports the mean scores per

133

country and per attribute. These national mean scores indicate the overall trends in attribute perceptions as obtained from the study sample, but may not be representative of the entire population in each country. In the following, the attributes that are interpreted were selected based on the criterion that they show relatively high or low values either in comparison to the other countries, or in comparison to the other attributes within the same country. On average, the Spanish and the Italians give similar scores. To them, TFP are characterised by a good and special taste, a high and consistent quality, a good appearance, a high nutritional value and healthiness. The Belgians perceive TFP as having a good taste, a high quality and a high availability. The French find TFP to be of high quality and rather expensive. The Poles characterise TFP by a good and special taste, a high yet not highly consistent quality, a high environmental friendliness, a good support for the local economy, a high preparation time and rather high prices. The Norwegians characterise TFP with a good taste, a high quality, a relatively low healthiness, a high safety and a long preparation time. A clear split in the group of countries is observed for six of the attributes. First, a special taste is attributed to TFP in Italy, Spain and Poland, but not particularly so in Norway, France and Belgium. Fig. 3 also reveals differences between Polish and Belgian consumers with regard to their perceptions of availability, ease of preparation and time of preparation. This may be explained by divergent conceptions of traditional food in the two countries. According to Vanhonacker, Verbeke, et al. (2010), while the Poles define traditional food mainly as specialty dishes consumed on festive occasions, the Belgians mostly consider traditional food as familiar food of daily character. Second, Norway demarks itself with a relatively low score on healthy and a relatively high score on safety. The low score on healthiness corroborates both Pieniak et al. (2009), who found a negative association of weight control with the general attitude to TFP in Norway, and Guerrero et al. (2009) who concluded from focus group discussions that traditional foods in Norway were recognised as rather fatty. As for safety, the high perception of this attribute is consistent with earlier studies where it was shown that

Fig. 3. Average scores per country per attribute of TFP.

134

V.L. Almli et al. / Food Quality and Preference 22 (2011) 129–138

Fig. 4. Average scores per country per category of attributes.

Norwegian consumers feel particularly confident that the government’s food controls secure safe food (Berg, 2005). Third, although consumers in all countries report that traditional food is not fast to prepare, scoring under mid-point in Norway and Poland, this trend is not marked in Belgium. This relates probably to the fact that Belgians define TFP rather as familiar food of daily character, which are often faster to prepare than festive dishes. Fourth, TFP are perceived as rather expensive in France, Poland and Norway, with scores of 3.9 in average, but not in the other countries. This may reflect the large presence of ‘‘Produits du terroir” in France and the definition of TFP as festive foods in Poland and Norway (Vanhonacker, Verbeke, et al., 2010). Last, Spanish respondents perceived the widest assortment of TFP with a score of 5.1 on average, while respondents in Poland and Norway scored moderately on the perception of broadness of assortment with mean scores of 4.2 and 4.4, respectively (Fig. 3). This variation in perceived assortment may be related to the number of products that fulfil the EU regulations for protected designations: 182 Spanish products, 34 Polish products (EC DOOR database, 2010) and 16 Norwegian products (not shown in the EC DOOR database as Norway is not a EU member) were registered by May 2010. It should be noted however that the DOOR database does not necessarily match the number of traditional food registrations at national levels. In Poland, 732 traditional products were registered nationally by May 2010 (Polish registry, 2010), a number in contradiction with the perception of a moderate assortment by Polish consumers. One possible explanation for this contradiction may be related to the wide food assortments of today, in which the relative share of traditional foods has decreased as compared to pre-liberalisation periods. Another possible explanation may be a mismatch between what food professionals and authorities define as traditional products, versus what consumers experience as traditional foods, maybe due to the high regional identity and diversity of this product category in Poland (Weichselbaum et al., 2009). Further, the less wide assortments perceived in Northern countries may be linked to their rougher climates, which restrict the possibilities for a fully varied agriculture compared to Southern nations like Spain (Jordana, 2000), or linked to different food quality policy orientations in the respective countries (Becker, 2009).

