Prospects for pro-environmental protein consumption in Europe: Cultural, culinary, economic and psychological factors

Appetite 121 (2018) 29e40

Contents lists available at ScienceDirect

Appetite journal homepage: www.elsevier.com/locate/appet

Prospects for pro-environmental protein consumption in Europe: Cultural, culinary, economic and psychological factors Joop de Boer*, Harry Aiking Institute for Environmental Studies, VU University, Amsterdam, The Netherlands

a r t i c l e i n f o

a b s t r a c t

Article history: Received 8 September 2017 Received in revised form 29 October 2017 Accepted 30 October 2017 Available online 2 November 2017

The current ratio between plant and animal protein in the Western diet is causing serious threats to both public health and the environment. Healthy, pro-environmental protein consumption requires a transition to a diet with more plant protein and considerably less animal protein. The present paper focuses on the prospects of this transition by analyzing consumer responses to some key options in the context of regional differences across Europe. The aim is to assess how responses to the options might be shaped by 1) cultural, culinary and economic spatial gradients (including GDP per capita) at regional level and 2) differences in environmental friendly behavior and gender at individual level. The study, covering all EU members in 2012, compares regional level statistics (food supply data) with individual level statistics (consumer survey data) and vice-versa. The south-north latitude gradient showed a decreasing trend in vegetable and pulse protein supplies and, in parallel, a decreasing trend in positive consumer responses to the key options, probably due to differences in meal experiences. The west-east longitude gradient showed decreasing levels of animal protein supplies and GDP per capita. Individuals' willingness to do something positive for the environment and their gender played a weak but consistent role in the responses. To effectively stimulate diet changes, it is important to seek ways in which culinary and environmental aspects can complement each other and to ensure that diet changes do not depend solely on individual decisions but become an integral part of regional social processes. © 2017 Elsevier Ltd. All rights reserved.

Keywords: Animal protein Plant protein Health Environment Consumers

1. Introduction The current ratio between plant and animal protein in the Western diet is causing serious threats to both public health and the environment (Aiking, 2014; Friel et al., 2009; Mertens, van 't Veer, Hiddink, Steijns, & Kuijsten, 2017; Reynolds, Buckley, Weinstein, & Boland, 2014; Westhoek et al., 2014). The dietary imbalance dates from the unprecedented growth of the intensive (high input) livestock industry during the twentieth-century, which made livestock the chief source of protein in many Western countries (Grigg, 1995a; Swatland, 2010). Although meat and dairy are valuable sources of nutrients, the protein produced by Western feedlot animals is quantitatively and qualitatively much less sustainable than the meat and other animal foods acquired by hunting and herding populations in the distant past. According to several reviews (see above), the world's future protein supply can

* Corresponding author. Institute for Environmental Studies, VU University, De Boelelaan 1087, 1081 HV Amsterdam, The Netherlands. E-mail address: [email protected] (J. de Boer). https://doi.org/10.1016/j.appet.2017.10.042 0195-6663/© 2017 Elsevier Ltd. All rights reserved.

only be ensured through a ‘reversed’ diet transitionda shift to a diet with more plant protein and considerably less animal protein. This would mean that many Western consumers will have to add healthy, pro-environmental protein consumption to the other positive things they do for the environment. Obviously, however, pro-environmental protein consumption is a new concept, which until now has little active support from policy-makers in government, industry, and even environmental NGOs (Laestadius, Neff, Barry, & Frattaroli, 2014). The aim of the present study is to assess how consumer responses to pro-environmental protein options (PPOs) in the European context might be shaped by 1) cultural, culinary and economic spatial gradients (including GDP per capita) at regional level (Grigg, 1995a, 1995b) and 2) differences in environmental friendly behavior and gender at individual level (Clayton & Myers, 2009). In order to make such a cross-regional comparison meaningful, it is necessary to use multiple types of data, which allow comparisons of regional level data with individual level data and vice-versa (Scheuch, 1989; de Boer, Helms, & Aiking, 2006). The present study combines food supply data with consumer survey data, focusing on the main spatial gradients in

30

J. de Boer, H. Aiking / Appetite 121 (2018) 29e40

European food patterns. 1.1. Conceptual model The conceptual model of our approach is presented in Fig. 1. One of the central aspects that require further development is the concept of pro-environmental protein consumption. The literature on environment and behavior shows that there are two general, interrelated types of pro-environmental behavior; the first type reflects a commitment to the overarching environmental goal of ‘using fewer natural resources’ and the second relates to the goal of ‘doing things in a different way’ and with a reduced environmental impact (Clayton & Myers, 2009, p. 144). In the case of protein, both types are relevant. The key background issue is that turning protein from feed crops into animal protein for human consumption is inherently resource-inefficient (Smil, 2000). Due to the large scale of agricultural production, reducing the inefficiency through a ‘reversed’ diet transition back to less animal protein could make a significant difference to the most important human-caused environmental issues, including biodiversity loss, nitrogen cycle disruption, and climate change (Aiking, 2014; Westhoek et al., 2014). As protein intakes of adults in Europe are often above the Population Reference Intake (PRI) (EFSA Panel on Dietetic Products Nutrition and Allergies, 2012), achieving a ‘reversed’ diet transition is a matter of ‘using less’ (e.g. leaving the meat out of the dish) and ‘doing things differently’ (e.g. replacing meat with beans). In fact, there is a broad variety of options for a transition, which are more or less different from existing eating patterns and may or may not involve novel foods (Elzerman, Hoek, van Boekel, & Luning, 2011; € sler, de Boer, & Boersema, 2012; Hartmann & Siegrist, 2017; Scho Vainio, Niva, Jallinoja, & Latvala, 2016). However, due to the overall popularity of meat (e.g. Graça, Calheiros, & Oliveira, 2015; Latvala et al., 2012; Macdiarmid, Douglas, & Campbell, 2016; Pohjolainen, Vinnari, & Jokinen, 2015) and the great variety of

Fig. 1. Conceptual model; the arrow numbers correspond to the research questions.

factors that influence food patterns, including in particular gender €sler, de Boer, (Rozin, Hormes, Faith, & Wansink, 2012; Scho Boersema, & Aiking, 2015), it may take a profound societal effort to achieve a ‘reversed’ diet transition. The prospects for PPOs may significantly vary by factors related to earlier diet transitions, such as cultural, culinary and economic factors related to geographic area variations in cuisine and diet. These background factors reflect differences between regions in physical characteristics and historic development (Montanari, 1994; Rokkan, 1999; Scheuch, 1989). Cuisine styles, such as the Mediterranean diet, are combinations of ingredients, condiments and food preparation procedures that lie within a common historical and territorial context (Turmo, 2012). Due to gradual differences between neighboring areas, in combination with cultural diffusion and adaptation processes, cuisine styles may vary in patterns linked to spatial gradients, such as a south-north latitude gradient and a west-east longitude gradient. Indeed, descriptions in the European historical literature show that there are large culturally and territorially defined differences between the classic (Roman and Greek) versus the Germanic and Celtic world, which contrast southern diets with a dominant vegetable component and northern diets with mainly milk, cheese and meat (Montanari, 1994). Factors such as economic development, greater trade in food, and more uniform regulations have since then contributed to a homogenization of food consumption patterns across Europe (Schmidhuber & Traill, 2006; Scholliers, 2007; Teuteberg & Flandrin, 1999). However, some of the regional food practices show a remarkable degree of historical continuity (Askegaard & Madsen, 1998; Fischler, 1999), which has been reflected by a south-north latitude gradient across the European continent in the direction of lower supplies of protein provided by vegetables and cereals in the north (de Boer et al., 2006; Grigg, 1995a, 1999). There also used to be a south-north gradient across the European continent in the direction of smaller supplies of animal-based protein in the south (de Boer et al., 2006; Grigg, 1995a, 1999). This gradient was partly the result of the twentieth-century nutritional transition that made animals rather than bread the chief source of protein in developed countries. As noted by Grigg (1995a), this transition was prompted by a long term rise in real incomes, which began at different dates in different regions. By the 1960s most countries in Western Europedsave the Mediterranean areadhad gone through this transition. In hindsight, the high-plant and lowmeat, Mediterranean type of diet of the early 1960s has been identified as a valuable reference point for the development of options to create more sustainable, healthful diets in other regions (Dernini & Berry, 2015; Tilman & Clark, 2014; van Dooren & Aiking, 2016). Since the 1960s, however, this Mediterranean type of diet has been supplemented by a higher level of meat protein supply (Leclercq, Arcella, Piccinelli, Sette, & Le Donne, 2009; Moreno, Sarría, & Popkin, 2002; Moreno-Altamirano et al., 2016; Turmo, 2012). In terms of Turmo (2012, p. 128), traditional stews and soups may have been incorporating “abundant animal protein”. The regions in Eastern Europe may not have gone through the nutritional transition yet. At this point it should be noted that, although the focus of the paper is on regions and individuals, it is necessary to use countries as administrative units. The west-east longitude gradient is not only related to differences between coastal and landlocked countries but also to differences between the 15 countries forming the European Union before the enlargements of 2004 and 2007 and the ‘new members’ from Eastern Europe. This difference corresponds to a split in the meat market for consumers, dividing countries with high incomes and high meat price levels from countries with medium or low incomes and meat price levels (GfK EU3C, 2012, p. 258e259). A notable exception to this pattern is the Netherlands, which belongs to the high-income

