Sustainable food consumption in the nexus between national context and private lifestyle: A multi-level study

Food Quality and Preference 55 (2017) 16–25

Contents lists available at ScienceDirect

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

Sustainable food consumption in the nexus between national context and private lifestyle: A multi-level study John Thøgersen ⇑ Aarhus University, Denmark

a r t i c l e

i n f o

Article history: Received 9 May 2016 Received in revised form 12 August 2016 Accepted 14 August 2016 Available online 16 August 2016 Keywords: Food-related lifestyle Sustainable consumption Organic food Consumer innovativeness Multi-level latent class analysis

a b s t r a c t This paper investigates how country of residence and food-related lifestyle (FRL) interact in shaping (un)sustainable food consumption patterns. An online survey was carried out in ten European countries (n  335 in each country), covering the five regions North, South, East, West and Central Europe. Multi-group CFA (AMOS22) was used to test the cross-national validity of the FRL instrument. After deleting a few items, it was found that the factorial structure of all five FRL domains is invariant with respect to factor configuration and factor loadings but not with respect to item intercepts. The segmentation analysis was performed by means of Latent Gold 5.1 and multi-level latent class analysis based on data from all ten countries and using the 23 FRL dimensions as input. A five-segment, three-country class solution was judged to produce the best compromise between fit and parsimony, confirming that cross-country FRL segments can be meaningfully identified, but that the segment structure differs across Europe’s regions. The joint effect of country class and FRL on sustainable food-related consumer behaviour was analysed by means of GLM (SPSS22). Both country class and FRL significantly account for variation in meat and organic food consumption and FRL in addition for variation in sustainable food product innovativeness. Further, there is significant interaction between country and FRL for all outcome variables. Hence, the impact of FRL on sustainability choices partly depends on country of residence. Ó 2016 Elsevier Ltd. All rights reserved.

1. Introduction Contemporary food production and consumption are not sustainable (Reisch, Eberle, & Lorek, 2013) but they contribute substantially to global problems such as climate change, biodiversity loss, and environmental degradation (IPCC, 2014). Food is one of the three consumption domains responsible for the largest share of environmental impact (the other two being housing and transportation, cf., e.g., Steen-Olsen & Hertwich, 2015; Tukker, 2015). The shift seen in recent decades towards a more meat-heavy diet (Popp, Lotze-Campen, & Bodirsky, 2010), especially in middleincome countries (Speedy, 2003), also links sustainable food consumption to challenges such as food security, poverty and inequality (Field & et al., 2014; Reisch et al., 2013). An emerging stream of research investigates the potential of lifestyle changes to drive the needed transition towards a lowcarbon future (e.g., Mont, Neuvonen, & Lähteenoja, 2014; Neuvonen et al., 2014). Others have investigated how national context shapes the sustainability of consumption patterns (cf. ⇑ Address: Department of Management, Bartholins Allé 10, 8000 Aarhus C, Denmark. E-mail address: [email protected] http://dx.doi.org/10.1016/j.foodqual.2016.08.006 0950-3293/Ó 2016 Elsevier Ltd. All rights reserved.

Thøgersen, 2010a). Also, lifestyle and national context are hardly mutually independent. For example, it is highly likely that national and cultural contexts are major factors shaping consumer lifestyles (Milfont & Markowitz, 2016). At the same time, there are forces that may ‘‘synchronize” consumer lifestyles across countries, such as mass media, advertising, international traveling, and the human tendency to emulate those that seem to be doing better, including those in other, economically more advanced countries (e.g., Cleveland, Laroche, & Papadopoulos, 2009; Thøgersen, 2010b; Wilk, 2002). Be that as it may, there is a lack of empirical research on the relative importance of, and the likely interplay between, a person’s private lifestyle and opportunities and constraints afforded by the context with regard to the sustainability of consumers’ food-related choices (cf. Milfont & Markowitz, 2016). Against this backcloth, this paper reports a study of the extent to which differences in consumers’ more or less sustainable foodrelated choices and their openness to new, more sustainable food products result from their private lifestyle versus the opportunities and constraints afforded by the wider context in which they live. Using Grunert and associates’ thoroughly validated instrument for measuring FRL (e.g., Grunert, 1993; Scholderer, Brunsø, Bredahl, & Grunert, 2004), national, and potentially crossnational, food-related lifestyle (FRL) segments are identified by

J. Thøgersen / Food Quality and Preference 55 (2017) 16–25

means of multi-level latent class analysis of survey data from representative samples of consumers from ten European countries. The approach employed enables both profiling the identified cross-national lifestyle segments and determining the joint impact of food-related lifestyle and national context for (self-reported) behaviour, in this case more sustainable food choices. 2. Lifestyle research Lifestyle research in marketing is primarily used for market segmentation (Plummer, 1974; Vyncke, 2002). With the development of the affluent consumer society, demographic characteristics became less and less predictive of consumer behaviour, and ‘‘psychographic” (Demby, 1974) or lifestyle segmentation was proposed as a more effective way to divide consumers into relatively homogeneous groups. Lifestyle segmentation is usually surveybased, where lifestyle groups or segments are identified by first using a data reduction technique, such as factor analysis, multidimensional scaling or correspondence analysis, followed by a cluster analysis based on the dimensions found in the data. In lifestyle research, it is increasingly acknowledged that people may not just have one, but can have several interconnected lifestyles. Researchers have suggested the existence of domainspecific lifestyles (van Raaij & Verhallen, 1994), of which especially food-related lifestyles have been thoroughly researched (Grunert, 1993). The basic proposition behind domain-specific lifestyles is that a person’s lifestyle need not be consistent across domains and therefore descriptions of lifestyles should be restricted to specific life domains (van Raaij & Verhallen, 1994). Grunert (1993) proposed a Food-Related Lifestyle (FRL) model, which has been further developed and applied in a wide range of countries all over the world (e.g., Grunert, Brunsø, Bredahl, & Bech, 2001; Grunert et al., 2011; Nie & Zepeda, 2011). Grunert (1993) characterizes his FRL model as a deductive, cognitive approach to lifestyle research. Lifestyle is conceived as a mental construct, which is different from but explains behaviour. Specifically, he defines domain-related lifestyle as ‘‘the system of cognitive categories, scripts, and their associations, which relate a set of products to a set of values” (Grunert, Brunsø, & Bisp, 1993, p. 13). Taking inspiration from psychological means-end chain theory (Gutman, 1982), Grunert’s (1993) FRL model views lifestyle as part of a hierarchical, cognitive-behavioural system functioning as an organizing and guiding construct in a person’s life. Lifestyles are conceived as a means to achieving personal superordinate goals or values (e.g., hedonism, tradition, self-direction), which are more abstract and trans-situational cognitive categories (Rokeach, 1973; Schwartz, 1992; Schwartz, 1994). In specific situations, lifestyle is assumed to be the backcloth that frames a consumer’s perceptions of products and services and guides her choices and behaviours. The system of cognitive structures that constitutes a FRL is assumed to include two types of cognitive schemas related to food, purchase motives and food quality aspects, as well as three broad cognitive scripts related to food provision, viz. ways of shopping, cooking methods and consumption situations. Together, these five cognitive elements are assumed to capture the key characteristics of an individual’s food-related lifestyle. Specifically, purchase motives refer to the consequences a consumer anticipates from a meal, including, for example, social aspects, hedonism, tradition, and security (Brunsø, Scholderer, & Grunert, 2004; de Boer, McCarthy, Cowan, & Ryan, 2004). Quality aspects refer to the general importance that consumers attach to food product attributes, such as healthy, tasty, natural, organic, and fresh (Brunsø et al., 2004). Consumers use both purchase motives and quality aspects to justify their purchases. Ways of shopping refers to how consumers actually shop for food, that is,