3.3.2. Patterns in attribute perceptions explored by PCA In order to study the patterns in attribute perceptions at a national level, six individual PCA models were run. Explained variances on the first two principal components range from 31% in Belgium to 39% in Poland. These models reveal that the three-category pattern of (1) Sensory, (2) Health & Ethics and (3) Purchase & Convenience attributes highlighted in the pan-European model also applies in each country separately (results not shown). This indicates a similar understanding and grouping of the selected product attributes across European food cultures and validates the translation work in elaborating the survey questionnaire in six languages. This finding is valuable as it establishes the relevance of building pan-European models and policies based on the survey and its outcomes. 3.4. Relating attribute perceptions to general image: pan-European results The PLSR models at national and pan-European levels confirm the main pattern of relationships between the attributes seen in the PCA models (Fig. 2). It is however important to note that unlike PCA, the PLSR models do not aim at exploring patterns in the attribute perceptions, but at identifying the attributes that contribute to a positive image of TFP. Therefore, the PLSR results differ somewhat from the PCA results, and more importantly, complement these. Further, the PLSR results complement the mean descriptive scores presented in Section 3.3.1. While the mean scores reported the consumers’ perception of traditional food in terms of 15 individual attributes, the PLSR models identify which of these attribute perceptions specifically shape the overall image of traditional food. The attributes’ associations to the positive general image of traditional food in six countries and for the pan-European sample are presented in Table 4. Significant negative regression coefficients are symbolised with a minus sign and indicate an association of general image to the negative anchor of the attribute. Positive coefficients are symbolised with plus and indicate associations to the positive anchor. All results in Table 4 are shown for one PLS-component models, except the model for France which includes two PLS-components due to a non-negligible contribution of explained variance on the second model dimension (Comp 1: 11.6%, cross-validation: 10%; Comp 2: 3.7%, cross-validation: 2%). The

135

V.L. Almli et al. / Food Quality and Preference 22 (2011) 129–138

Table 4 Attributes’ associations to the positive general image of traditional food. ‘‘+” indicates associations to positive attribute anchors; ‘‘ ” indicates associations to negative attribute anchors; ‘‘ns” stands for non-significant. Belgium (n = 801)

Francea (n = 783)

Italy (n = 793)

Norway (n = 796)

Poland (n = 796)

Spain (n = 796)

Total sample (n = 4765)

Sensory Taste (bad/good) Taste (ordinary/special) Appearance (bad/good) Quality (low/high) Quality (inconsistent/consistent)

+ + ns + ns

ns ns + ns

+ + + + +

+ + + + +

+ + ns + +

+ + ns + +

+ + + + +

Health & Ethics Healthiness (unhealthy/healthy) Safety (unsafe/safe) Nutritional value (low/high) Environmental friendliness (unfriendly/friendly) Support of local economy (not supportive/supportive)

ns + ns ns ns

ns ns ns ns ns

ns + ns ns +

ns ns ns ns ns

+ + + + ns

+ + ns ns ns

+ + + ns ns

Attributes ( /+) associated to general image

Purchase & Convenience Ease of preparation (difficult/easy) Preparation time (time-consuming/not time-consuming) Availability (low/high) Assortment (narrow/wide) Price (expensive/inexpensive) a

ns ns ns

ns +

ns

ns

ns ns

ns

Results from a two-component PLSR model. All other results based on one-component PLSR models.