J. de Boer, H. Aiking / Appetite 121 (2018) 29e40

countries but still has a lower meat price level. According to the market report, the exception can be understood by the prominent position of the Netherlands as a net exporter of pork and poultry, and its position in transit trade as an importer of raw materials from outside the EU, which leads to relatively low feed prices (GfK EU3C, 2012, p. 258). In addition to the west-east gradient, more specific information about average income levels is captured by the purchasing power adjusted Gross Domestic Product (GDP in PPS) per capita. Generally, meat consumption increases with national income but the increase over time tends to be an S-shaped saturation curve (Keyzer, Merbis, Pavel, & van Wesenbeeck, 2005; Sans & Combris, 2015; Vranken, Avermaete, Petalios, & Mathijs, 2014). Given the position of the high-income European countries in the saturated part of the curve, the consumption of meat will not increase much with a continued rise of incomes in these countries (de Boer et al., 2006). However, this may be different in the medium and low income countries. At the individual level, various studies have been done to assess whether consumers have an awareness of the impacts of meat consumption on the environment. In several European and nonEuropean countries, including Australia (Lea & Worsley, 2008), Finland (Latvala et al., 2012; Pohjolainen, Tapio, Vinnari, Jokinen, & €nen, 2016), Flemish Belgium (Vanhonacker, Van Loo, Gellynck, R€ asa & Verbeke, 2013), Germany (Cordts, Nitzko, & Spiller, 2014), the Netherlands (de Boer, de Witt, & Aiking, 2016), Portugal (Graça, Oliveira, & Calheiros, 2015), Switzerland (Siegrist, Visschers, & Hartmann, 2015; Tobler, Visschers, & Siegrist, 2011), the United Kingdom (Apostolidis & McLeay, 2016; Clonan, Wilson, Swift, Leibovici, & Holdsworth, 2015) and the United States (de Boer et al., 2016), significant minorities of consumers surveyed acknowledged the environmental impacts of meat eating. The studies also found that acknowledging these impacts of meat eating was positively associated with a lower level of meat consumption and greater willingness to reduce the current level (de Boer et al., 2016; Graça, Oliveira, et al., 2015; Hartmann & Siegrist, 2017; Latvala et al., 2012; Tobler et al., 2011). However, current work has not yet considered the prospects of pro-environmental protein consumption in a systematic way, taking due account of cultural, culinary, economic and psychological factors. Also, the studies have been done in a few European countries only. This is unfortunate as the report on the meat market for consumers in the European Union shows that the availability and purchases of various meat products vary widely by country and purchase channel (GfK EU3C, 2012). 1.2. Research strategy To provide a broader picture and more analytical insight into relevant issues, the overall research strategy of this study is that regional level data will be compared with individual level data and vice-versa. For this purpose, differences across countries will be examined in relation to their ranking along the south-north and west-east gradients and their GDP ranking. Apart from this ranking, it is useful to combine the south-north and west-east gradients and to categorize the EU member countries into six zones, namely three culinary zones of high-income countries (Northern zone, Western Central zone, and high-income Mediterranean zone) and three culinary zones of medium-to-low-income countries (North-Eastern zone, Eastern Central zone, and medium-to-low-income Mediterranean zone). This framework of cultural, culinary and economic factors enables a systematic analysis of differences in prospects for pro-environmental protein consumption, using data at the regional level (zones and countries) and at the level of consumers within regional areas. In line with the conceptual model (Fig. 1), three research

31

questions were developed to clarify the prospects for the adoption of PPOs. The first research question checks the regional differences in the amounts of different sources of protein supply (arrow 1 in Fig. 1). It focuses on the country level, using food supply data from FAOSTAT, which represent the average per capita quantity of a range of protein sources available for consumption (FAO, 2017). These data can be related to the country's location along southnorth and west-east gradients as well as its GDP ranking. The second research question examines the relation between regional and individual level data (arrows 2 and 3 in Fig. 1). The data at the level of individual consumers were derived from a survey into attitudes of Europeans towards building the single market for green products, which was carried out in 27 member countries on behalf of the European Commission (European Commission, 2013a). The data include consumers' frequency of meat consumption, their responses to some key options for pro-environmental protein consumption, their self-reported general pro-environmental behavior, and their gender. The question is how average frequency of meat consumption (arrow 2) and responses to PPOs (arrow 3) per country are related to its location along the south-north and westeast gradients as well as its GDP ranking. The third research question focuses on the relations between the individual level variables (arrows 4 and 5) in the context of regional level differences (arrows 1 and 3). The question is: (a) What are the relations between consumers' responses to options for pro-environmental protein consumption and the variables gender, frequency of meat consumption, and levels of self-reported pro-environmental actions and (b) how do these relations vary across the different zones? 2. Method 2.1. Data sets FAOSTAT supply data were used to reveal country level differences in the significance of plant-based and animal-based dietary protein (FAO, 2017). The data refer to national per capita supply at retail level (¼ production þ imports  exports) and it should be noted that, due to food wasted beyond retail level, supply data are limited indicators of the food actually consumed by individuals. Across EU-27, estimates of both food consumption and food waste tend to be significantly higher in higher-income countries (Monier et al., 2010, p. 108). However, although the EU countries are synchronizing their national dietary surveys in order to create a combined database (EFSA Panel on Dietetic Products Nutrition and Allergies, 2012), there are as yet no better published data available about specific sources of protein consumption. The FAOSTAT supply data are probably the most used data concerning human nutrition (Smil, 2000) and comparing supply patterns as indicators of consumption patterns is generally considered feasible (Grigg, 1995b; Moreno-Altamirano et al., 2016; Sans & Combris, 2015; de Boer et al., 2006). Following de Boer et al. (2006), detailed and aggregated data on the main sources of daily protein supply (protein content x volume of supply in g/person/day) were collected. The detailed items are cereals (excluding beer), vegetables, potato, pulses, stimulants (primarily cocoa and coffee), fruit, nuts, milk (including cheese), pork, poultry, fish, beef, eggs, offal, mutton, and goat. The aggregated categories are total plant protein, total meat protein, total animal protein, and total protein. As the consumer survey (see below) was carried out at the end of 2012, this year was chosen for the supply data. Also, the same population figures were used for international weighting to aggregate the per capita country data to a regional scale. The second data source is the survey into attitudes of Europeans towards building the single market for green products, carried out by TNS Political & Social network in the 27 member countries of the

32

J. de Boer, H. Aiking / Appetite 121 (2018) 29e40

European Union between 4th December and 10th December 2012 (European Commission, 2013a). Country-based samples of about 1000 respondents (500 in the three smallest countries) were interviewed via telephone (landline and mobile phone) in their mother tongue. The basic sample design applied in all countries is multi-stage random (probability). In each household, the respondent was drawn at random following the ‘last birthday rule’. For each country, TNS Opinion & Social carried out a comparison between the sample and the universe (post weighting). The national weighting procedure, using marginal and intercellular weighting, was carried out based on gender, age, region and size of locality. For international weighting (e.g. comparing multi-country zones), official population figures were applied (European Commission, 2013b). The countries' latitude and longitude coordinates were derived from Central Intelligence Agency (2013). Consumers' willingness to do something positive for the environment was derived from their self-reported participation in general pro-environmental actions. This was assessed by the item “Have you done any of the following during the past month for environmental reasons?” The eight activities (multiple answers possible) were: Chosen a more environmentally friendly way of traveling (on foot, by bicycle, by public transport), reduced the consumption of disposable items (plastic bags, certain kinds of packaging, etc.), separated most of your waste for recycling, cut down your water consumption (for example not leaving water running when washing the dishes or taking a shower), cut down your energy consumption (for example turning down air conditioning or heating, not leaving appliances on stand-by, buying energy saving light bulbs, buying energy efficient appliances), bought environmentally friendly products marked with an environmental label, chosen locally produced products or groceries, used your car less without replacing it by another means of travelling. The activities reflect both types of pro-environmental behavior (i.e. ‘using fewer natural resources’ and ‘doing things in a different way’), but the number of items is too small to develop separate subscales for each type. Therefore, the sum score can be used as a basis for comparing consumers' willingness to do something positive for the environment. Frequency of meat consumption was assessed by the question: “How many times a week do you eat meat (beef, pork, poultry, etc.)?” The answer categories were never, less than once a week, once a week, two or three times a week, four or five times a week, more than five times a week. Participants who reported to eat meat were asked to respond to three options for pro-environmental protein consumption, which were introduced by a statement read out by the interviewer: “Some people say large scale meat production has a negative impact on the environment. Would you be willing to do the following for environmental reasons … ?” The PPOs were: “Replace most of the meat you eat by vegetables” (yes/no), “Eat less meat but of certified origin” (yes/no), “Replace beef or pork with poultry or fish” (yes/no). As answering “yes” to these questions in a telephone survey is relatively easy, the percentage of positive responses should be interpreted cautiously (Van Vaerenbergh & Thomas, 2013). This makes it important to take due account of the differences between the items. Replacing most of the meat by vegetables implies a partial diet shift from animal-based to plant-based protein; eating less meat but of a certified origin and replacing beef or pork with poultry or fish are changes within the animal-based protein category. Therefore, the options will be treated as separate items, which will be subsequently correlated with the other variables. In addition, the measure of self-reported pro-environmental actions will be used to provide a benchmark for correlation values.