17

do they find the task enjoyable, do they make shopping lists, do they deliberate extensively (or not) when making a decision, how much do they consider the price and other product information, or do they rely on the advice of experts, like friends or sales staff, do they prefer one-stop shopping or use specialty food shops? Cooking methods include preparation time and how the products purchased are transformed into meals, is the preparation characterized by efficiency or by indulgence, is cooking planned or spontaneous, is it a social event or the sole responsibility of one person? Finally, the consumption situation refers to issues such as the number of set meals, snacking habits, eating out and the social aspects of sharing a meal (Brunsø et al., 2004; de Boer et al., 2004). The overall FRL model (see Fig. 1) is thus a system of interacting elements in which personal values are (part of) the foundation from which purchasing motives are derived; quality aspects, consumption situations, ways of shopping and cooking methods frame our view of food products, services, and other food-related activities and thus affect our behaviour, including food choices and preparation and how we, for example, deal with food and foodrelated waste. The cross-cultural validity of the Grunert and associates FRL instrument presented below has been thoroughly tested and confirmed (O’Sullivan, Scholderer, & Cowan, 2005; Scholderer et al., 2004). Specifically, these studies found that the FRL instrument possesses metric invariance in a European context, but scalar invariance only across limited samples of countries (O’Sullivan et al., 2005), not in general across all European countries (Scholderer et al., 2004). As predicted by the model, food-related lifestyle has been found to completely mediate the relationship between basic value priorities and food-related behaviour (Brunsø et al., 2004). On the behaviour side, FRL has been successfully applied to the study of the risk of general lifestyle diseases, such as obesity (Pérez-Cueto et al., 2010). FRL has also been found to predict a range of specific food-related behaviours, including how consumers respond to new food products (Cullen & Kingston, 2009), meat consumption (Grunert, 2006), and preferences for a vegetarian diet (Hoek, Luning, Stafleu, & de Graaf, 2004). However, up till now these empirical studies have mostly been carried out on a country-bycountry basis. When multiple countries are studied, comparative analyses generally are of an add-on nature. To my knowledge, the present study is the first one to employ a multi-level analytical approach to a multi-country study of food-related lifestyles. The most important benefit of such an approach is the opportunity to obtain an integrated picture of how behavioural outcomes are co-determined by interacting personal-level (i.e., FRL) and broader contextual (i.e., country of residence) factors (Milfont & Markowitz, 2016).

Source: Adapted from Grunert (2006). Fig. 1. The food-related lifestyle model.

18

J. Thøgersen / Food Quality and Preference 55 (2017) 16–25

3. Method 3.1. Data collection and sample With the aim of mapping lifestyles and related constructs, a survey study was carried out in ten European countries (N  1000 in each country),1 covering the five regions North, South, East, West and Central Europe. A random third of respondents completed a questionnaire which included the food-related lifestyle (FRL) instrument. The following analyses are limited to these respondents (N  335 in each country). In addition to lifestyles, the survey also contained questions about a range of relevant background characteristics and behavioural outcomes related to sustainable foodconsumption as well as questions not pertinent to the present study. The questionnaire was developed in English and translated into the nine other national languages. In order to check the translations, they were back-translated into English by the organizations doing the translation (but by a different person). The author checked the back-translations, comparing them with the original in English and settled all uncertainties and ambiguities with the translators. Before implementing the surveys, the final online questionnaire was further checked by a knowledgeable, native speaker of the language in question, which led to limited adjustments. A professional market research company (YouGov) collected the survey data; YouGov sampled respondents from their own and their partners’ panels in the other countries, administered the data collection as Computer Assisted Web Interviews (CAWI) and organized the data into files for statistical analysis. The samples from each country are representative for the 18–65-year age group with regard to gender, age and geography.

and organic food), using a five-point rating scale with the end points ‘‘never” (1) and ‘‘always/every time” (5) and the labelled midpoint ‘‘about half the times” (3). A ‘‘Not applicable” option was also given. The items are: (a) How often do you have beef for supper at home? (b) How often do you have a meatless supper at home? (c) How often do you buy organic tomatoes? (c) How often do you buy organic eggs? (d) How often do you buy organic milk? The three latter items were combined and used in the following as an indicator for organic food consumption. The threeitem scale shows a splendid construct reliability (Cronbach’s alpha = 0.84). In addition, we adapted Goldsmith and colleagues’ (Goldsmith & Hofaker, 1991) six-item consumer innovativeness scale to measure innovativeness with regard to environmentally-friendly food products. In the middle part of the interview, participants read the following introduction before responding to six consumer innovativeness items, which were interspersed with other items: ‘‘Please indicate your opinion about the statements below by checking the box that comes closest to your opinion. Use the scale from 1 to 7 for your answer. Mark 1 if you totally disagree with the statement and mark 7 if you totally agree. You may use the numbers from 2 to 6 to graduate your response if you neither totally agree nor totally disagree, depending on how much you agree or disagree with the statement.” An exploratory factor analysis revealed that the six items loaded on two factors. Four of the six items loaded on the first factor (e.g., ‘‘In general, I am among the first in my circle of friends to buy a new environmentally friendly food product when it appears”), which is used in the following as the indicator of consumer innovativeness. The four-item scale shows an excellent construct reliability (Cronbach’s alpha = 0.91).

3.2. Measuring food-related lifestyle

3.4. Removing mischievous respondents

For measuring the five FRL components mentioned earlier, the FRL instrument contains 69 items, which were placed in the first part of the interview. Before responding to these items, which were presented in individualized, random order, the respondents read the following introduction: ‘‘To begin with, we would like to get a better picture of your opinion, habits and routines on a broad range of issues that are related to eating and cooking. Please indicate to what extent you agree or disagree with the statements below. Use the scale from 1 to 7 for your answer. Mark 1 if you totally disagree with the statement and mark 7 if you totally agree. You may use the numbers from 2 to 6 to graduate your response if you neither totally agree nor totally disagree, depending on how much you agree or disagree with the statement.” The item wordings and the mean scores by country, after deleting mischievous respondents (see below), are reported in Table A1.

The reliability and validity of survey research is threatened by response biases, respondent carelessness and, especially, mischievous respondents (MRs), i.e. respondents who knowingly make phony responses meant to cheat the researcher (Hyman & Sierra, 2012). Whereas the two former can be reduced through careful research and questionnaire designs (Dillman, Smyth, & Christian, 2009), good design alone cannot eliminate MRs. According to Hyman and Sierra (2012), it is virtually guaranteed that MRs comprise a subset of any survey sample, especially when respondents are paid for participating in the survey (irrespective of whether the payment is money or course credit). Hence, the survey data were cleared of MRs by means of Hyman and Sierra’s (2012) distribution-free, sample-size-unconstrained, backward-stepping MR algorithm using the lowest variance deletion rule. Specifically, respondents were considered mischievous if the variance of their responses to the 69 items in the FRL instrument was below 0.30 (6.1% of the sample overall, varying from 2.4% in Finland to 12.3% in France). This reduced the active sample size to N = 3216.