pan-European model explains 9% (cross-validation: 8.5%) of the variations in general image. In the following, the pan-European results are reported and discussed. National results are presented in Section 3.5. The eight intrinsic product attributes (taste, taste specialty, appearance, quality, quality consistency, healthiness, safety and nutritional value), the convenience attributes (ease and time of preparation) and two purchase attributes (availability and price) show statistical significance in their influence on the general image for the pan-European sample (Table 4). Only three out of the 15 attributes do not show significance at pan-European level. These are the two ethical attributes (environmental friendliness and supportive of local economy) and purchase attribute assortment. Among the Purchase & Convenience properties, it is particularly interesting to observe that negative anchors expensive, low availability, difficult to prepare and time-consuming to prepare are associated to a positive general image of TFP. This means that the respondents who scored relatively low on attributes price, availability, ease of preparation and preparation time also were the ones who scored relatively high on general image. This gives evidence that the general image of TFP is not necessarily damaged by factors of inconvenience in purchasing or preparing this category of foods. Prior studies found that for quality-seeking consumers, a high price signals a high quality product (Grunert, Skytte, Esbjerg, Poulsen, & Hviid, 2002; Juhl, Høg, & Poulsen, 2000; Verbeke, Vermeir, & Brunsø, 2007; Zeithaml, 1988). Our results show that high quality is a unanimously significant attribute in all countries and for the pan-European model (Table 4), suggesting that consumers of TFP qualify as quality-seekers. For them, a high price may therefore re-enforce the good image of TFP rather than deteriorate it. To conclude, our results give evidence that European consumers may trade-off some degree of inconvenience in the purchase, expensiveness and preparation of TFP in order to enjoy the specific taste, quality, appearance, nutritional value, healthiness and safety they find in TFP. Further, the attributes associated to the general image of traditional food can be interpreted in relation to the dual appropriateness of use observed by Vanhonacker, Verbeke, et al. (2010). These authors reported that traditional food consumption can be associated both to daily or to festive meal occasions. In our results, several attributes associated to the general image of TFP may typically describe festive foods rather than everyday foods, for

example: special taste, high quality, difficult and time-consuming to prepare, low availability and expensive. Conclusively, festive consumptions of traditional food may be the ones that most strongly shape the general image of TFP in the consumers’ minds. That is to say that despite a higher frequency of consumption in ordinary and everyday meals, reflexions and feelings associated to traditional food may be forged in festive settings. 3.5. Relating attribute perceptions to general image: national results in six countries The country-specific models explain from 6% (cross-validation: 5%) of the general image of TFP in Italy and Spain to 16% (cross-validation: 15%) in Norway. The low explained variances of these models indicate: (a) that the relationships between the general image of TFP and their attributes are very individual, thus no common model will fit all respondents; (b) that our 15 attributes are not sufficient to draw the full picture of TFP’s general image. However, all of the 15 selected attributes are relevant as each of them is significantly correlated to general image in at least one country. A first general observation is that the results are very similar from one country to another. In the following, we discuss the attributes that contribute to the positive image of TFP for each of the attribute categories defined in Section 3.2. 3.5.1. Sensory attributes Attribute perceptions relating to good taste (significant only on PLS-component 1 in France) and high quality contribute to a positive general image of TFP in each of the six countries, corroborating the importance of sensory characteristics in traditional foods (Caporale, Policastro, Carlucci, & Monteleone, 2006; Iaccarino, Di Monaco, Mincione, Cavella, & Masi, 2006; Platania & Privitera, 2006). A good appearance is less necessary to generate a good image, as it is a positive driver in Italy and Norway only. Consistent quality is significantly related to the positive image of TFP in most countries, except Belgium and France. It is interesting to observe that consistent quality significantly impacts the general image of TFP in Poland, even though this country gave a lower score than Belgium and France to that attribute. This suggests that not necessarily the most typical attributes of TFP, but rather the most valued ones within each country, contribute to shaping the positive general image of TFP.

136

V.L. Almli et al. / Food Quality and Preference 22 (2011) 129–138

3.5.2. Health & Ethics Health & Ethics attributes do not create unanimity in the six countries. Healthy is a significant driver of the positive image of TFP in Poland and Spain only, and may partly explain why these two countries scored the highest on general image of TFP. It is interesting to observe that even though Italian consumers described TFP as healthy with a mean score of 4.5 (Fig. 3), health perception is not a driver of the positive image of TFP in this country. This is because both Italians who have a very positive image and Italians who have a less positive image of TFP may perceive TFP as healthy. Similarly, safety is significant in all countries except in Norway and France. We have observed above that consumers in Norway scored highest on safety as a perceived attribute of TFP (Fig. 3), yet it is not a driver of the positive image in this country. This can be explained by the fact that the Norwegians perceive Norwegian foods as safe independently of their general positivity or negativity towards TFP. Further, high nutritional value and environmental friendliness are drivers of positivity towards TFP in Poland only. This is coherent with the fact that Polish consumers scored highest on environmental friendliness (Fig. 3). Last, Italy is the only country where the attribute supportive of our local economy significantly contributes to a positive image of TFP. This is consistent with the fact that in the scope of the same survey, the Italians were found to be the most familiar respondents with Protected Designation of Origin (PDO; known by 95% of the Italian respondents), Protected Geographical Indication (PGI; known by 84% of the Italian respondents) and Traditional Speciality Guaranteed (TSG; known by 34% of the Italian respondents) labels (Vanhonacker et al., 2008). 3.5.3. Purchase & Convenience Attribute perceptions relating to time-consuming to prepare and expensive unanimously correlate with a positive general image of TFP in all of the six countries, indicating that consumers do not typically spare their time and money when it comes to TFP. This supports previous findings showing that traditional food consum-