2.2. Data analysis The first research question was addressed by calculating the main sources of daily protein supply (g/person) in the six culinary and economic zones. In addition, the amounts of different sources of protein supply per country were ranked in order to calculate rank correlations (Spearman's r) between these rankings and the country's location along the south-north and west-east gradients and its GDP ranking. The rankings were assigned in ascending order. As the 27 countries cannot be understood as a random sample, the statistical significance of the rank correlations has no meaning. To distinguish between major and minor correlations, a nominally significant test result (r ¼ 0.38, n ¼ 27, p < 0.05, two-sided) was used as a reference value. The second research question was treated in the same way with the average per country responses to the survey items as variables. The third research question focuses on the relations between the survey responses. Levels of self-reported pro-environmental actions were measured by counting the number of reported activities, which forms an adequate scale (Cronbach's alpha ¼ 0.62, weighted sample n ¼ 25,568). The variables gender, frequency of meat consumption, self-reported pro-environmental actions, and responses to the PPOs were correlated in each zone, using Pearson's correlation coefficient. Given the large sizes of the samples per zone, the standard errors of the correlations were very small and it was not considered meaningful to test whether any difference between two correlations is statistically significant. All analyses were conducted with SPSS 23 for Windows. 3. Results Table 1 provides an overview of the six culinary and economic zones and the countries categorized into each of the zones. It shows large differences in population size; the Western Central zone had the largest population (187,168,000) and the Northern zone the smallest (20,517,000). The Northern zone had the highest index of GDP per capita (123.8) relative to the European Union average (100), whereas the Eastern Central zone had the lowest level (55.2). The analysis at the country level shows that the west-east longitude gradient often went together with lower levels of GDP (r ¼ 0.60), but the south-north gradient was not much related to GDP (r ¼ 0.26). Table 1 also shows that the differences in GDP between the zones closely corresponded to differences in mean meat price in V per kg (V8.32 versus V3.66, respectively). At the country level, the rank correlation between GDP and mean meat price was very high: r ¼ 0.83. 3.1. Check of the spatial gradients in sources of protein supply (arrow 1 in Fig. 1) Research question 1 focuses on regional differences in the importance of plant-based and animal-based dietary protein. The differences were examined at the level of the six zones and at the level of the 27 countries that can be ranked. The main sources of daily protein supply (g/person) in each of the six zones are shown in Table 2 and the correlations between rankings of the countries are displayed in Table 3. Compared to the overall mean, Table 2 shows that the two Mediterranean zones had a relatively high level of protein supply, whereas the low-income North-Eastern and Eastern Central zones had a lower level. In addition, Table 3 reveals that total protein supply, ranging from the lowest levels (69.1 g (Slovenia) and 80.7 g (Hungary)) to the highest levels (116.4 g (Finland) and 124.2 g (Lithuania)), often went together with higher

J. de Boer, H. Aiking / Appetite 121 (2018) 29e40

33

Table 1 Overview of the 6 culinary and economic zones (countries involved, population size, median south-north latitude and west-east longitude, index of GDP per capita and mean meat price). Zones

High-income zones

Medium-to-low income zones

Total

Northern zone

Western Central zone

Mediterranean zone

North-Eastern zone

Eastern Central zone

Mediterranean zone

Countries (official abbreviations)

Finland (FI) Sweden (SE) Denmark (DK)

France (FR) Italy (IT)

Latvia (LV) Lithuania (LT) Estonia (EE) Poland (PL) Slovakia (SK) Czech Republic (CZ)

Hungary (HU) Bulgaria (BG) Romania (RO)

Portugal (PT) Spain (ES) Malta (MT) Slovenia (SI) Greece (EL) Cyprus (CY)

Total population (* 1000) Median latitude (south-north) of the countries Median longitude (west-east) of the countries Index of GDP in PPS per capita1) Mean meat price in V per kg2)

20,517 62.0

Ireland (IE) United Kingdom (UK) Netherlands (NL) Belgium (BE) Luxembourg (LU) Germany (DE) Austria (AT) 187,168 51.0

124,822 44.2

60,696 60.0

39,009 46.0

72,109 39.1

15.0

5.4

7.2

22.0

25.0

14.4

123.8 V8.32

119.6 V6.75

104.4 V7.40

70.7 V3.91

55.2 V3.66

85.7 V5.42

504,321

100.0 V6.20

a

) Source Eurostat. The volume index of GDP per capita in Purchasing Power Standards (PPS) is expressed in relation to the European Union average set to equal 100 (in 2012). PPS is a common currency that eliminates the differences in price levels between countries allowing meaningful volume comparisons of GDP between countries. 2 ) Source GfK EU3C (2012, p. 263). The mean meat prices are based on price data collected in the course of a mystery shopping exercise (averaged over four meat categories and several meat types) in each country.

Table 2 Main sources of daily protein supply (g/person/day) in 6 European zones (2012). Sources

Plant-based protein Cereals (excluding beer) Vegetables Potato Pulses Stimulants Fruits Tree nuts Animal-based protein Milk (including cheese) Pork Poultry Fish Beef and veal Eggs Offal Mutton, goat Totals Totala plant protein Totala meat protein Totala animal protein Totala protein

High-income zones

Medium-to-low income zones

Total

Northern zone

Western Central zone

Mediterranean zone

North-Eastern zone

Eastern Central zone

Mediterranean zone

27.9 3.0 2.4 0.9 2.4 1.3 0.8

26.6 3.2 3.3 1.1 1.5 1.4 0.9

32.3 4.1 1.8 2.1 1.4 1.7 0.8

32.4 3.0 4.3 1.2 0.7 0.8 0.4

35.1 3.7 3.1 1.5 0.7 0.9 0.2

25.9 4.4 2.6 2.9 1.3 1.4 1.0

29.4 3.6 2.9 1.6 1.3 1.2 0.8

26.2 10.1 7.6 8.7 9.0 4.0 1.2 0.3

22.6 11.3 9.4 5.3 5.8 3.7 0.8 0.7

20.0 9.7 8.5 7.9 9.1 3.8 2.2 0.7

16.5 14.0 9.0 5.2 1.5 2.8 1.5 0.0

19.2 8.1 8.2 1.7 1.9 3.6 1.6 0.8

16.6 11.5 11.2 10.8 5.4 3.7 2.2 1.4

20.2 11.0 9.2 6.5 5.8 3.6 1.5 0.7

38.9 27.7 68.9 107.8

41.5 29.0 61.8 103.3

45.6 30.2 64.5 110.1

45.2 25.2 51.8 97.0

47.6 19.5 45.9 93.4

41.7 30.4 63.9 105.5

43.4 28.2 60.5 103.9

a Including the small contributions of ‘other sources’. Source: FAO (2017). The data were weighted by population size for each country of the zone.

levels of GDP (r ¼ 0.42). However, the position of Lithuania, which had a high supply of fish and seafood (16.6 g), indicates that other factors also played a role. In five of the six zones, plant-derived foods provided the smallest part of the total; this applied in particular to the highincome Northern zone (38.9 g). The significance of plant-based protein was relatively high in the low-income Eastern Central zone (47.6 g). At the country level, Table 3 reveals that the total amount of plant-based protein was not related to the countries' geographic and economic gradients. However, the detailed items did show considerable differences. Of all plant sources, cereals (mostly wheat) were the largest supplier of plant protein in all the six zones, but the supply of protein from cereals was higher in countries with a lower level of GDP (r ¼ 0.46). At the level of

zones and at the country level, the North was associated with a higher supply of potato protein (r ¼ 0.51), but a lower supply of vegetable protein (r ¼ 0.45) and protein from pulses (r ¼ 0.55). The East was associated with a lower supply of protein from fruits (r ¼ 0.45) and vegetable protein (r ¼ 0.40), although there were some exceptions, such as Greece (with a relatively high 6.5 g). Finally, the level of GDP was associated with a higher supply of protein from fruits (r ¼ 0.59), nuts (r ¼ 0.58) and stimulants (r ¼ 0.48). Table 2 shows that the total amount of animal-based protein varied more between the zones than that of plant-based protein. In four zones, the supply of animal protein was higher than 60 g, but it was much less in the low-income Northeastern and Eastern Central zones (51.8 and 45.9 g, respectively). In addition, Table 3 displays

34

J. de Boer, H. Aiking / Appetite 121 (2018) 29e40

Table 3 Country ranking by sources of daily protein supply (g/person/day) and correlations of the rankings with spatial gradients and GDP. Sources