3.3. Food-related innovativeness and sustainable everyday food consumption Reisch et al. (2013, p. 1) argue that the most effective ways of reducing the environmental impact of food consumption are ‘‘to reduce consumption of meat and dairy products (especially beef), to favour organic fruits and vegetables, and to avoid goods that have been transported by air on both individual and institutional levels . . .” As a measure of their everyday sustainable food consumption, participants in the final part of the interview responded to five items capturing the two first of these (i.e., meat avoidance 1 Due to a mistake by the market research company, a Danish sample about three times the size of the other samples was collected. In order to avoid biases emanating from this difference in sample size, a random sample of the Danish respondents of the same size as the other country samples was drawn and used in the analyses.

4. Results 4.1. Measurement invariance test In the first step, the cross-national measurement invariance of the 23 lifestyle dimensions was investigated by means of multigroup confirmatory factor analysis (CFA, AMOS22), following the procedure applied by Scholderer et al. (2004). Five items were deleted (marked with a superscript ‘‘d” in Table A1) because in some countries, they had a low factor loading. For each of the five lifestyle elements, fit indices of the multi-group CFAs for the final models after these deletions are reported in Table 1.

19

J. Thøgersen / Food Quality and Preference 55 (2017) 16–25 Table 1 Analysis of the cross-cultural validity of the FRL instrument: goodness-of-fit statistics. No. Model

c2

df

RMSEA

AIC

TFI

CFI

Ways of shopping (6 factors) A1.1 Configural invariance A1.2 Metric invariance A1.3 Scalar invariance

2485 2627 4931

760 832 967

0.02 0.02 0.03

3718 3710 5732

0.89 0.90 0.80

0.92 0.92 0.82

Quality aspects (6 factors) A2.1 Configural invariance A2.2 Metric invariance A2.3 Scalar invariance

2673 2877 4462

900 981 1125

0.02 0.02 0.03

3973 4007 5290

0.92 0.93 0.89

0.94 0.94 0.89

Cooking methods (6 factors) A3.1 Configural invariance A3.2 Metric invariance A3.3 Scalar invariance

1843 2054 4441

750 831 966

0.02 0.02 0.03

3098 3139 5244

0.92 0.92 0.81

0.95 0.94 0.82

Consumption situations (2 factors) A4.1 Configural invariance 375 A4.2 Metric invariance 463 A4.3 Scalar invariance 1601

80 116 170

0.03 0.03 0.05

762 777 1805

0.88 0.91 0.73

0.94 0.93 0.70

Purchasing motives (3 factors) A5.1 Configural invariance 1162 A5.2 Metric invariance 1381 A5.3 Scalar invariance 2643

240 294 375

0.03 0.03 0.04

1779 1887 2982

0.82 0.83 0.72

0.88 0.86 0.71

The values of the chi-square relative to degrees of freedom and the RMSEA show that all five basic factor analysis models (i.e., their factor configuration) are acceptable and, hence, possess configural invariance. Also, many fit indices improve when restraining factor loadings to be equal across countries, which shows that the models also possess metric invariance (Steenkamp & Baumgartner, 1998). However, when constraining measurement intercepts to be identical across all countries (i.e., assuming scalar invariance), all the fit indices deteriorate substantially. Hence, the FRL instrument is not invariant across countries with respect to item intercepts (cf. also Scholderer et al., 2004). Some nation-specific response biases are thus caused by something other than variations in the underlying factor. Hence, the metric invariance model is used as point of departure for the following analyses. 4.2. Identifying national and cross-national FRL segments Based on the factor weights produced by the metric invariance model, factor scores were calculated for each of the 23 dimensions of the final FRL instrument for each country separately. These factor scores were used to identify national and possible crossnational FRL segments; this was done by means of multilevel latent class cluster analysis using Latent Gold 5.1 (Vermunt, 2003; Vermunt, 2008). To identify the best fitting model, 45 possible models were estimated, varying the number of clusters from two to ten and the number of country groups from one to five. Latent Gold estimates a range of information-based fit criteria, but in the present case they all move in concert, so it suffices to report one. Hence, the one most commonly used, the Bayesian Information Criterion based on the log-likelihood (BICLL), is reported in Table 2. This information criterion is often used together with the estimated classification error to pick the best model. The classification error increases with the number of clusters and country classes therefore favouring a more parsimonious model. Since parsimony also makes the model more interpretable, a more parsimonious model is preferable as long as the costs in terms of model fit are insignificant. Regarding model fit, the rule of thumb is that the model with the lowest BICLL (and similar information criteria) should be selected. However, when using large samples, information criteria such as the BICLL tend to show a continuous decline as the number of clusters increases (Paas, 2014). As can be seen in Table 2,

Table 2 The Bayesian Information Criterion based on the log-likelihood (BICLL) for models with 1–5 country classes and 2–10 clusters. Clusters/C. Classes

1 Country class

2 Country classes

3 Country classes

4 Country classes

5 Country classes

2 Clusters 3 Clusters 4 Clusters 5 Clusters 6 Clusters 7 Clusters 8 Clusters 9 Clusters 10 Clusters

207708 201146 195573 192126 190324 189011 187760 186659 185907

207505 200880 195239 191811 190017 188670 187393 186446 185473

207496 200845 195173 191742 189958 188628 187361 186198 185463

207505 200852 195158 191739 189982 188636 187457 186248 185420

207519 200867 195163 191789 189947 188676 187346 186274 185415

this is also the case here, although the BICLL decreases at a slower pace with an increasing number of clusters. Hence, with a sample this large, it is necessary to strike a balance between information criteria (i.e., BICLL) and classification error, in order to achieve a reasonably parsimonious and thereby interpretable model. Therefore, in the present case the ‘‘minimize BICLL” criterion is replaced by the more pragmatic criterion ‘‘the decrease in BICLL produced by a further increase in the number of clusters is marginal.” Specifically, the ‘‘marginal decrease in BICLL” criterion is operationalized as ‘‘a less than one percent decrease in BICLL”, which leads to favouring a five-cluster solution (i.e., BICLL for a five-cluster solution is more than one percent lower than for a four-cluster solution and less than one percent higher than for a six-cluster solution) irrespective of the number of country groups. With the number of FRL segments specified at five (i.e., a five-cluster solution), BICLL decreases when moving from one to three country classes, it is practically identical for three and four country classes, and it increases when increasing the number of country classes even further. Hence, when combining this and the classification error criteria, the preferred model becomes the five-clusters (i.e., FRL-segments), three country-classes solution (BICLL = 191742; classification error = 0.05). Table 3 presents the parameter estimates for this model. The upper part of this table shows that four of the ten countries belong to Country Class (CC) 1: Spain, Italy, Poland and Hungary, four to CC 2: Germany, France, Netherlands, and the UK, and two countries belong to CC 3: Denmark and Finland. Hence, it appears that there are three main food-related lifestyle regions in Europe: a central region, a northern region and a south-and-eastern region. It also appears that these three country classes have very different distributions of FRL segments. The smallest country class, CC 3, is furthest from the general average, segments 2 and 1 being heavily overrepresented and segments 3 and 5 heavily underrepresented. The central European CC 2 is closest to the general average while in CC 1, segment 3 is heavily overrepresented and segment 2 heavily underrepresented. The lower part of Table 3 reports the means of the 23 FRL dimensions for each of the FRL segments. As expected, the table shows that there are large differences in the scores on lifestyle dimensions between FRL segments. A major distinction seems to be how interested and engaged consumers are in food and food provision, how important food and food provision are to them. The segment profiles in terms of FRL dimensions are summarized in Table 4, together with profiles in terms of background characteristics and their differences with regard to more sustainable food choices and food innovativeness. At this stage, we can conclude that there are large differences in lifestyle dimensions across FRL segments and in lifestyles across country classes. The analyses producing the latter results are presented below.