ers in Europe use a relatively high percentage of their income on food and spend a long time making dinner both on ordinary days and on special occasions (Vanhonacker, Lengard, et al., 2010). Consumers with a positive image of TFP also perceive TFP as difficult to prepare (not significant in France and Spain) and characterised by a low availability (not significant in France and Belgium). An explanation for the apparent contradiction between negative attributes expensive, time-consuming and difficult to prepare and a positive general image of TFP may be brought by the fact that traditional food consumers consider making dinner as taking care of their family (Vanhonacker, Lengard, et al., 2010). For these consumers, more time and effort spent on preparing TFP for their family may lead to positive feelings for these products. When it comes to assortment, disparities are revealed as a narrow perceived assortment is significantly correlated to a positive general image of TFP in Norway and Poland, possibly due to a notion of exclusivity associated with TFP in these countries, whereas a wide assortment is significant in France, relating to the variety of TFP of daily character. 3.5.4. Dual definition of TFP revealed by PLSR in France In the model for France, two PLS-components are needed to properly report the structured variation in the data against one PLS-component only for all other countries. With the data from France, a model with one PLS-component (12% Y-explained variance; cross-validation: 10%) identifies good taste, good appearance, high quality, safe, supportive of local economy, difficult to prepare, time-consuming to prepare, low availability and expensive as significant drivers of a positive general image (results not shown). PLScomponent 1 may therefore be interpreted as the dimension representative of consumers who define TFP as festive and/or seasonal foods. A model with two PLS-components (16% Y-explained variance, cross-validation: 12%) reports ordinary taste, high quality, time-consuming to prepare, wide assortment and expensive as significant drivers of a positive general image (Fig. 5). PLS-component 2 may therefore be interpreted as the dimension representative of

X−loading weights and Y−loadings 0.6

Wide assortment

0.4

Component 2 (X:6%, Y:4%)

High availability

High quality

Easy to prepare

TFP IMAGE

0.2

0

Not time−consuming to prepare

Healthy Supportive economy

Good appearance Environ. friendly Safe

−0.2

Inexpensive

High nutrition

Good taste Consistent quality

−0.4 Special taste

−0.6 −0.8

−0.6

−0.4

−0.2

0

0.2

0.4

0.6

0.8

Component 1 (X:17%, Y:12%) Fig. 5. General image of traditional food (TFP IMAGE) related to 15 product attributes in France. Statistically significant attributes are indicated in bold.

V.L. Almli et al. / Food Quality and Preference 22 (2011) 129–138

consumers who define TFP as frequently-consumed foods rather than specialty foods. This model shows that the drivers of general image of TFP may depend upon one’s own definition of TFP in terms of context of consumption: festive vs. daily foods. It must be noted that this duality of consumption is present in all countries, yet comes out significantly toward general image only in the model for France. For the other countries and for the pan-European sample, the general image of TFP relates to attributes typical of festive traditional foods rather than traditional foods of daily character. Moreover, it is interesting to observe that three of the attributes remain significant in either context of consumption: high quality, time-consuming to prepare and expensive. These are also the only three attributes that significantly correlate to general image in all six countries.

137

5. Conclusion

A measure of the general image of TFP revealed a positive general perception of TFP across Europe, yet some of the perceived attributes of TFP scored just above average on their scales (Fig. 3). Accordingly, these results imply that there is a potential for further image improvement and subsequent sales growth. Based on our results, the following tentative suggestions may be exposed. Firstly, the industry is encouraged to keep surfing on the current high quality, high-value image of TFP which satisfies quality-seekers. Secondly, the introduction of TFP with increased convenience may appeal to new consumers. Thirdly, an enlarged assortment may allow increasing sales volumes within the existing consumer group, though notions of exclusivity are important for some consumers. In addition, Norway may focus on developing healthier TFP, such as low-fat and/or salt-reduced meat products. Finally, the strong environmental friendly, supportive of local economy model from Poland may be exported to other countries, especially Norway and Belgium. This may perhaps be implemented hand-in-hand with a further development and marketing effort of local foods and PDO, PGI, TSG labelling.