Plant-based protein Cereals (excluding beer) Vegetables Potato Pulses Stimulants Fruits Tree nuts Animal-based protein Milk (including cheese) Pork Poultry Fish Beef and veal Eggs Offal Mutton, goat Totals Totala plant protein Totala meat protein Totala animal protein Totala protein

Ranking of the countries

Spearman correlations

Lowest rank (country)

Highest rank (country)

South-north latitude

West-east longitude

GDP

19.7 (CY) 1.6(BG) 0.8 (CY) 0.0 (LV) 0.5 (NL) 0.6 (SI) 0.0 (HU)

40.1 (RO) 7.1 (MT) 5.3 (LV) 3.1 (ES) 5.6 (LU) 2.1 (IT) 1.4 (EL)

0.00 0.45 0.51 0.55 0.20 0.13 0.10

0.28 0.40 0.00 0.06 0.15 0.45 0.29

0.46 0.13 0.20 0.34 0.48 0.59 0.58

9.5 6.7 4.4 1.4 1.1 2.3 0.2 0.0

30.9 (FI) 16.1 (PL) 12.9 (UK) 16.6 (LT) 12.1 (LU) 5.0 (DK) 3.6 (LV) 4.8 (EL)

0.50 0.17 0.21 0.11 0.07 0.27 0.08 0.48

0.03 0.21 0.40 0.34 0.55 0.22 0.11 0.31

0.45 0.26 0.04 0.36 0.76 0.37 0.28 0.20

55.0 (RO) 35.5 (NL) 76.3 (NL) 124.2 (LT)

0.10 0.09 0.26 0.08

0.07 0.72 0.50 0.36

0.27 0.66 0.63 0.42

(SK) (EL) (SI) (HU) (PL) (PL) (NL) (PL)

32.6 15.6 32.1 69.1

(CY) (SK) (SI) (SI)

a Including the small contributions of “other sources”. Source: (FAO, 2017).

that the supply of animal protein, ranging from 32.1 g (Slovenia) to 76.3 g (the Netherlands), decreased along the west-east gradient (r ¼ 0.50) and increased with higher levels of GDP (r ¼ 0.63). Similar results were found regarding total meat protein supply, which strongly decreased along the west-east gradient (r ¼ 0.72) and increased with higher levels of GDP (r ¼ 0.66), with the highest supply in the Netherlands (35.5 g). At the level of countries, the south-north gradient was not related to the total amounts of animal-based protein and meat protein; Table 2 shows that both were relatively high in the two Mediterranean zones. The detailed items disclosed some additional differences. Directly behind the total category of meat protein, milk (including cheese) was the second most important supplier of animal proteins in all six zones. At the level of countries, the supply of milk protein, ranging between 9.5 g (Slovakia) and 30.9 g (Finland), increased from south to north (r ¼ 0.50) and from lower to higher levels of GDP (r ¼ 0.45). The supply of fish protein varied between 1.4 g in land-locked Hungary and 16.6 g in coastal Lithuania, but was not much related to the two gradients or the level of GDP. The different varieties of meat showed a mixed pattern of relative importance. Table 2 displays that pork was the largest source of meat protein in five zones, with a peak in the Northeastern zone, but that the differences with poultry were often small, especially in the two Mediterranean zones; beef was relatively important in four zones. The importance of poultry decreased along the west-east gradient (r ¼ 0.40) and the same applied to the importance of beef (r ¼ 0.55). In contrast, the importance of beef strongly increased with the level of GDP (r ¼ 0.76) and a higher average meat price level (r ¼ 0.81). The only considerable association of the southnorth gradient with animal-based protein was for the supply of mutton and goat meat, which decreased from south to north (r ¼ 0.46), ranging from 4.8 g (Greece) to less than 0.1 g (Poland). Overall, these results show a strong influence of location and GDP on protein diets, with mainly south-north differences in sources of plant protein (and milk supply) and west-east (plus GDP) differences in sources of meat protein.

3.2. Relations between the regional and individual level data (arrows 2 and 3 in Fig. 1) Research question 2 examines how the reported frequencies of meat consumption and the responses to the PPOs were related to the south-north and west-east gradients as well as the GDP ranking of the country. Following the previous section, the survey responses were averaged in each of the six zones (see Table 4) and also in each country in order to facilitate comparisons between the countries (see Table 5). The question on the frequency of meat (beef, pork, poultry, etc.) consumption resulted in very different answers. Table 4 shows the percentages that reported to eat meat four or more days per week. This percentage was relatively high in the Northern zone (57%) and low in the medium-to-low-income Mediterranean zone (24%). At the country level, Table 5 reveals that frequency of meat consumption (the full answer scale) was moderately strongly associated with the south-north gradient (r ¼ 0.61), but not related to the west-east gradient or the level of GDP. As mentioned above, the south-north gradient was not related to the total amounts of meat protein supply. Hence, it should be noted that the rank correlation between reported frequency of meat consumption and meat protein supply was r ¼ 0.05. Instead, the reported frequency of meat consumption was associated with a higher supply of potato protein (r ¼ 0.45). In this context, the position of the two Mediterranean zones is important; the reported frequency of meat consumption is low, the supply of potato protein is low, but the level of meat protein supply is relatively high. Regarding the PPOs, Table 4 shows that replacing most of the meat by vegetables was responded to more positively in the two Mediterranean zones (55% and 57%) than in the Northern zone (38%). At the country level, Table 5 displays that the percentage of positive responses, ranging between 70% (Romania) and 29% (the Netherlands), strongly decreased along the south-north gradient (r ¼ 0.76) and to a lesser degree, with higher levels of GDP (r ¼ 0.41). The option of eating less meat but of a certified origin was the most popular one, with on average positive responses from

J. de Boer, H. Aiking / Appetite 121 (2018) 29e40

35

Table 4 Responses to the survey questions in 6 European zones (2012). Questions

How many times a week do you eat meat? (% 4 times or more) Willingness to replace meat (% yes) Willingness to eat less but better meat (% yes) Willingness to replace beef or pork (% yes)

High-income zones

Medium-to-low income zones

Total

Northern zone (n ¼ 1005)

Western Central zone (n ¼ 8823)

Mediterranean zone (n ¼ 5943)

North-Eastern zone (n ¼ 3064)

Eastern Central zone (n ¼ 1982)

Mediterranean zone (n ¼ 3523)

57%

37%

33%

39%

30%

24%

35%

38% 77% 69%

42% 78% 67%

55% 86% 78%

54% 83% 79%

63% 86% 81%

57% 86% 78%

51% 82% 74%

Table 5 Country ranking by responses to the survey questions and correlations of the rankings with spatial gradients and GDP. Questions

Frequency of meat consumption per week (% 4 times or more) Frequency of meat consumption per week (full rating scale) Willingness to replace meat (% yes) Willingness to eat less but better meat (% yes) Willingness to replace beef or pork (% yes)

Ranking of the countries

Spearman correlations

Lowest rank (country)

Highest rank (country)

South-north latitude

West-east longitude

GDP

12% (EL)

81% (DK)

29% (NL) 51% (EE) 63% (DE)

70% (RO) 91% (PT) 89% (PT)

0.61 0.76 0.60 0.53

0.22 0.22 0.04 0.06

0.22 0.41 0.20 0.42

82% of the participants (Table 4). In this case, the percentage of positive responses moderately decreased along the south-north gradient (r ¼ 0.60) but was almost unrelated to levels of GDP (r ¼ 0.20). The option of replacing beef or pork with poultry or fish had an in between position in each zone, with on average positive responses from 74% of the participants (Table 4). Again, the percentage of positive responses decreased along the south-north gradient (r ¼ 0.53) and with higher levels of GDP (r ¼ 0.42). In sum, the south-north gradient was related to the responses to all the options, although in different degrees, the level of GDP was related to the responses to two options, but the west-east gradient was not related to any of the responses (Table 5). In addition, it is worth noting that, at the country level, positive responses to the option of replacing most of the meat by vegetables increased with the supply of vegetable protein (r ¼ 0.43) and the supply of pulses (r ¼ 0.41), but that they were almost unrelated to the total amount of meat or animal protein supply. Positive responses to the option of replacing beef or pork with poultry or fish weakly increased with the supply of vegetable protein (r ¼ 0.34) and considerably increased with the supply of pulses (r ¼ 0.53), but were not related to the supply of either beef, pork, poultry or fish. In order to provide some reference values, it is useful to consider how the level of self-reported pro-environmental actions was related to the south-north and west-east gradients as well as the GDP ranking of the country. The average level of self-reported proenvironmental actions decreased somewhat along the south-north gradient (r ¼ 0.38), decreased more strongly along the west-east gradient (r ¼ 0.51) and was unrelated to the level of GDP (r ¼ 0.17). In sum, the pattern of these correlations was different from the pattern found for the PPOs; factors related to the southnorth gradient made a difference that was specific for the latter.