20

J. Thøgersen / Food Quality and Preference 55 (2017) 16–25

Table 3 Country class, FRL-segment conditional on country class and FRL segment-specific response probabilities obtained with the model with three country classes and five FRL segments.

Country Class Country Class Country Class FRL segments

1 2 3 share

Country Class Size

Cluster1

Cluster2

Cluster3

Cluster4

Cluster5

0.39 0.39 0.21

0.35 0.34 0.44 0.37

0.09 0.25 0.38 0.22

0.27 0.11 0.05 0.16

0.16 0.18 0.12 0.16

0.13 0.12 0.01 0.10

3.80 4.13 1.40 5.70 4.91 5.55 5.02 3.60 4.58 2.17 1.62 2.23 3.10 3.57 4.31 1.95 1.80 4.78 4.04 3.36 3.63 2.92 4.50

2.80 2.88 1.40 4.37 3.30 4.33 3.63 2.15 3.72 2.08 2.56 2.62 2.37 2.80 2.66 1.67 1.77 3.99 3.33 2.11 2.56 2.21 3.01

4.54 5.04 1.62 6.27 5.92 6.35 5.99 4.64 5.09 2.77 1.35 2.63 3.72 4.33 5.49 2.48 2.22 5.53 4.68 4.65 4.62 3.94 5.84

3.23 3.53 1.63 4.41 4.05 4.35 4.17 3.20 3.55 3.16 2.24 3.28 3.12 3.67 3.76 2.32 2.62 4.23 3.48 3.46 3.55 3.46 4.14

3.99 4.50 2.03 5.22 5.05 5.25 5.15 4.18 4.25 4.17 2.53 4.28 3.84 4.63 4.68 2.99 3.45 5.14 4.09 4.50 4.51 4.57 5.25

Means Social relationships Self-fulfilment in food Security Freshness Health Price/quality relation Novelty Organic Taste Woman’s task Interest in cooking (neg.) Convenience Whole family Planning Looking for new ways Social event Snacks vs meals Shopping list Price criteria Specialty shops Enjoyment from shopping Attitude to advertising Importance of product information

Note: Country Class 1 = Poland, Spain, Italy and Hungary. Country Class 2 = Germany, France, the Netherlands, UK. Country Class 3 = Denmark and Finland.

Table 4 FRL segment profiles in terms of FRL dimensions, background characteristics and food-related outcome variables. Segment

Profile in terms of FRL dimensions

Background characteristics

Meat avoidance

Organic food

FRL segment 1 (37% of the sample): ‘‘everyday food providers”

Generally score average on most FRL dimensions. They score particularly low on convenience, snacking, and woman’s task. High on interest in cooking.

+0

0

FRL segment 2 (21% of the sample): ‘‘food ignoramuses”

Generally score lowest on most FRL dimensions, except convenience and taste, where they are more average.

+

–

FRL segment 3 (16% of the sample): ‘‘enthusiastic food consumers”

Score highest on quality aspects, interest in food shopping, interest in cooking and new ways.

++

+

+

FRL segment 4 (16% of the sample): ‘‘uninvolved food consumers”

Generally score average to low on all FRL dimensions. They score particularly low (but not lowest) on quality dimensions and interest in shopping.

0

0

0

FRL segment 5 (10% of the sample): ‘‘traditional family oriented food consumers”

Score average to high on most FRL dimensions, highest on woman’s task, convenience, whole family, planning, social events, snacking and advertisement, but low interest in cooking,

Women, Older Fewer children Average singles Fewer home owners Car, Mid urban Men, Older Fewest children Most singles Fewest home owners Fewest cars Mid urban Women, Older More children Fewest singles Home owners Car, Mid urban Men, Younger More children Average singles Fewer home owners Car, Least urban Men, Younger Most children Fewest singles Home owners Multiple cars Most urban

+

+

++

4.3. Further profiling segments The FRL segments are profiled further by investigating the relationship between class membership and various background characteristics. This is done by means of a three-step approach in which the estimated latent class model was the first step. In step 2, indi-

Food innovativeness

viduals are assigned to latent classes using their class membership probabilities. (For the present purpose, individuals are assigned to the class with the largest membership probability, i.e., modal classification.) Finally, in step 3, the association between the assigned class memberships and relevant background characteristics are investigated. For analysing background characteristics as covariates,

J. Thøgersen / Food Quality and Preference 55 (2017) 16–25

the step 3 submodule in Latent Gold 5.1 yields a multivariate logistic regression analysis while correcting for classification errors (Vermunt, 2010). Distributions and means for each of the significant covariates as well as Wald tests are reported in Table 5. It appears from Table 5 that the FRL segments differ significantly on a range of background characteristics. In FRL segments 1 and 3 women are overrepresented and in FRL segments 2, 4 and 5 men are overrepresented. Members of FRL segment 1, 2 and 3 tend to be older and younger in segments 4 and 5. The likelihood of having children in the household is lowest in FRL segment 2 and highest in segments 5, 4 and 3. The likelihood of being single is highest in segment 2 and lowest in segments 3 and 5. FRL segments 3 and 5 are most likely to own their home. The likelihood of owning more than one car is particularly high in segment 5. Segment 5 is the most and segment 4 the least urban. These results are summarized in the ‘‘background characteristics” column in Table 4.

4.4. The influence of FRL and country class on food innovativeness and sustainable food choice The joint effects of country of residence (i.e., country class) and FRL on sustainable food choices and environmentally-friendly food innovativeness were analysed in a general linear model (GLM, SPSS22). The analysis revealed that the outcome variables vary significantly across FRL segments, cf. Table 6. Further, after controlling for FRL, the direct effect of country class is highly significant for meatless suppers and marginally significant for buying organic food, but non-significant for eating beef and food innovativeness. The country class  FRL segment interactions are significant as well, except with regard to meatless suppers their impact is marginal. The significant interactions with regard to meatless suppers are especially due to segment 5 being the one that most frequently has meatless suppers in CC 1 and 2, but least frequently in CC 3. Table 5 Distributions and means for each of the significant covariates and Wald tests for differences between FRL segments. Covariates

