Biotechnology is indubitably one of the most innovative scientific-technological areas in which scientists and public and private enterprises have worked in the past decade. Nevertheless, although discoveries in genetics have been widely reported by the mass media and raised hopes for the solution of many of mankind's problems (hunger, disease, pollution of the planet), the goods produced are regarded with great suspicion by Europeans (Gaskell et al. Reference Gaskell, Allum, Bauer, Durant, Allansdottir, Bonfadelli, Boy, Cheveigné, Fjaestad, Gutteling, Hampel, Jelsøe, Correia Jesuino, Kohring, Kronberger, Midden, Hviid Nielsen, Przestalski, Rusanen, Sakellaris, Torgersen, Twardowski and Wagner2000, Reference Gaskell, Thompson, Allum, Bauer and Gaskell2002; Bauer and Gaskell Reference Bauer, Gaskell, Bauer and Gaskell2002; Midden et al. Reference Midden, Boy, Eisiendel, Fjæsrad, Liakopoulos, Miller, Öhman, Wagner, Bauer and Gaskell2002; Nielsen et al. Reference Nielsen, Jelsøe, Öhman, . Bauer and Gaskell2002; Bucchi and Neresini Reference Bucchi and Neresini2002, Reference Bucchi and Neresini2004; Bonny Reference Bonny2003; Gaskell et al. Reference Gaskell, Allum and Stares2003; Loner Reference Loner2006). The data available, most notably those collected by the Eurobarometer work group, have repeatedly evidenced attitudes towards this type of technological innovation (INRA, EB 35.1 of 1991; EB 39.1 of 1993; EB 46.1 of 1996; EB 52.1 of 2000; EB 58.0 of 2002). Numerous surveys have considered, among other things, public policy, media coverage, and public perception. All such analyses have confirmed the distrust of Europeans in biotechnology (Bauer and Gaskell Reference Bauer, Gaskell, Bauer and Gaskell2002). Similar results have been obtained in regard to GM foods (Grunert et al. Reference Grunert, Lähteenmäki, Nielsen, Poulsen, Ueland and Åström2001, Reference Grunert, Bredahl and Scholderer2003; Frewer et al. Reference Frewer, Scholderer and Brendahl2003, Reference Frewer, Kettlitz, Scholderer, Beekman and Berdal2004; Gaskell et al. Reference Gaskell, Allum, Wagner, Kronberger, Torgersen, Hampel and Bardes2004; Poortinga and Pidgeon Reference Poortinga and Pidgeon2005; Scholderer Reference Scholderer2005; Cook et al. Reference Cook, Robbins and Pieri2006). As Scholderer notes, “attitudes towards GM foods appear to be utterly resistant to persuasion” (2005, p. 270).
This study contributes to the debate on the possible introduction of GM foods by investigating how the attitudes of Europeans towards such foods are shaped. The article examines the data of Eurobarometer 58.0 of 2002, paying particular attention to the opinion of respondents who evinced specific attitudes. It is important, in fact, to observe the characteristics of those individuals who differ from the majority of Europeans by being less mistrustful of GM foods. Is suspicion generalized, or are some categories of people more willing to accept the new foods?
Research studies have stressed that the causes of rejection are manifold. The hypothesis that the possession of greater information favours acceptance, and that the matter can be thus explained in terms of information deficit has not been confirmed. This has been shown by various studies reporting that acceptance of the new biosciences by no means increases with the amount of information possessed (Loner Reference Loner2006). Other analyses have emphasised the importance of the type of education, and in particular of interest in scientific subjects. For instance, in a survey of young Finns, Saher et al. (Reference Saher, Lindeman and Kovisto Hursti2006) note that students enrolled on degree courses in science or technology are more favourable towards GM foods.
Some authors stress that people assess biotechnology products by giving great weight to socio-cultural factors. On this view, it may be that the fear of dangers to humanity and the environment, and the scant transparency of decision-making processes concerning innovative technologies, have created a certain alarm in regard to genetics (Gaskell et al. Reference Gaskell, Thompson, Allum, Bauer and Gaskell2002; Bucchi and Neresini Reference Bucchi and Neresini2002, Reference Bucchi and Neresini2004; Grunert et al. Reference Grunert, Bredahl and Scholderer2003; Verbeke et al. Reference Verbeke, Frewer, Scholderer and De Brabander2007). This seems mainly due to the fact that the innovations introduced by genetics, for instance in the agro-food sector, may have unexpected and harmful consequences (Scholderer Reference Scholderer2005).