3.3. Relations between the individual level data (arrows 4 and 5 in Fig. 1) The third research question focuses on the relations between the responses to the PPOs and the variables gender, frequency of meat consumption, and self-reported pro-environmental actions,

in the context of regional differences. The correlations between the responses to the three options in the total weighted sample of participants who reported to eat meat (n ¼ 23,646) were r ¼ 0.29, r ¼ 0.31 and r ¼ 0.22, which allows for treating the options as separate items. The correlations in Table 6 show that women in all zones reacted more often positively to the PPOs than men did. The correlations with gender, which ranged from r ¼ 0.22 (p < 0.001) to r ¼ 0.07 (p < 0.05), displayed two general tendencies; the correlations were higher in the Northern zone, Western Central zone, and NorthEastern zone, and they were higher for the option of replacing most of the meat by vegetables than for the others. In contrast, a higher frequency of meat consumption was associated with less positive responses to the options for pro-environmental protein consumption. In this case, the significant correlations ranged between r ¼ 0.35 (p < 0.001) and r ¼ 0.08 (p < 0.01), with one nonsignificant correlation (r ¼ 0.03). Again, the correlations were higher for the option of replacing most of the meat by vegetables than for the others, in particular in the Western Central zone. In all zones, the correlations of either gender or frequency of meat consumption with the responses to the option of replacing most of the meat by vegetables were somewhat higher than the correlations between gender and frequency of meat consumption (see Table 6, men reported slightly higher frequencies of meat consumption). Levels of self-reported pro-environmental actions were associated with more positive responses to the PPOs in all zones, although the correlations were small. The correlations ranged between r ¼ 0.26 (p < 0.001) and r ¼ 0.06 (p < 0.01) and were somewhat lower in the low-income zones. In order to provide some reference values, it is useful to consider the correlations between the eight items of the self-reported pro-environmental actions measure; the correlations between each of the eight items and the sum of the seven other items were often higher than 0.30 and ranged between r ¼ 0.23 (recycling) and r ¼ 0.36 (reduced disposables). In comparison with these values, those in Table 6 often are smaller than desirable for item-total correlations. The smallest correlations were found for the option of replacing beef or pork with poultry or fish. The highest correlation (r ¼ 0.26) was for the

36

J. de Boer, H. Aiking / Appetite 121 (2018) 29e40

Table 6 Correlations of gender, frequency of meat consumption and level of pro-environmental action with the options for pro-environmental protein consumption in each zone. Variables

High-income zones Northern zone (n ¼ 1005)

Correlations of gender (man ¼ 1, woman ¼ 2) with Willingness to replace meat (% yes) 0.22*** Willingness to eat less but better meat (% 0.13*** yes) Willingness to replace beef or pork (% 0.07* yes) Frequency of meat consumption -0.16*** Level of pro-environmental action 0.18*** Correlations of frequency of meat consumption with Willingness to replace meat (% yes) -0.25*** Willingness to eat less but better meat (% -0.11** yes) Willingness to replace beef or pork (% -0.12** yes) Level of pro-environmental action -0.19*** Correlations of level of pro-environmental action with Willingness to replace meat (% yes) 0.22*** Willingness to eat less but better meat (% 0.19*** yes) Willingness to replace beef or pork (% 0.15*** yes) *

p < 0.05;

**

p < 0.01;

***

Medium-to-low income zones

Western Central zone (n ¼ 8823)

Mediterranean zone (n ¼ 5943)

North-Eastern zone (n ¼ 3064)

Eastern Central zone (n ¼ 1982)

Mediterranean zone (n ¼ 3523)

0.22*** 0.12***

0.15*** 0.05***

0.22*** 0.17***

0.18*** 0.08***

0.15*** 0.09***

0.09***

0.10***

0.10***

0.10***

0.13***

-0.16*** 0.15***

-0.06*** 0.07***

-0.17*** 0.12***

-0.13*** 0.03

-0.09*** 0.05**

-0.35*** -0.19***

-0.26*** -0.08***

-0.28*** -0.14***

-0.22*** -0.08**

-0.23*** -0.03

-0.08***

-0.09***

-0.16***

-0.11***

-0.13***

-0.17***

-0.04**

-0.11***

-0.07*

-0.14***

0.21*** 0.26***

0.20*** 0.19***

0.13*** 0.14***

0.13*** 0.12***

0.20*** 0.15***

0.14***

0.12***

0.06**

0.11***

0.16***

p < 0.001.

option “eat less meat but of certified origin” in the Western Central zone. Overall, however, the correlations indicate that the responses to the three options were only weakly affected by willingness to do something positive for the environment. 4. Discussion This paper aimed to clarify the prospects for the adoption of PPOs by analyzing how consumer responses to some key options might be shaped by cultural, culinary and economic factors at regional and individual level, including regional differences in the importance of plant-based and animal-based dietary protein sources, as well as regional and individual level differences in reported frequencies of meat consumption and self-reported proenvironmental actions. The results show a strong influence of location and GDP on protein availability, with south-north trends in sources of plant protein (but not animal protein except for milk and mutton meat) and west-east (plus GDP) trends in sources and amounts of animal and meat protein. Along the south-north gradient, there were also trends in consumers' reported frequency of meat consumption (which increased) and their responses to the PPOs (which became less positive). The pattern of decreasing positive responses to the PPOs along the south-north gradient and with higher levels of GDP was different from the pattern of the self-reported pro-environmental actions (i.e. the benchmark), which means that it was specific for the PPOs. Overall, the results indicate that spatial variations in consumer diets and the historical context behind them might have complex impacts on the prospects for a future transition and that it is important for our understanding of these impacts to compare regional and individual level data. In the following, we first elaborate on the role of the south-north and west-east gradients, which requires attention for some specific cultural and culinary topics, such as the format of €sler et al., meals (Douglas, 1972) and the familiarity of dishes (Scho 2012). Then, we focus on differences between the PPOs, also taking the role of gender into account. Finally, we suggest directions for future research and policy-making.

4.1. Elaboration on the spatial gradients The south-north gradient showed a number of intriguing contrasts. For instance, in the South, the reported frequency of meat consumption was lowdalthough the level of meat protein supply was relatively highdand there were more positive responses to the PPOs; in the North, consumers reported a higher frequency of meat consumption and, in parallel, more negative responses to the PPOs. These differences suggest that at the individual level the low-meat Mediterranean diet of the 1960 still existed in the mindsets of many consumers, whereas data at the regional level show otherwise, in agreement with the literature (Leclercq et al., 2009; Moreno et al., 2002; Moreno-Altamirano et al., 2016; Turmo, 2012). One of the potential explanations for this contrast is the continuous impact of cuisines and meal formats, which may contribute to the culinary assimilation of new foods or larger portions without consumers being aware of it. As mentioned before, Turmo (2012, p. 128) notes in her chapter about changes in the Mediterranean diet that traditional stews and soups may in recent years have been incorporating vegetables that were unknown until recently, as well as abundant animal protein that was lacking in the 1960s diet. Higher amounts of food waste, in relation to higher levels of disposable incomes, may also be a factor (see Monier et al., 2010, p. 56, for an overview of estimated household food waste per country). Importantly, regional level data about the past transition of the Mediterranean diet show that the strong increase in meat protein was often not accompanied by a corresponding decrease in plant protein (de Boer et al., 2006). Apparently, consumers did not made a deliberate trade-off between plant and animal sources of protein, but just incorporated more meat, thinking they were eating “the same as always” (Turmo, 2012). The role of meal formats is also important for the interpretation of the results in the North. Consumers in the northern and western parts of Europe often have a hot meal with a staple, typically potatoes, a vegetable and a protein component, typically meat €sler et al., 2012). This (Douglas, 1972; Meiselman, 2009; Scho format, which renders the meal recognizable and familiar, may