FRL1

FRL2

FRL3

FRL4

FRL5

Wald

p

Gender Female Male

0.59 0.41

0.46 0.54

0.66 0.34

0.39 0.61

0.40 0.61

118.9903

<.001

Age 18–29 30–39 40–49 50–59 60–68 Mean age

0.17 0.20 0.19 0.22 0.23 43.53

0.16 0.18 0.18 0.22 0.26 44.39

0.12 0.19 0.20 0.24 0.25 44.98

0.29 0.26 0.19 0.16 0.11 37.79

0.29 0.29 0.20 0.14 0.08 36.69

134.9587

<.001

Children Yes

0.33

0.27

0.40

0.41

0.51

19.8135

<.001

Single Yes

0.38

0.47

0.33

0.41

0.34

15.6745

.004

Own home Yes

0.66

0.59

0.74

0.64

0.71

22.4812

<.001

Single family house Yes 0.53

0.55

0.51

0.53

0.54

16.3793

.003

Number of cars 0 0.18 1 0.52 2 0.26 3 or more 0.04 Mean 1.16

0.23 0.53 0.21 0.03 1.05

0.15 0.52 0.29 0.05 1.24

0.14 0.48 0.31 0.06 1.29

0.10 0.45 0.36 0.08 1.58

32.3069

<.001

Place of residence Capital (1) 0.16 Urban (2) 0.68 Rural (3) 0.16

0.18 0.68 0.14

0.16 0.68 0.16

0.12 0.70 0.18

0.18 0.70 0.12

13.9431

.008

21

It also appears from the F-values reported in Table 6 that FRL accounts for a substantially larger variation than country class in all outcome variables except meatless suppers. It also appears that meatless suppers are most common in CC 1 (South and East) and least common in CC 3 (the two Nordic countries) whereas consumers in CC 2 (central Europe) tend to buy less organic food than in the two other country classes. FRL segments 3 and 5 most often buy organic food, most often have a meatless supper (except for CC 3) and are the most innovative with regard to environmentally-friendly food products. FRL segment 2 scores lowest in all of these respects. Segments 1, 2 and 3 eat beef least often while segment 5 eats beef most often. In sum, FRL segment 3, the ‘‘enthusiastic food consumers”, seem to be not only most engaged in food and food provision but also the most sustainability-oriented food consumers. 5. Discussion This paper presents a food-related lifestyle segmentation study based on survey data from representative samples of consumers from ten European countries. Theoretically, the study is based on Grunert’s (1993) well-established, cognitive approach to foodrelated lifestyle segmentation, but methodologically a new, multi-level approach is applied to the segmentation of food consumers from multiple countries. The FRL instrument was obtained from Grunert’s research group and thoroughly screened and tested for construct reliability and measurement invariance across countries, following a state-of-the-art procedure (Scholderer et al., 2004). After discarding items that appeared to lack sufficient cross-national measuring invariance, the final 23 FRL dimensions were used to segment consumers in the ten countries using multi-level latent class analysis. The analysis confirmed that the revised FRL instrument is able to identify meaningful crossnational lifestyle segments in the countries covered, but that the pattern of segments differs significantly between different European regions (Central Europe vs. Northern Europe vs. Southern and Eastern Europe). The results also show that the segments identified differ significantly and substantially with regard to the sustainability of everyday food choices and also with regard to their openness to new environmentally-friendly food products (i.e., environmentally-friendly food product innovativeness). This shows that the FRL segmentation instrument possesses predictive validity with regard to important consumer responses. Further, it confirms the instrument’s practical relevance as a segmentation tool for food marketers as well as campaigners promoting a more sustainable food consumption pattern. 5.1. Conclusions and implications The current study presents and validates a new approach for multi-national segmentation in terms of food-related lifestyle. The constructs measured by the proposed instrument are derived from previous research and tested for construct validity and reliability using factor analysis and for metric measurement invariance across the ten countries using multi-group confirmatory factor analysis. The multi-level latent class analysis revealed a meaningful segmentation of consumers in ten European countries and that the segmentation pattern is shared across countries within the same (broad) region of Europe, but not across Europe as a whole. The proposed instrument also has predictive validity with regard to relevant outcome variables, namely everyday sustainable food choices and consumer innovativeness with regard to environmentally-friendly food products. This further reinforces both the theoretical and the practical relevance of the FRL instrument. Lifestyle segmentation is useful for obtaining a better understanding of consumers, in this case food consumers. However, its

22

Table 6 Differences between FRL segments in terms of sustainable food consumption (meat and organic food) and innovativeness with regard to environmentally-friendly food products. Beef for supper. R2-adj. = 0.10 Mean

a

2.33 2.39a 2.41a 2.77 3.32 2.50a 2.51a 2.59a

S.E.

t/F

3.00

0.39 0.40 0.40 0.43 0.41

7.613 25.775 1.167 1.072 1.556 0.155

<0.001 <0.001 0.243 0.284 0.120 0.877

0.40 0.40

0.233 0.954 0.639

0.792 0.340 0.523

0.41 0.41 0.42 0.41 0.44 0.44 0.42 0.42

2.130 1.820 1.114 1.524 1.372 0.773 0.139 1.243 1.182

0.030 0.069 0.265 0.128 0.170 0.439 0.890 0.214 0.237

0.46 0.43 0.67 0.06

p

0⁄ 0.38 0.26 0⁄ 0.74 0.46 0.63 0.56 0.34 0.06 0.52 0.50

Mean

a

2.90 2.54 3.32b 2.96a 3.51b 3.32 2.85 2.49

Organic food. R2-adj. = 0.15

B

S.E.

t

p

2.33

0.42 0.43 0.43 0.46 0.44

5.496 14.012 0.343 0.097 1.247 0.881

<0.001 <0.001 0.731 0.923 0.212 0.378

0.43 0.43

39.313 2.810 2.740

<0.001 0.005 0.006

0.44 0.44 0.45 0.44 0.47 0.48 0.45 0.45

3.364 0.846 2.053 1.449 2.373 1.504 1.946 1.687 2.317

0.001 0.398 0.040 0.148 0.018 0.133 0.052 0.092 0.021

0.15 0.04 0.58 0.39 0⁄ 1.21 1.19 0⁄ 0.37 0.90 0.65 1.05 0.71 0.93 0.76 1.05

Mean

a

2.59 1.93 3.12b 2.61a 3.26b 2.69a 2.50 2.64a

Note: Means marked by same letter are not significantly different between clusters or country classes at the 0.05 level (Bonferroni). 1 For tests of between-subjects effects, F-values are reported in this column/row, else t-values. CC = Country class. 2 This parameter is set at zero because it is redundant.

Food innovativeness. R2-adj. = 0.34

B

S.E.

t

p

3.28

0.45

0.34 1.13 0.22 0.63 0⁄

0.45 0.46 0.49 0.47

7.288 74.227 0.758 2.484 0.444 1.351

<0.001 <0.001 0.448 0.013 0.657 0.177

0.46 0.46

2.740 0.022 0.055

0.065 0.982 0.956

0.46 0.47 0.47 0.47 0.50 0.51 0.48 0.48

1.976 1.149 0.762 0.563 0.753 0.820 0.617 0.087 0.204

0.046 0.251 0.446 0.573 0.452 0.412 0.537 0.931 0.838

0.01 0.03 0⁄ 0.53 0.36 0.27 0.35 0.41 0.31 0.04 0.10

Mean

B

S.E.

t

p

0.84

0.33 0.34 0.34 0.36 0.35

2.533 202.747 2.727 4.989 0.119 2.250

0.011 <0.001 0.006 <0.001 0.905 0.025

0.15 0.86 0.72 0.16 0.89

0.92 1.68 0.04 0.78 0⁄

0.19 0.06 0.29

0.08 0.01 0⁄

0.34 0.34

0.113 0.224 0.044

0.893 0.823 0.965

0.23 0.07 0.06 0.12 0.22 0.04 0.07 0.06

0.34 0.35 0.35 0.35 0.37 0.38 0.36 0.36

2.573 0.658 0.206 0.159 0.345 0.584 0.102 0.198 0.159

0.008 0.511 0.837 0.874 0.730 0.559 0.919 0.843 0.874

J. Thøgersen / Food Quality and Preference 55 (2017) 16–25

Intercept [clu] 1 [clu# = 1] [clu# = 2] [clu# = 3] [clu# = 4] [clu# = 5] [CC] [CC# = 1]2 [CC# = 2] [CC# = 3] [clu⁄CC] [clu# = 1] ⁄[CC# = 1] [clu# = 1] ⁄[CC# = 2] [clu# = 2] ⁄[CC# = 1] [clu# = 2] ⁄[CC# = 2] [clu# = 3] ⁄[CC# = 1] [clu# = 3] ⁄[CC# = 2] [clu# = 4] ⁄[CC# = 1] [clu# = 4] ⁄[CC# = 2]