This paper used information collected through a cross-cultural survey in six European countries to measure the general image of traditional food and to identify which attribute perceptions contribute to a positive image of this group of products. It was revealed that traditional food products benefit from a positive general image in all countries with the highest scores in Spain and Poland. Further, TFP present satisfying sensory, health and ethical properties, but less positive purchase and convenience attributes. Common to all countries, the positive general image of TFP is significantly correlated to high quality, time-consuming to prepare and expensive. Additional attributes contribute significantly to a positive general image of TFP when the pan-European sample is studied as a whole: special taste, good appearance, consistent quality, healthy, safe, high nutritional value, difficult to prepare and low availability. Our results give evidence that European consumers tradeoff some degree of inconvenience in the purchase and preparation of TFP in order to enjoy their specific taste, quality, appearance, nutritional value, healthiness and safety. Further, results indicate that not necessarily the most typical attributes of TFP, but rather the most valued ones, contribute to shaping a positive general image of TFP. Finally, our results also suggest that festive consumptions of traditional foods, rather than daily consumptions, are the ones that most strongly shape the general image of TFP in the European consumers’ minds. The use of multivariate models has illustrated some of the modelling and visualisation capabilities of these methods. PCA’s exploratory abilities were used to discover patterns in the attribute perceptions of TFP, while PLSR’s explanatory skills were used to relate the product attributes to the general image of TFP. The PLSR model for France allowed uncovering a duality in the respondents’ answers, which we would otherwise have missed out. In our experience, multivariate methods are valuable for the analysis of attitudinal and behavioural data. Finally, this paper contributes to understanding the general image of TFP in Europe, and calls for confirmative and complementary work by expanding the research to additional countries and to other product attributes.

4.2. Limitations in relating general image to product attributes

Acknowledgements

The current research was limited to a selection of 15 product attributes drawn from the results of previous qualitative studies (Guerrero et al., 2009, 2010). Only 9% of the variations in general image were explained by the 15 attributes in the pan-European regression model. That is to say that although the attributes that are identified contribute significantly to explaining the general image of TFP, these are far from enough to explain it wholly and fully. Additional factors which were beyond the scope of this study may also play a role in the general image of TFP, such as branding, guarantee of origin labelling (Caporale & Monteleone, 2001), manufacturing process information (Caporale & Monteleone, 2004; Iaccarino et al., 2006), packaging, advertising and distribution networks (Vannoppen, Verbeke, Van Huylenbroeck, & Viaene, 2001). In addition, referring to the definition of TFP in Vanhonacker, Verbeke, et al. (2010) we believe that TFP trigger consumers’ emotions and values (‘‘. . .specific celebrations, transmitted from one generation to another, gastronomic heritage, associated to a certain area. . .”) which are not reflected by rational attributes such as nutritional value or price. Nonetheless, the attributes that come out significant in the pan-European regression model have been robust to crosslanguage and cross-cultural differences, respondent effects and statistical cross-validation. This supports the reliability of the findings.

We would like to thank TRUEFOOD (Traditional United Europe Food) for financial support. TRUEFOOD is an Integrated Project financed by the European Commission under the Sixth Framework Programme (Contract No. FOOD-CT-2006-016264). The information in this document reflects only the authors’ views and the Community is not liable for any use that may be made of the information contained therein. The authors are grateful to the referees for their comments and suggestions.