J. de Boer, H. Aiking / Appetite 121 (2018) 29e40

contribute to understandings of meat as the main part of the dish, which is not easy to replace. The typical meal format may indirectly explain that, at the regional level, the average reported frequency of meat consumption was associated with a higher supply of potato protein. In the North and West, the reported frequency of meat consumption was high, just as the level of meat protein supply. Moreover, in contrast to the South, the responses to the PPOs were less positive, especially where it concerned replacing meat with vegetables. This difference in responses may be partly explained by the fact that those who are well aware of their frequent meat consumption will be less inclined to give it up (de Boer et al., 2016). Another important factor is the traditional familiarity with dishes based on plant-based protein, such as vegetables and pulses, among consumers in the Mediterranean zones (Montanari, 1994), which agrees with the higher supply of these protein sources. This contrasts with the lack of familiarity with plant-based diets as selfevident alternatives to meat diets in countries such as the United Kingdom, Finland and the Netherlands (Ensaff et al., 2015; Jallinoja, €sler et al., 2012). Focusing on food Niva, & Latvala, 2016; Scho practices, the study in the Netherlands indicated that consumers who reported regularly eating hot meals without meat mostly replaced the meat by either other sources of animal protein, such as cheese, or by commercial meat substitutes, which enabled them to € sler maintain their familiar meal patterns and familiar dishes (Scho et al., 2012). In contrast, traditional Mediterranean meat-free dishes, such as Moroccan couscous with chick peas and vegetable, are complete dishes that are based on a broad variety of mainly plant proteins and not on the idea of deliberate meat substitution. Hence, consumers from the North and the South may respond differently to PPOs, depending on their focus on either the idea of having to replace meat or the idea of a plant-based dish itself. Differences across the west-east gradient were dominated by economic and market factors. Higher average incomes at the country level were correlated with higher average meat price levels and higher meat supplies, in particular of beef. The Netherlands, a high-income country where the average meat price level was relatively low (GfK EU3C, 2012), had the highest level of meat protein supply. The higher meat protein supplies in the Western zones indicate that a ‘reversed’ diet transition is especially necessary here. The current levels of protein supply in the 15 early EU member countries, which largely belong to the Western zones, can be compared to the results of an earlier study (de Boer et al., 2006). This comparison reveals that the total level of protein supply in EU15 has slightly decreased from 108.6 g in 1999 to 106.7 g in 2012 (1.9 g), which parallels a slight decrease in the supply of meat protein from 30.4 g to 28.9 g (1.5 g). Although these aggregated figures mask shifts between protein categories, they seem relatively stable over time, despite the economic crises in the early 2000s (von Braun, Algieri, & Kalkuhl, 2014). An economic recovery might lead to an increasing supply of animal protein, especially in the North-Eastern and Eastern Central zones, which had the lowest levels. However, recent worldwide studies report that, at a country level, higher incomes do not uniformly lead to increased meat consumption (Sans & Combris, 2015; Vranken et al., 2014). Based on model calculations, these studies provide some evidence that from a certain point onwards higher levels of average income lead to lower levels of average meat consumption. Although more evidence on this topic is needed, it is important, as Scholliers (2007) notes, that meat, which was highly valued for a long time and functioned as a clear marker of social boundaries (e.g. the traditional British roast dinner), seems to have lost this symbolic significance in several developed countries.

37

4.2. Differences between the PPOs Differences between the responses to the PPOs may be interpreted in terms of balances between, on the one hand, their perceived culinary consequences and, on the other hand, their recognized and appreciated environmental benefits. In each zone, changes within the animal-based protein category received more positive responses than a diet shift from animal-based to plantbased protein. Moreover, at the level of consumers within zones, replacing meat with vegetables was the option that had the strongest correlations with the variables frequency of meat consumption and gender. The three variables might be interrelated because, in particular, traditional types of men do not just tend to report higher frequencies of meat consumption, but also prefer€ sler et al., ences for larger meat portion sizes than women (see Scho 2015). From such a “quantity” perspective, high meat-eaters might have focused primarily on the culinary consequences of the meat substitution issue, perhaps partly responding less positively to this option because they were afraid of not having enough to eat. The two other options showed smaller correlations with frequency of meat consumption and gender in all zones, which indicates that less focus on the meat substitution issue may help to engage more consumers. Also in each zone, positive responses to the PPOs were, to some degree, associated with a higher level of self-reported pro-environmental actions. However, the weak strength of these correlations suggests that a focus on the environmental aspects of food (apart from its packaging) is not self-evident to many consumers. That is, for consumers who are buying or consuming food protecting the environment will not be their focal goal (de Boer et al., 2016). Despite that, however, they may have particular principles guiding their food choices, such as “buy quality if it's worth it”, which could reflect culinary- and environmentally-inspired background goals. This may explain that the option “eat less meat but of certified origin” appeared relatively popular and showed somewhat stronger correlations with the level of self-reported proenvironmental actions. As the certified origin of the meat was unspecified, consumers may have interpreted it in their own ways, for instance in terms of country of origin certification, which is the most familiar certificate on the European meat market (GfK EU3C, 2012). It has been shown that consumers often attribute environmental (and other) benefits to locally or nationally produced products (Lazzarini, Visschers, & Siegrist, 2017; de Boer et al., 2016). Hence, it may have been relatively easy for them to recognize and appreciate the environmental benefits of this option. In contrast, the option of replacing beef or pork with poultry or fish “for environmental reasons” may have been the least familiar one for many consumers, as it showed the weakest correlations with selfreported pro-environmental actions. Indeed, the literature suggests that a preference for white meat over red meat is normally associated with pickiness about meat and taste-related reasons, but not with environmental considerations (de Boer & Aiking, 2011). 4.3. Future directions for research and policy-making There are no reference points for the overall mean of 103.9 g dietary protein supply; the dietary reference values for protein intake provided by the EFSA Panel (2012, p. 32) are 62 (g/d) for European men (reference body weight 74.6 kg), and 52 (g/d) for European women (reference body weight 62.1 kg), but it should be emphasized that supply is not the same as intake and that average body weight is country dependent. As differences between protein supply and protein intake are largely caused by waste (beyond the

38

J. de Boer, H. Aiking / Appetite 121 (2018) 29e40

retail level) and protein intakes of adults in Europe are often above the reference intake (according to the EFSA Panel), achieving a ‘reversed’ diet transition is indeed a matter of ‘using less’, and ‘doing things differently’. A variety of contextual and individual factors can be used as a starting point for improving the prospects of pro-environmental protein consumption, taking into account that there is no single approach for this problem and that it is important to develop approaches in which culinary and environmental aspects can complement each other in different combinations that may bridge a dietary transition (de Boer & Aiking, 2017). Our work suggests a number of important directions for research and policy-making. One of the main strategic questions change agents face is deciding to focus on the meat substitution issue or on the benefits of the meat-free dish itself. The former strategy may generate negative responses among consumers who see themselves as frequent meat eaters (e.g. in the North and West), but it also offers the opportunity to develop meat substitutes that are tastier, less expensive and environmentally more effective than the current ones (Elzerman et al., 2011; Hoek et al., 2011). Less focus on the meat substitution issue, such as promoting changes within the animal-based protein category from red meat to white meat, may help to engage more consumers, but it will reduce the environmental effectiveness of the change strategy (Hoolohan, BernersLee, McKinstry-West, & Hewitt, 2013; de Boer & Aiking, 2011). Focusing on the benefits of the meat-free dish may be easier among consumers who are already familiar with the quality of plant-based dishes (e.g. in the South) or who want to become more familiar with ethnically-inspired meat-free dishes (e.g. adventurous minorities in the North and West). Findings on meat replacement in the Netherlands showed that consumer preferences for dishes with conventional meat substitutes were clearly different from preferences for authentic, ethnically-inspired dishes, such as Moroccan couscous with chick peas and vegetable, that happen to be meat€sler free, but are more than just a meat substitution dish (Scho et al., 2012). This finding may be linked to the broadening appeal of ‘ethnic’ food and restaurants since the 1970s (Scholliers, 2007). A related strategic question is whether the environmental benefits of PPOs will by themselves be convincing enough to stimulate and maintain effective diet changes. In agreement with €sler et al., 2012; de the literature (Hartmann & Siegrist, 2017; Scho Boer & Aiking, 2011), the present results suggest that a focus on the environmental aspects of food (apart from its packaging) is not self-evident to many consumers, although some of them might have environmentally-inspired background goals and show some inclination to increase the share of plant-based protein in their meal. As the topic of pro-environmental protein consumption is relatively new, there may be a lack of awareness and elementary knowledge of the options, as well as a lack of willingness to adopt the most effective options. A recent multi-country study on consumer perception of options to mitigate climate change showed that the outstanding effectiveness of the ‘eating less meat’ option (in comparison with buying either local or organic products) was recognized by merely 6e12% of the consumers (de Boer et al., 2016). Indeed, it has been argued that one potential solution to this problem is increasing consumer awareness of the environmental impact of meat production (Hartmann & Siegrist, 2017). Another potential solution is to consider how culinary and environmental mindsets can be combined in ways that create synergy. Culinary mindsets often show opposing polarities between, for instance, savory and sweet, novelty and tradition, indulgence and health, or convenience and care (Leschziner, 2006; € sler, 2016). These Scholliers, 2007; Warde, 1997; de Boer & Scho polarities may match in various combinations with environmental background goals, such as using fewer resources or doing things