These fears have been fuelled by the failure of the European (and national) public institutions to manage the crises caused by technology-connected risks (Frewer et al. Reference Frewer, Kettlitz, Scholderer, Beekman and Berdal2004; Scholderer Reference Scholderer2005). For instance, Torgersen et al. (Reference Torgersen, Hampel, von Bergmann-Winberg, Bridgman, Durant, Eisiedel, Fjæstad, Gaskell, Grabner, Hieber, Jelsøe, Lassen, Marouda-Chatjoulis, Nielsen, Rusanen, Sakellaris, Seifert, Smink, Twardowski, Wambu Kamara, Bauer and Gaskell2002) recall the Bovine Spongiform Enchephalitis (BSE) crisis, which although not imputable to biotechnology, notably eroded public trust in experts. In this regard, Bonny (Reference Bonny2003, p. 58-59) notes that genetically modified organisms have assumed the role of “scapegoats” symbolizing the harmful effects of globalization in the collective imagination.
Moreover, some surveys have shown that the attitudes of Europeans towards GM foods is associated with a general image of technologies, more than with a judgment on the actual properties of such foods (Bredahl Reference Bredahl2001; Midden et al. Reference Midden, Boy, Eisiendel, Fjæsrad, Liakopoulos, Miller, Öhman, Wagner, Bauer and Gaskell2002). It is, in fact, necessary to bear in mind that Europeans at-large have never had direct experience of such products because of the de facto moratorium imposed by the European Union on their importing in application of the so-called one “precautionary principle”Footnote 1.
An interpretation of the effect of the media on public opinion put forward by Bauer (Reference Bauer2005a; Reference Bauer2005b) centres on the differences between “red” and “green” biotechnologies. Bauer's analysis concentrates on the distinction drawn in the 1990s between medical and agro-food biotechnology. According to Bauer, there is an association between media coverage – and therefore the emphasis on biotechnology's possible health benefits or risks – and the medium-period trend in public opinion. Green biotechnology, in fact, is largely described by the media in terms of its harmful aspects, and this has fostered the public's negative attitude towards it. Media reports on red applications of biotechnology have instead extolled its possible positive impact on human health and well-being, thus helping to create a positive image of it. To this can be added that medical applications are perceived as more controllable – because they are restricted to a small section of the population – and as more acceptable because the patient's survival is at stake. This hypothesis has been partially confirmed by other studies (Bonfadelli Reference Bonfadelli2005; Gutteling Reference Gutteling2005), which, however, stress the need to clarify the theoretical framework in which to insert the role of the media and that of associations campaigning against the negative effects of genetics. Gaskell moreover, emphasises that opposition is stronger against GM foods and animal cloning, while the production of seeds and the cloning of human cells for therapeutic purposes instead receive moderate support (Gaskell et al. Reference Gaskell, Allum, Bauer, Durant, Allansdottir, Bonfadelli, Boy, Cheveigné, Fjaestad, Gutteling, Hampel, Jelsøe, Correia Jesuino, Kohring, Kronberger, Midden, Hviid Nielsen, Przestalski, Rusanen, Sakellaris, Torgersen, Twardowski and Wagner2000). This therefore seems to bear out Bauer's thesis.
This study divides into three parts which analyse the attitudes of Europeans towards GM foods and the mechanisms that shape opinions on the issue. The first part considers the opinions of Europeans on transgenic foods. It does so also by constructing an index showing the extent of acceptance (or rejection) of this technological innovation. The second part examines the index of openness to GM foods, paying particular attention to the factors that favour their acceptance. It is, in fact, of interest to determine whether distrust varies according to the characteristics of respondents, or to other attitudes such as fears (and hopes) in regard to the consequences of biotechnologies for the health, trust in the ability of science to improve human well-being, or broader aspects like political orientations or fears for the health of the planet. The third part further deepens the analysis by applying a non-standard technique (Guttman's error analysis). To investigate the phenomenon more thoroughly, this perspective assumes that it is necessary to focus mainly on respondents who judge GM foods from points of view different from – and at times opposed to – those of the majority. It is, in fact, important for sociological search to seek to identify the mechanisms that shape people's opinions. In this specific case, this entails analysing the features and the motivations of those who, contrary to the general pattern, would accept GM foods.