4. Implications and limitations 4.1. Implications for the traditional food sector

References Becker, T. (2009). European food quality policy: The importance of geographical indications, organic certification and food quality assurance schemes in European countries. The Estey Centre Journal of International Law and Trade Policy, 10(1), 111–130. Berg, L. (2005). Tillitens triangler om forbrukertillit og matsikkerhet [Triangle of trust: On consumer trust and food safety]. Fagrapport nr. 1-2005. ISSN: 1502-6760, ISBN: 82-7063-403-4. Bertozzi, L. (1998). Tipicidad alimentaria y dieta mediterránea. In A. Medina, F. Medina, & G. Colesanti (Eds.), El color de la alimentación mediterránea. Elementos sensoriales y culturales de la nutrición (pp. 15–41). Barcelona: Icaria. CAMO Software AS. The Unscrambler v9.8. Oslo, Norway. . Caporale, G., & Monteleone, E. (2001). Effect of expectations induced by information on origin and its guarantee on the acceptability of a traditional food: Olive oil. Sciences Des Aliments, 21(3), 243–254.

138

V.L. Almli et al. / Food Quality and Preference 22 (2011) 129–138

Caporale, G., & Monteleone, E. (2004). Influence of information about manufacturing process on beer acceptability. Food Quality and Preference, 15(3), 271–278. Caporale, G., Policastro, S., Carlucci, A., & Monteleone, E. (2006). Consumer expectations for sensory properties in virgin olive oils. Food Quality and Preference, 17(1–2), 116–125. Chambers, S., Lobb, A., Butler, L., Harvey, K., & Traill, B. (2007). Local, national and imported foods: A qualitative study. Appetite, 49, 208–213. Chittleborough, C. R., Baum, F., Taylor, A. W., & Hiller, J. E. (2008). Missing data on retrospective recall of early-life socio-economic position in surveillance systems: An additional disadvantage? Public Health, 122(11), 1152–1166. Deliza, R., & MacFie, H. J. H. (1996). The generation of sensory expectation by external cues and its effect on sensory perception and hedonic ratings: A review. Journal of Sensory Studies, 11, 103–128. EC DOOR database. (2010). Database of Origin and Registration of the European Commission, Agriculture and Rural Development. (visited May 2010). Esbensen, K. H. (2001). Multivariate data analysis – In practice (5th ed.). Oslo, Norway: CAMO ASA. ISBN: 82-993330-3-2. EU. (2006). Council Regulation (EC) No. 509/2006 of 20 March 2006 on agricultural products and foodstuffs as traditional specialties guaranteed. Official Journal of the European Union, L 93/1. Fitzsimons, G. M., Chartrand, T. L., & Fitzsimons, G. J. (2008). Automatic effects of brand exposure on motivated behavior: How Apple makes you ‘‘Think Different”. Journal of Consumer Research, 35(1), 21–35. Grunert, K. G. (2005). Food quality and safety: Consumer perception and demand. European Review of Agricultural Economics, 32, 369–391. Grunert, K. G., Skytte, H., Esbjerg, L., Poulsen, C. S., & Hviid, M. (2002). Dokumenteret kødkvalitet. MAPP project paper no. 2. Guerrero, L., Claret, A., Verbeke, W., Enderli, G., Zakowska-Biemans, S., Vanhonacker, F., et al. (2010). Perception of traditional food products in six European countries using free word association. Food Quality and Preference, 21(2), 225–233. Guerrero, L., Guardia, M. D., Xicola, J., Verbeke, W., Vanhonacker, F., ZakowskaBiemans, S., et al. (2009). Consumer-driven definition of traditional food products and innovation in traditional foods. A qualitative cross-cultural study. Appetite, 52(2), 345–354. Guinard, J. X., Uotani, B., & Schlich, P. (2001). Internal and external mapping of preferences for commercial lager beers: Comparison of hedonic ratings by consumers blind versus with knowledge of brand and price. Food Quality and Preference, 12(4), 243–255. Guinot, C., Latreille, J., & Tenenhaus, M. (2001). PLS path modelling and multiple table analysis. Application to the cosmetic habits of women in Ile-de-France. Chemometrics and Intelligent Laboratory Systems, 58(2), 247–259. Iaccarino, T., Di Monaco, R., Mincione, A., Cavella, S., & Masi, P. (2006). Influence of information on origin and technology on the consumer response: The case of soppressata salami. Food Quality and Preference, 17(1–2), 76–84. Jordana, J. (2000). Traditional foods: Challenges facing the European food industry. Food Research International, 33(3–4), 147–152. Juhl, H. J., Høg, E., & Poulsen, C. S. (2000). Forbrugernes vurdering af nogle udvalgte kvalitetsmærkninger for hakket oksekød: Undersøgelsens design, gennemførelse og resultater. MAPP project paper no. 2. Malhotra, N. K., & Peterson, M. (2006). Basic marketing research: A decision-making approach (p. 190). Upper Saddle River: Pearson Education/Prentice Hall.