Meatless supper. R2-adj. = 0.12 1

B

23

J. Thøgersen / Food Quality and Preference 55 (2017) 16–25

primary use is for campaign planning, both in the commercial and in the public service sector. The present study focuses on the implications of food-related lifestyle for everyday sustainable food choices and for consumer innovativeness with regard to environmentally-friendly food products, i.e., how open and willing they are to accept new opportunities for environmentally-friendly food consumption. A strong impact of food-related lifestyle was found especially with regard to the latter, which strongly supports the usefulness of the FRL instrument for creating more effective campaigns in this important area, through a more effective segmentation of the target groups for such campaigns. However, the FRL segmentation is not only valuable for public service or social marketing campaigns. Commercial marketers in the food markets and a wide range of food-related products might find it equally useful. Acknowledgements This research is part of the large-scale project ‘Sustainable Lifestyles 2.0: End User Integration, Innovation and Entrepreneurship (EU-InnovatE)’. The project has received funding from the European Union’s Seventh Framework Programme for research, technological development and demonstration under grant agreement no 613194. The survey instrument used for this study was developed by the author with input from Geertje Schuitema, lecturer in Consumer Behaviour and Technology Adoption, University College Dublin, and WP1 leader Professor André Habisch and his team at the Katholische Universität Eichstatt-Ingolstadt. The Danish

translation of the survey instrument was done by MA Merete Elmann and Senior clerk Birgitte Steffensen (Diploma in LSP, English), Aarhus University. All other translations were done by KERN AG, Sprachendienste, Ingolstadt, directed by Project Manager, MA Ulyana Lohinava. Final control to ensure that the questions had the same meaning in all the covered languages was done by the following native language speakers: Giacomo Cattaneo, Maria Csutora, André Habisch, Liisa Lähteenmäki, René Rohrbeck, Boleslaw Rok, Geertje Schuitema, Helle Søndergaard, and John Thøgersen. I am grateful to all these collaborators, translators and controllers for their invaluable help in securing the quality of the data collection. The survey data were collected by the market research company YouGov, sampling respondents from their own and partners’ panels in the other countries, administered the data collection as CAWI interviews and organized and presented the data in SPSS files. At YouGov, the proposal was developed by Research Team Manager Katrine Ludvigsen and the survey was set up and managed by Senior Project Manager Dorthe Sortkjær, assisted by Research Consultant Kristian Preben Nielsen. I am grateful to them all and to YouGov in general for the competent work and the excellent collaboration throughout this rather complicated data collection task. I am also grateful to Senior clerk Birgitte Steffensen (Diploma in LSP, English), Aarhus University for correcting the English of this article. Appendix

Table A1 Transport-related lifestyle responses by country. Code q1_1 q1_2 q1_3 q1_4 q1_5 q1_6 q1_7 q1_8 q1_9d q1_10 q1_11 q1_12 q1_13 q1_14 q1_15d q1_16 q1_17 q1_18 q1_19d q1_20 q1_21 q1_22 q1_23 q1_24 q1_25 q1_26 q1_27

I prefer to buy natural products, i.e. products without preservatives To me the naturalness of the food that I buy is an important quality I try to avoid food products with additives I always try to get the best quality for the best price I compare prices between product variants in order to get the best value for money It is important for me to know that I get quality for all my money I love to try recipes from foreign countries I like to try new foods that I have never tasted before Well-known recipes are indeed the best I always buy organically grown food products if I have the opportunity I make a point of using natural or organic food products I don’t mind paying a premium for organic products I find taste in food products important When cooking, I first and foremost consider taste It is more important to choose food products for their nutritional value rather than for their taste I prefer fresh products to canned or frozen products It is important to me that food products are fresh I prefer to buy fresh meat and vegetables rather than pre-packed I like to have ample time in the kitchen Cooking is a task that is best over and done with I don’t like spending too much time on cooking I like to try out new recipes I look for ways to prepare unusual meals Recipes and articles on food from other culinary traditions make me experiment in the kitchen Frozen foods account for a large part of the food products I use in our household We use a lot of ready-to-eat foods in our household I use a lot of mixes, for instance baking mixes and powder soups

DE N = 351

DK N = 380

ES N = 356

FI N = 349

FR N = 355

HU N = 353

IT N = 355

NL N = 342

PL N = 367

UK N = 335

4.56

4.55

5.23

4.19

5.08

5.56

5.36

4.23

5.40

4.57

4.97

4.78

5.27

4.50

5.16

5.35

5.61

4.42

5.00

4.70

4.46 5.15 5.16

4.22 5.56 5.15

4.70 5.56 5.28

5.32 5.19 5.25

4.74 5.37 5.24

5.19 5.70 5.44

5.20 5.82 5.47

4.07 5.14 5.02

4.94 5.29 5.19

4.35 5.58 5.28

5.49

5.34

5.61

4.01

5.43

5.81

5.58

5.53

5.43

5.57

4.68 4.87 4.32 3.89

4.50 4.90 3.76 3.60

5.00 5.12 4.85 4.16

4.58 3.94 3.63 4.19

4.78 5.02 4.52 3.71

4.90 5.17 4.53 3.34

4.73 5.22 4.62 4.47

4.86 4.73 4.16 3.26

5.05 5.23 4.35 4.18

4.81 4.82 4.33 3.45

3.70

4.17

4.44

3.59

4.59

4.53

5.14

3.72

4.69

3.86

4.13 5.86 5.56 3.58

4.07 5.81 5.22 3.35

4.18 5.84 5.36 4.47

5.83 5.31 3.62 5.56

3.98 5.48 5.32 4.01

3.82 6.06 5.39 3.95

4.54 5.96 5.45 4.44

3.57 5.42 5.16 3.51

4.34 5.75 5.50 3.81

3.64 5.59 5.21 4.03

5.19 5.52 5.12

5.32 5.60 5.19

5.56 5.61 5.69

5.56 5.06 4.17

5.19 5.37 5.11

6.02 6.09 6.01

5.57 5.82 5.63

5.02 5.24 5.00

5.53 5.81 5.72

5.12 5.36 5.04

5.20 3.75 3.85 5.10 4.38 4.58

4.72 3.58 3.64 5.08 3.68 4.23

4.59 3.48 3.91 5.35 4.60 4.73

4.07 4.37 4.68 3.49 3.87 3.46

4.92 3.67 3.88 5.05 4.47 4.71

5.35 3.31 4.12 5.35 5.03 4.64

4.97 4.00 3.68 5.52 4.65 4.86

4.27 3.64 3.62 4.87 4.92 4.56

4.14 3.13 3.83 5.21 4.74 4.74

4.63 3.84 4.06 4.88 4.11 4.28

3.02

3.21

3.62

2.81

3.93

2.75

3.60

3.42

3.36

3.85

2.83 3.19

2.54 2.35

3.59 3.50

2.45 3.35

3.45 3.12

3.11 2.58

3.13 3.55

2.73 3.29

3.09 2.84

3.45 3.28

(continued on next page)

24

J. Thøgersen / Food Quality and Preference 55 (2017) 16–25

Table A1 (continued) Code q1_28

q1_29 q1_30 d

q1_31 q1_32 q1_33 q1_34 q1_35

q1_36d

q1_37 q1_38

q1_39 q1_40 q1_41 q1_42 q1_43 q1_44 q1_45 q1_46 q1_47 q1_48 q1_49 q1_50 q1_51 q1_52 q1_53 q1_54 q1_55 q1_56 q1_57 q1_58 q1_59 q1_60 q1_61 q1_62 q1_63 q1_64 q1_65 q1_66 q1_67 q1_68 q1_69