differently for environmental reasons (Clayton & Myers, 2009). For instance, health and care may be prime candidates for synergy with environmental goals. A recent Swiss study found that in the perception of consumers the healthiness and the environmental friendliness of various protein products go together very well, even where this is unwarranted (Lazzarini, Zimmermann, Visschers, & Siegrist, 2016). A complicating factor is that other polarities may play a role as well. For instance, although pulses are seen as healthy and environmental friendly, they are also considered as oldfashioned sources of flatulence, likely to increase body weight (Schneider, 2002). In addition to emphasizing the benefits of pulses, therefore, it is important to modernize their image and to create savory dishes with convenient and varied healthy pulse products (Jallinoja et al., 2016; Schneider, 2002; Vainio et al., 2016). Instead of focusing on one type of product, it is feasible to develop a broader approach to create synergy between culinary and environmental aspects of food by focusing on new, nutritionally healthy, environmentally friendly, and culturally acceptable, regional diets. Such an approach can be adapted to the eating practices, climate, and agricultural tradition of a particular region or country as an alternative to the promotion of ‘ethnic’ food from elsewhere. Examples are the new Nordic diet (Mithril et al., 2012; Saxe, 2014) and the Low Lands Diet (van Dooren & Aiking, 2016). A regional approach could create multiple synergies, for instance, by involving a range of commercial stakeholders, such as farmers, food processors, retailers, restaurant owners and new kinds of food networks. In turn, the stakeholders may support consumers who find it difficult to adopt the diet and who need help with ‘protein’ literacy. Otherwise, as observed by Micheelsen, Holm, and Jensen (2013) in the context of the new Nordic diet, some consumers who are trying a plant-based dish may simply serve it with meat. Hence, a regional approach is necessary to bring gastronomic, nutritional, and environmental specialists together and to bridge the gap between the specialists and the daily life of consumers. 4.4. Limitations This study has several limitations. One of the main limitations is that the use of FAO supply data at the country level is not the most appropriate means to examine the role of cuisines. That requires a more sophisticated analysis of randomly selected, actual meals in smaller geographical areas. Similarly, it was not possible to go into details concerning the functioning of the European meat market. Another limitation is the dichotomous way in which the responses to the PPOs were measured. This made it important to cautiously interpret the percentage of positive responses (Van Vaerenbergh & Thomas, 2013), in particular because people in the South may have a more expressive communication style than people in the North and West (Hui & Triandis, 1989). To address this potential problem, it was essential to use various types of data and to make comparisons at multiple levels. Overall, a key strength of the study is that regional level data were compared with individual level data and vice-versa, which has the advantage that it provides a broader picture (i.e. the impacts of culinary and economic gradients) and more analytical insight into relevant issues (i.e. the effects of an individual's focus). 5. Conclusions The present work indicates that cultural, culinary and economic factors related to spatial variations in consumer diets might have complex impacts on the prospects for a European transition towards a low meat diet and that it is important for our understanding of these impacts to compare regional and individual level data. Our study shows, on the one hand, the impacts of contextual

J. de Boer, H. Aiking / Appetite 121 (2018) 29e40

factors, and, on the other hand, the complexities of individual behavior. As to the first, there was a strong influence of location and GDP on protein supplies, with mainly south-north trends in sources of plant protein (but not animal protein except for milk and mutton meat) and west-east (plus GDP) trends in sources and amounts of animal and meat protein. As to the second, the results suggest that consumers from the North and the South responded differently to PPOs, presumably depending on their focus on either the meat substitution issue or the quality of the meat-free dish itself. The results also indicate that willingness to do something positive for the environment and gender played a weak but consistent role in the responses. The variety of contextual and individual factors can be used as a starting point for improving the prospects of proenvironmental protein consumption, taking into account that there is no single approach for this problem and that it is important to develop approaches in which culinary and environmental aspects can complement each other in different combinations. A significant point for change agents is that consumers may not make a deliberate trade-off between plant and animal sources of protein, but may just incorporate more meat. Overall, a regional approach is necessary to ensure that the diet changes do not depend solely on individual decisions but become an integral part of social processes. References Aiking, H. (2014). Protein production: Planet, profit, plus people? The American Journal of Clinical Nutrition, 100(suppl), 483Se489S. Apostolidis, C., & McLeay, F. (2016). Should we stop meating like this? Reducing meat consumption through substitution. Food Policy, 65, 74e89. Askegaard, S., & Madsen, T. K. (1998). The local and the global: Exploring traits of homogeneity and heterogeneity in European food cultures. International Business Review, 7, 549e568. de Boer, J., & Aiking, H. (2011). On the merits of plant-based proteins for global food security: Marrying macro and micro perspectives. Ecological Economics, 70(7), 1259e1265. de Boer, J., & Aiking, H. (2017). Pursuing a low meat diet to improve both health and sustainability: How can we use the frames that shape our meals? Ecological Economics, 142, 238e248. de Boer, J., de Witt, A., & Aiking, H. (2016). Help the climate, change your diet: A cross-sectional study on how to involve consumers in a transition to a lowcarbon society. Appetite, 98, 19e27. de Boer, J., Helms, M., & Aiking, H. (2006). Protein consumption and sustainability: Diet diversity in EU-15. Ecological Economics, 59, 267e274. €sler, H. (2016). Food and value motivation: Linking consumer de Boer, J., & Scho affinities to different types of food products. Appetite, 103, 95e104. von Braun, J., Algieri, B., & Kalkuhl, M. (2014). World food system disruptions in the early 2000s: Causes, impacts and cures. World Food Policy, 1(1), 1e22. Central Intelligence Agency. (2013). The world factbook 2013-14. Washington, DC: Central Intelligence Agency. Available at: https://www.cia.gov/library/ publications/the-world-factbook/index.html. (Accessed 12 March 2017). Clayton, S., & Myers, G. (2009). Conservation psychology: Understanding and promoting human care for nature. John Wiley & Sons. Clonan, A., Wilson, P., Swift, J. A., Leibovici, D. G., & Holdsworth, M. (2015). Red and processed meat consumption and purchasing behaviours and attitudes: Impacts for human health, animal welfare and environmental sustainability. Public Health Nutrition, 18(13), 2446e2456. Cordts, A., Nitzko, S., & Spiller, A. (2014). Consumer response to negative information on meat consumption in Germany. International Food and Agribusiness Management Review, 17(A), 83e106. Dernini, S., & Berry, E. M. (2015). Mediterranean diet: From a healthy diet to a sustainable dietary pattern. Frontiers in Nutrition, 2, 15. van Dooren, C., & Aiking, H. (2016). Defining a nutritionally healthy, environmentally friendly, and culturally acceptable Low Lands Diet. The International Journal of Life Cycle Assessment, 21(5), 688e700. Douglas, M. (1972). Deciphering a meal. Daedalus, 101, 61e81. EFSA Panel on Dietetic Products Nutrition and Allergies. (2012). Scientific opinion on dietary reference values for protein. EFSA Journal, 10(2), 2557. Elzerman, J. E., Hoek, A. C., van Boekel, M. A. J. S., & Luning, P. A. (2011). Consumer acceptance and appropriateness of meat substitutes in a meal context. Food Quality and Preference, 22(3), 233e240. Ensaff, H., Coan, S., Sahota, P., Braybrook, D., Akter, H., & McLeod, H. (2015). Adolescents' food choice and the place of plant-based foods. Nutrients, 7(6), 4619e4637. European Commission. (2013a). Attitudes of Europeans towards building the single market for green products (Flash Eurobarometer 367). Brussels: European Commission, Directorate-General for Communication. European Commission. (2013b). Flash Eurobarometer 367 (Attitudes of Europeans