Martens, H. (2001). Reliable and relevant modelling of real world data: A personal account of the development of PLS Regression. Chemometrics and Intelligent Laboratory Systems, 58(2), 85–95. Martens, H., & Martens, M. (2000). Modified Jack-knife estimation of parameter uncertainty in bilinear modelling by partial least squares regression (PLSR). Food Quality and Preference, 11(1–2), 5–16. Martens, H., & Martens, M. (2001). Multivariate analysis of quality. An introduction. John Wiley & Sons Ltd.. ISBN: 0-491-97428-5. Næs, T., Lengard, V., Bølling Johansen, S., & Hersleth, M. (2010). Alternative methods for combining design variables and consumer preference with information about attitudes and demographics in conjoint analysis. Food Quality and Preference, 21, 368–378. Ophuis, P., & van Trijp, H. C. M. (1995). Perceived quality – A market driven and consumer oriented approach. Food Quality and Preference, 6(3), 177–183. Pieniak, Z., Verbeke, W., Vanhonacker, F., Guerrero, L., & Hersleth, M. (2009). Association between traditional food consumption and motives for food choice in six European countries. Appetite, 53(1), 101–108. Platania, M., & Privitera, D. (2006). Typical products and consumer preferences: The ‘‘soppressata” case. British Food Journal, 108(5), 385–395. Pohjanheimo, T., & Sandell, M. (2009). Explaining the liking for drinking yoghurt: The role of sensory quality, food choice motives, health concern and product information. International Dairy Journal, 19(8), 459–466. Polish registry. (2010). Register of traditional foods. Polish Ministry of Agriculture and Rural Development. (visited May 2010). Stone, M. (1974). Cross-validatory choice and assessment of statistical predictions. Journal of the Royal Statistical Society, Series B: Methodological, 36(2), 111–147. Tenenhaus, M., Vinzi, V. E., Chatelin, Y. M., & Lauro, C. (2005). PLS path modeling. Computational Statistics & Data Analysis, 48(1), 159–205. Trichopoulou, A., Soukara, S., & Vasilopoulou, E. (2007). Traditional foods: A science and society perspective. Trends in Food Science & Technology, 18(8), 420–427. Vanhonacker, F., Lengard, V., Hersleth, M., & Verbeke, W. (2010). Profiling European traditional food consumers. British Food Journal, 112(8), 871–886. Vanhonacker, F., Verbeke, W., Guerrero, L., Claret, A., Sulmont-Rossé, C., Raude, J., et al. (2010). How European consumers define the concept of traditional food: Evidence from a survey in six countries. Agribusiness. doi:10.1002/agr.20241. Vanhonacker, F., Verbeke, W., Sulmont-Rossé, C., Issanchou, S., Raude, J., Contel, M., et al. (2008). National and cross-national reports of the consumer survey. TRUEFOOD deliverable, D1.2.1.3. Vannoppen, J., Verbeke, W., Van Huylenbroeck, G., & Viaene, J. (2001). Consumer valuation of short market channels for fresh food through laddering. Journal of International Food & Agribusiness Marketing, 12(1), 41–69. Verbeke, W., Frewer, L. J., Scholderer, J., & De Brabander, H. F. (2007). Why consumers behave as they do with respect to food safety and risk information. Analytica Chimica Acta, 586, 2–7. Verbeke, W., Vermeir, I., & Brunsø, K. (2007). Consumer evaluation of fish quality as basis for fish market segmentation. Food Quality and Preference, 18, 651–661. Weichselbaum, E., Benelam, B., & Soares Costa, H. (2009). Traditional foods in Europe. In European Food Information Ressource (EuroFIR) consortium, EU 6th framework food quality and safety thematic priority. Contract FOOD-CT-2005513944. . Zeithaml, V. A. (1988). Consumer perceptions of price, quality and value – A meansend model and synthesis of evidence. Journal of Marketing, 52(3), 2–22.