The kids or other members of the family always help in the kitchen, for example they peel potatoes and cut vegetables My family helps with other mealtime chores, such as setting the table and doing the dishes When I do not really feel like cooking, I can get one of the other members of my family to do it What we are going to have for supper is often a lastminute decision Cooking needs to be planned in advance I always plan what we are going to eat a couple of days in advance I consider the kitchen to be the woman’s domain It is the woman’s responsibility to keep the family healthy by serving a nutritious diet Nowadays the responsibility for shopping and cooking ought to lie just as much with the husband as with the wife Being praised for my cooking adds a lot to my selfesteem Eating is to me a matter of touching, smelling, tasting and seeing, all the senses are involved. It is a very exciting sensation I am an excellent cook I dislike everything that might change my eating habits I only buy and eat foods which are familiar to me A familiar dish gives me a sense of security Dining with friends is an important part of my social life When I have friends over for dinner, the most important thing is that we are together Over a meal one may have a lovely chat To me product information is of high importance. I need to know what the product contains I compare labels to select the most nutritious food I compare product information labels to decide which brand to buy I have more confidence in food products that I have seen advertised than in unadvertised products I am influenced by what people say about a food product Information from advertising helps me to make better buying decisions Shopping for food does not interest me at all I just love shopping for food Shopping for food is like a game to me I do not see any reason to shop in specialty food stores I like buying food products in specialty stores where I can get expert advice I like to know what I am buying, so I often ask questions in stores where I shop for food I always check prices, even on small items I notice when the price of products I buy regularly changes I look for ads in the newspaper for store specials and plan to take advantage of them when I go shopping Before I go shopping for food, I make a list of everything I need I make a shopping list to guide my food purchases Usually I do not decide what to buy until I am in the shop I eat before I get hungry, which means that I am never hungry at meal times I eat whenever I feel the slightest bit hungry In our house, nibbling has taken over and replaced set eating hours Going out for dinner is a regular part of our eating habits We often get together with friends to enjoy an easyto-cook, casual dinner I enjoy going to restaurants with my family and friends