39

towards building the single market for green products). Cologne: GESIS Data Archive. https://doi.org/10.4232/1.11737. ZA5798 Data file Version 1.0.0. FAO. (2017). Food balance sheets. Rome: Food and Agriculture Organization. Available at: http://faostat3.fao.org/download/FB/FBS/E. (Accessed 12 March 2017). Fischler, C. (1999). The 'McDonaldization' of culture. In J.-L. Flandrin, M. Montanari, & A. Sonnenfeld (Eds.), Food: A culinary history from antiquity to the present (Histoire de l'alimentation) (pp. 530e547). New York: Columbia University Press (Original work published in 1996) (C. Botsford et al., Trans). Friel, S., Dangour, A. D., Garnett, T., Lock, K., Chalabi, Z., Roberts, I., … Haines, A. (2009). Public health benefits of strategies to reduce greenhouse-gas emissions: Food and agriculture. Lancet, 374(9706), 2016e2025. GfK EU3C. (2012). Final report - Consumer market study on the functioning of the meat market for consumers in the European Union SANCO/2009/B1/010. Available at: http://ec.europa.eu/consumers/consumer_evidence/market_studies/meat_ market/index_en.htm. (Accessed 26 January 2016). Graça, J., Calheiros, M. M., & Oliveira, A. (2015). Attached to meat? (Un)Willingness and intentions to adopt a more plant-based diet. Appetite, 95, 113e125. Graça, J., Oliveira, A., & Calheiros, M. M. (2015b). Meat, beyond the plate: Datadriven hypotheses for understanding consumer willingness to adopt a more plant-based diet. Appetite, 90(1), 80e90. Grigg, D. (1995a). The nutritional transition in Western Europe. Journal of Historical Geography, 21, 247e261. Grigg, D. (1995b). The pattern of world protein consumption. Geoforum, 26, 1e17. Grigg, D. (1999). Food consumption in the Mediterranean region. Tijdschrift voor Economische en Sociale Geografie, 90, 391e409. Hartmann, C., & Siegrist, M. (2017). Consumer perception and behaviour regarding sustainable protein consumption: A systematic review. Trends in Food Science & Technology, 61, 11e25. Hoek, A. C., Luning, P. A., Weijzen, P., Engels, W., Kok, F. J., & de Graaf, C. (2011). Replacement of meat by meat substitutes. A survey on person- and productrelated factors in consumer acceptance. Appetite, 56(3), 662e673. Hoolohan, C., Berners-Lee, M., McKinstry-West, J., & Hewitt, C. N. (2013). Mitigating the greenhouse gas emissions embodied in food through realistic consumer choices. Energy Policy, 63, 1065e1074. Hui, H. C., & Triandis, H. C. (1989). Effects of culture and response format on extreme response style. Journal of Cross-cultural Psychology, 20, 296e309. Jallinoja, P., Niva, M., & Latvala, T. (2016). Future of sustainable eating? Examining the potential for expanding bean eating in a meat-eating culture. Futures, 83, 4e14. Keyzer, M. A., Merbis, M. D., Pavel, I. F. P. W., & van Wesenbeeck, C. F. A. (2005). Diet shifts towards meat and the effects on cereal use: Can we feed the animals in 2030? Ecological Economics, 55, 187e202. Laestadius, L. I., Neff, R. A., Barry, C. L., & Frattaroli, S. (2014). “We don't tell people what to do”: An examination of the factors influencing NGO decisions to campaign for reduced meat consumption in light of climate change. Global Environmental Change, 29, 32e40. €, J., Pouta, E., Heikkila €, J., Kotro, J., et al. (2012). DiverLatvala, T., Niva, M., M€ akela sifying meat consumption patterns: Consumers' self-reported past behaviour and intentions for change. Meat Science, 92(1), 71e77. Lazzarini, G. A., Visschers, V. H., & Siegrist, M. (2017). Our own country is best: Factors influencing consumers' sustainability perceptions of plant-based foods. Food Quality and Preference, 60, 165e177. Lazzarini, G. A., Zimmermann, J., Visschers, V. H., & Siegrist, M. (2016). Does environmental friendliness equal healthiness? Swiss consumers' perception of protein products. Appetite, 105, 663e673. Lea, E. J., & Worsley, A. (2008). Australian consumers' food-related environmental beliefs and behaviours. Appetite, 50(2e3), 207e214. Leclercq, C., Arcella, D., Piccinelli, R., Sette, S., & Le Donne, C. (2009). The Italian national food consumption survey INRAN-SCAI 2005e06: Main results in terms of food consumption. Public Health Nutrition, 12(12), 2504e2532. Leschziner, V. (2006). Epistemic foundations of cuisine: A socio-cognitive study of the configuration of cuisine in historical perspective. Theory and Society, 35(4), 421e443. Macdiarmid, J. I., Douglas, F., & Campbell, J. (2016). Eating like there's no tomorrow: Public awareness of the environmental impact of food and reluctance to eat less meat as part of a sustainable diet. Appetite, 96, 487e493. Meiselman, H. L. (Ed.). (2009). Meals in science and practice: Interdisciplinary research and business applications. Oxford: Woodhead Publishing Limited, Inc. Mertens, E., van 't Veer, P., Hiddink, G. J., Steijns, J. M., & Kuijsten, A. (2017). Operationalising the health aspects of sustainable diets: A review. Public Health Nutrition, 20(4), 739e757. Micheelsen, A., Holm, L., & Jensen, K. O. D. (2013). Consumer acceptance of the new nordic diet. An exploratory study. Appetite, 70, 14e21. Mithril, C., Dragsted, L. O., Meyer, C., Blauert, E., Holt, M. K., & Astrup, A. (2012). Guidelines for the new Nordic diet. Public Health Nutrition, 15(10), 1941e1947. Monier, V., Mudgal, S., Escalon, V., O'Connor, C., Gibon, T., Anderson, G., et al. (2010). Preparatory study on food waste across EU 27-Final report. Brussels: European Commission (DG ENV) Directorate C-Industry. Montanari, M. (1994). The culture of food (La fame e l'abbondanza: Storia dell'alimentazione in Europa). Oxford: Blackwell (Original work published in 1993) (Carl Ipsen, Trans.). ndez-Montoya, D., Soto-Estrada, G., García-García, J. J., Moreno-Altamirano, L., Herna Silberman, M., Capraro, S., et al. (2016). Changes in Mediterranean dietary patterns in Italy from 1961 to 2011. Mediterranean Journal of Nutrition and Metabolism, 9(3), 171e181.

40

J. de Boer, H. Aiking / Appetite 121 (2018) 29e40

Moreno, L. A., Sarría, A., & Popkin, B. M. (2002). The nutrition transition in Spain: A European Mediterranean country. European Journal of Clinical Nutrition, 56(10), 992e1003. Pohjolainen, P., Tapio, P., Vinnari, M., Jokinen, P., & R€ as€ anen, P. (2016). Consumer consciousness on meat and the environment - exploring differences. Appetite, 101, 37e45. Pohjolainen, P., Vinnari, M., & Jokinen, P. (2015). Consumers' perceived barriers to following a plant-based diet. British Food Journal, 117(3), 1150e1167. Reynolds, C. J., Buckley, J. D., Weinstein, P., & Boland, J. (2014). Are the dietary guidelines for meat, fat, fruit and vegetable consumption appropriate for environmental sustainability? A review of the literature. Nutrients, 6(6), 2251e2265. Rokkan, S. (1999). State formation, nation-building, and mass politics in Europe: The theory of Stein Rokkan: Based on his collected works. Oxford: Clarendon Press. Rozin, P., Hormes, J. M., Faith, M. S., & Wansink, B. (2012). Is meat male? A quantitative multimethod framework to establish metaphoric relationships. Journal of Consumer Research, 39(3), 629e643. Sans, P., & Combris, P. (2015). World meat consumption patterns: An overview of the last fifty years (1961-2011). Meat Science, 109, 106e111. Saxe, H. (2014). The New Nordic Diet is an effective tool in environmental protection: It reduces the associated socioeconomic cost of diets. The American Journal of Clinical Nutrition, 99(5), 1117e1125. Scheuch, E. K. (1989). Theoretical implications of comparative survey-research why the wheel of cross-cultural methodology keeps on being reinvented. International Sociology, 4, 147e167. Schmidhuber, J., & Traill, W. B. (2006). The changing structure of diets in the European Union in relation to healthy eating guidelines. Public Health Nutrition, 9, 584e595. Schneider, A. V. (2002). Overview of the market and consumption of pulses in Europe. British Journal of Nutrition, 88(S3), 243e250. Scholliers, P. (2007). Novelty and tradition. The new landscape for gastronomy. In P. Freedman (Ed.), Food. The history of taste (pp. 332e357). Berkeley, CA: University of California Press. €sler, H., de Boer, J., & Boersema, J. J. (2012). Can we cut out the meat of the dish? Scho Constructing consumer-oriented pathways towards meat substitution. Appetite, 58(1), 39e47. €sler, H., de Boer, J., Boersema, J. J., & Aiking, H. (2015). Meat and masculinity Scho among young Chinese, Turkish and Dutch adults in The Netherlands. Appetite,

89, 152e159. Siegrist, M., Visschers, V. H. M., & Hartmann, C. (2015). Factors influencing changes in sustainability perception of various food behaviors: Results of a longitudinal study. Food Quality and Preference, 46, 33e39. Smil, V. (2000). Feeding the world: A challenge for the twenty-first century. Cambridge, MA: The MIT Press. Swatland, H. J. (2010). Meat products and consumption culture in the West. Meat Science, 86(1), 80e85. Teuteberg, H. J., & Flandrin, J.-L. (1999). The transformation of the European diet. In J.-L. Flandrin, M. Montanari, & A. Sonnenfeld (Eds.), Food: A culinary history from antiquity to the present (Histoire de l'alimentation) (pp. 442e456). New York: Columbia University Press (Original work published in 1996) (C. Botsford et al., Trans). Tilman, D., & Clark, M. (2014). Global diets link environmental sustainability and human health. Nature, 515(7528), 518e522. Tobler, C., Visschers, V. H. M., & Siegrist, M. (2011). Eating green. Consumers' willingness to adopt ecological food consumption behaviors. Appetite, 57(3), 674e682. Turmo, I. G. (2012). The Mediterranean diet: Consumption, cuisine and food habits. In F. Mombiela (Ed.), Mediterra 2012: The Mediterranean diet for sustainable regional development (pp. 115e132). Paris: Presses de Sciences Po. Vainio, A., Niva, M., Jallinoja, P., & Latvala, T. (2016). From beef to beans: Eating motives and the replacement of animal proteins with plant proteins among Finnish consumers. Appetite, 106, 92e100. Van Vaerenbergh, Y., & Thomas, T. D. (2013). Response styles in survey research: A literature review of antecedents, consequences, and remedies. International Journal of Public Opinion Research, 25(2), 195e217. Vanhonacker, F., Van Loo, E. J., Gellynck, X., & Verbeke, W. (2013). Flemish consumer attitudes towards more sustainable food choices. Appetite, 62, 7e16. Vranken, L., Avermaete, T., Petalios, D., & Mathijs, E. (2014). Curbing global meat consumption: Emerging evidence of a second nutrition transition. Environmental Science & Policy, 39, 95e106. Warde, A. (1997). Consumption, food and taste: Culinary antinomies and commodity culture. London, UK: SAGE. Westhoek, H., Lesschen, J. P., Rood, T., Wagner, S., De Marco, A., MurphyBokern, D., … Oenema, O. (2014). Food choices, health and environment: Effects of cutting Europe's meat and dairy intake. Global Environmental Change, 26, 196e205.