DE N = 351

DK N = 380

ES N = 356

FI N = 349

FR N = 355

HU N = 353

IT N = 355

NL N = 342

PL N = 367

UK N = 335

3.85

3.63

4.03

3.95

3.74

4.25

4.12

3.37

4.14

3.30

4.51

4.66

5.09

3.88

4.30

4.90

4.92

4.18

4.53

4.04

3.81

3.94

4.63

3.36

3.85

3.73

4.95

3.87

4.55

3.79

3.75

3.82

4.06

4.52

4.01

4.47

4.28

3.40

4.63

3.87

4.28 3.95

4.28 3.34

4.09 4.19

3.73 2.38

4.34 4.05

4.81 3.75

4.68 3.85

4.27 3.94

4.42 3.53

4.35 3.94

2.86 3.13

2.10 2.17

2.62 3.08

2.50 5.82

3.17 3.39

3.22 4.02

3.82 4.12

2.62 3.12

3.38 3.75

3.04 3.15

5.58

5.81

5.83

4.12

5.47

4.99

4.76

5.51

5.40

5.24

4.71

4.36

4.94

4.88

4.80

4.80

5.34

5.17

5.38

4.67

5.49

4.67

5.32

4.05

4.99

5.28

5.33

4.59

5.05

4.75

4.38 2.88

4.33 2.69

4.44 3.85

2.64 3.30

4.43 3.43

4.42 3.06

4.66 3.73

4.16 3.20

4.58 3.62

4.29 3.34

3.96 4.64 4.15

3.25 4.45 4.21

4.35 5.20 4.89

4.61 3.82 4.64

4.15 4.80 4.30

4.22 4.72 4.16

4.34 4.78 4.97

3.47 4.61 3.99

4.11 4.64 3.85

3.78 4.84 4.01

4.81

5.27

5.38

5.05

5.33

5.24

5.30

4.57

5.00

4.66

5.35 4.56

5.48 4.80

5.65 5.07

4.82 4.12

5.03 4.87

5.25 5.27

5.76 5.48

4.97 4.30

5.46 5.12

5.09 4.68

4.07 4.37

3.86 4.31

4.64 4.89

4.32 2.93

4.14 4.69

4.48 4.73

4.95 5.15

3.75 3.96

4.69 4.76

4.16 4.33

2.57

2.18

3.62

3.32

3.28

2.75

3.65

2.67

3.34

3.51

3.33

3.70

4.02

3.76

3.77

4.37

3.48

3.94

4.22

3.90

3.78

3.20

4.08

2.92

3.62

3.82

4.10

3.71

3.76

3.69

2.77 4.87 3.86 3.81

2.78 3.86 3.93 3.31

2.84 4.97 4.35 3.80

3.66 4.20 4.26 2.89

3.56 4.38 4.22 4.15

2.44 4.85 3.97 3.22

3.33 4.91 4.58 3.81

3.18 4.40 3.56 4.24

2.62 4.49 3.98 4.12

3.29 4.41 3.60 4.15

3.72

3.88

4.17

2.96

3.74

3.99

4.48

2.98

4.06

3.46

3.35

3.23

4.57

5.09

3.84

3.91

4.64

3.06

4.26

3.59

4.75 5.31

4.65 5.03

5.24 5.19

4.99 4.51

5.15 5.10

5.24 5.64

5.70 5.48

4.99 4.88

4.70 5.16

5.32 5.29

4.81

3.89

3.83

4.44

4.07

4.61

5.17

4.77

4.02

3.67

4.98

4.78

5.17

4.72

5.06

4.91

5.26

4.87

4.91

4.56

4.98 3.44

5.01 3.26

5.07 3.67

3.26 2.66

5.06 3.63

4.93 3.05

5.18 4.25

4.93 3.75

4.85 3.87

4.63 3.67

2.77

2.40

3.10

3.51

2.86

2.78

3.63

3.12

3.55

3.13

3.27 3.03

3.07 2.26

3.71 2.88

3.08 3.21

3.49 2.79

3.57 2.56

4.09 2.66

3.56 2.20

3.77 3.14

3.70 3.08

3.11

3.31

4.04

2.97

3.88

2.79

3.90

3.32

2.73

3.67

3.56

3.67

4.43

4.86

4.12

3.47

4.61

3.20

3.75

3.46

4.68

5.36

5.38

4.19

5.01

3.98

5.02

4.59

4.06

5.02

J. Thøgersen / Food Quality and Preference 55 (2017) 16–25

References Brunsø, K., Scholderer, J., & Grunert, K. G. (2004). Closing the gap between values and behavior - a means-end theory of lifestyle. Journal of Business Research, 57, 665–670. Cleveland, M., Laroche, M., & Papadopoulos, N. (2009). Cosmopolitanism, consumer ethnocentrism, and materialism: An eight-country study of antecedents and outcomes. Journal of International Marketing, 17(1), 116–146. Cullen, F., & Kingston, H. (2009). Analysis of rural and urban consumer behavior toward new food products using a food-related lifestyle instrument. Journal of Foodservice Business Research, 12(1), 18–41. de Boer, M., McCarthy, M., Cowan, C., & Ryan, I. (2004). The influence of lifestyle characteristics and beliefs about convenience food on the demand for convenience foods in the Irish market. Food Quality and Preference, 15, 155–165. Demby, E. (1974). Psychographics and from where it came. In W. D. Wells (Ed.), Life style and psychographics (pp. 9–30). Chicago, IL: American Marketing Association. Dillman, D., Smyth, J., & Christian, J. (2009). Internet, mail, and mixed-mode surveys: The tailored design method (3rd edition). Hoboken, NJ: Wiley. Field, C. B. et al. (2014). Climate change 2014: Impacts, adaptation, and vulnerability. Volume I: Global and Sectoral Aspects. In Working group II contribution to the fifth assessment report of the Intergovernmental Panel on Climate Change. Cambridge, UK and New York, NY: Cambridge University Press. Goldsmith, R. E., & Hofaker, C. F. (1991). Measuring consumer innovativeness. Journal of the Academy of Marketing Science, 19, 209–221. Grunert, K. G. (1993). Towards a concept of food-related lifestyle. Appetite, 21, 51–55. Grunert, K. G. (2006). Future trends and consumer lifestyles with regard to meat consumption. Meat Science, 74, 149–160. Grunert, K. G., Brunsø, K., & Bisp, S. (1993). Food-related life style: Development of a cross-culturally valid instrument for market surveillance. Aarhus: MAPP; The Aarhus School of Business. Grunert, K. G., Brunsø, K., Bredahl, L., & Bech, A. C. (2001). Food-related lifestyle: A segmentation approach to European food consumers. In L. J. Frewer, E. Risvik, H. N. J. Schifferstein, & R. von Alvensleben (Eds.), Food, people and society: A European perspective of consumers’ food choices (pp. 211–230). London: Springer Verlag. Grunert, K. G., Perrea, T., Zhou, Y., Huang, G., Sørensen, B. T., & Krystallis, A. (2011). Is food-related lifestyle (FRL) able to reveal food consumption patterns in nonWestern cultural environments? Its adaptation and application in urban China. Appetite, 56, 357–367. Gutman, J. (1982). A means-end chain model based on consumer categorization processes. Journal of Marketing, 46(Spring), 60–72. Hoek, A. C., Luning, P. A., Stafleu, A., & de Graaf, C. (2004). Food-related lifestyle and health attitudes of Dutch vegetarians, non-vegetarian consumers of meat substitutes, and meat consumers. Appetite, 42, 265–272. Hyman, M. R., & Sierra, J. J. (2012). Adjusting self-reported attitudinal data for mischievous respondents. International Journal of Market Research, 54(1), 129–145. IPCC (Ed.). (2014). Climate Change 2014: Mitigation of Climate Change. Contribution of Working Group III to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change [Edenhofer, O., R. Pichs-Madruga, Y. Sokona, E. Farahani, S. Kadner, K. Seyboth, A. Adler, I. Baum, S. Brunner, P. Eickemeier, B. Kriemann, J. Savolainen, S. Schlömer, C. von Stechow, T. Zwickel and J.C. Minx (eds.)]. Cambridge, UK and New York, NY: IPCC & Cambridge University Press. Milfont, T. L., & Markowitz, E. (2016). Sustainable consumer behavior: a multilevel perspective. Current Opinion in Psychology, 10, 112–117. Mont, O., Neuvonen, A., & Lähteenoja, S. (2014). Sustainable lifestyles 2050: stakeholder visions, emerging practices and future research. Journal of Cleaner Production, 63, 24–32.

25

Neuvonen, A., Kaskinen, T., Leppänen, J., Lähteenoja, S., Mokka, R., & Ritola, M. (2014). Low-carbon futures and sustainable lifestyles: A backcasting scenario approach. Futures, 58, 66–76. Nie, C., & Zepeda, L. (2011). Lifestyle segmentation of US food shoppers to examine organic and local food consumption. Appetite, 57, 28–37. O’Sullivan, C., Scholderer, J., & Cowan, C. (2005). Measurement equivalence of the food related lifestyle instrument (FRL) in Ireland and Great Britain. Food Quality and Preference, 16, 1–12. Paas, L. J. (2014). Comments on: Latent Markov models: A review of a general framework for the analysis of longitudinal data with covariates. TEST, 23, 473–477. Pérez-Cueto, F. J. A., Verbeke, W., de Barcellos, M. D., Kehagia, O., Chryssochoidis, G., Scholderer, J., et al. (2010). Food-related lifestyles and their association to obesity in five European countries. Appetite, 54, 156–162. Plummer, J. T. (1974). The concept and application of life style segmentation. Journal of Marketing, 38(1), 33–37. Popp, A., Lotze-Campen, H., & Bodirsky, B. (2010). Food consumption, diet shifts and associated non-CO2 greenhouse gases from agricultural production. Global Environmental Change, 20, 451–462. Reisch, L., Eberle, U., & Lorek, S. (2013). Sustainable food consumption: an overview of contemporary issues and policies. Sustainability: Science, Practice & Policy, 9 (2), 1–19. Rokeach, M. (1973). The nature of human values. New York, NY: Free Press. Scholderer, J., Brunsø, K., Bredahl, L., & Grunert, K. G. (2004). Cross-cultural validity of the food-related lifestyles instrument (FRL) within Western Europe. Appetite, 42, 197–211. Schwartz, S. H. (1994). Are there universal aspects in the structure and content of human values? Journal of Social Issues, 50(4), 19–45. Schwartz, S. H. (1992). Universals in the content and structure of values: Theoretical advances and empirical tests in 20 countries. In M. P. Zanna (Ed.), Advances in experimental social psychology (pp. 1–65). San Diego, CA: Academic Press. Speedy, A. W. (2003). Global production and consumption of animal source foods. Journal of Nutrition, 133, 4048S–4053S. Steenkamp, J. B. E. M., & Baumgartner, H. (1998). Assessing measurement invariance in cross-national consumer research. Journal of Consumer Research, 25, 78–90. Steen-Olsen, K., & Hertwich, E. (2015). Life cycle assessment as a means to identify the most effective action for sustainable consumption. In L. A. Reisch & J. Thøgersen (Eds.), Handbook of research on sustainable consumption (pp. 131–144). Cheltenham, UK: Edward Elgar Publishing. Thøgersen, J. (2010a). Country differences in sustainable consumption: The case of organic food. Journal of Macromarketing, 30, 171–185. Thøgersen, J. (2010b). Pro-environmental consumption. In K. M. Ekström (Ed.), Consumer behaviour – a nordic perspective (pp. 95–115). Lund: Studentlitteratur. Tukker, A. (2015). Priorities for sustainable consumption policies. In L. A. Reisch & J. Thøgersen (Eds.), Handbook of research on sustainable consumption (pp. 145–160). Cheltenham, UK: Edward Elgar Publishing. van Raaij, W. F., & Verhallen, T. M. M. (1994). Domain-specific market segmentation. European Journal of Marketing, 28(10), 49–66. Vermunt, J. K. (2003). Multilevel latent class models. Sociological Methodology, 33(1), 213–239. Vermunt, J. K. (2008). Latent class and finite mixture models for multilevel data sets. Statistical Methods in Medical Research, 17, 33–51. Vermunt, J. K. (2010). Latent class modeling with covariates: Two improved +threestep approaches. Political Analysis, 18, 450–469. Vyncke, P. (2002). Lifestyle segmentation: from attitudes, interests and opinions, to values, aesthetic styles, life visions and media preferences. European Journal of Communication, 7, 445–463. Wilk, R. R. (2002). Consumption, human needs, and global environmental change. Global environmental change, 12, 5–13.