Constructing an index of acceptance of GM foods
The first part of this study examines the propensity of Europeans to purchase GM foods. The analysis considers data from the Eurobarometer 58.0 survey of 2002, which covered the main European countries (Austria, Belgium, Denmark, Finland, France, West and East Germany, Great Britain, Greece, Ireland, Northern Ireland, Italy, Luxembourg, Holland, Portugal, Spain and Sweden), with a total of 16040 casesFootnote 2.
The following questions were used for the analysis:
Q. 14. For each of the following statements, please tell me if you tend to agree or tend to disagree? (Possible answers: “Tend to Agree”, “Tend to Disagree”).
1. I would buy genetically modified food if it contained less fat than ordinary food (LESSFAT)
2. I would buy genetically modified food if it were cheaper than ordinary food (CHEAPER)
3. I would buy genetically modified food if it contained less pesticide residues than ordinary food (NOPESTIC)
4. I would buy genetically modified food if it were grown in a more environmentally friendly way than ordinary food (ENVFRIEND)
5. I would buy genetically modified foods if it tasted better than ordinary food (TASTEBET)
Preliminary examination of the replies to these questions showed that GM foods are disliked by Europeans. Fully two-thirds of respondents would not purchase them if they were cheaper, and around the same proportion would not do so if they contained less fat (CHEAPER = 66.2 %, LESSFAT = 65.2 %, table 1).
Table i Acceptance of GM Foods (Column %, Eurobarometer 58.0 of 2002, n. =16040)
Nor would superior quality persuade the most reluctant respondents (TASTEBET = 56.3 %). The difference between those subjects for and against purchasing GM foods would only decrease if they were produced without pesticides (-7.5 %)Footnote 3.
Mokken Scale Analysis (MSA), a nonparametric probabilistic IRT model, derived from Guttman's cumulative scale analysis (Guttman Reference Guttman, Stouffer, Guttman, Suchman, Lazarsfeld, Star and Clausen1950; Mokken Reference Mokken1971; Giampaglia Reference Giampaglia1990; Molenaar et al. Reference Molenaar, Van Schuur, Sijtsma and Mokken2000; Sijtsma and Molenaar Reference Sijtsma and Molenaar2002; Van Schuur Reference Van Schuur2003; for a parametric counterpart, see Andrich Reference Andrich1988; Embretson and Reise Reference Embretson and Reise2000; Bond and Fox Reference Bond and Fox2001), performed on all the questions, revealed the existence of a “strong” one-dimensional scale (H-scale = 0.76, table 2)Footnote 4. The result of the analysis enabled the items to be ordered according to decreasing difficulties: CHEAPER was the most difficult item and its average score was 26 (table 2, column 2). Since all the items had been dichotomized, this value also indicated the percentage of respondents who replied to the question in the affirmativeFootnote 5. This signifies that only just over one respondent in every four would purchase GM products if they cost less than other foods. There then followed in order: LESSFAT (27), TASTEBET (34), ENVFRIEND (41) and NOPESTIC (42), which was the easiest (i.e. popular) step of the scale.
Table ii Scale analysis of acceptance of GM foods (%; n. =11647; Eurobarometer 58.0 of 2002).
As predictable, the reasons that would induce the majority of respondents to purchase GM foods were the absence of pesticides and environmental friendliness. The positive replies obtained – from well below half of the respondents – on these items nevertheless indicate the unpopularity of GM foods.
Since the values of the coefficients of scalability (H-item, table 2, last column) were very highFootnote 6, and analysis did not show violations of the model's assumptions of monotonicity, the instrument was one-dimensional according to the requirements established by Mokken (Reference Mokken1971)Footnote 7.
On the basis of these results it was therefore possible to construct an additive index representing, by and large, the degree of acceptance of genetically modified foods. Summing the responses to the five items (with values equal to zero if the interviewee agreed [tend to agree] and to one if s/he disagreed [tend to disagree]), each respondent received a score ranging from five, for those who always gave positive replies, to zero for those who, conversely, would not in any circumstances purchase GM foods. The index confirmed the hostility of Europeans to these types of foods. Only 16.4 % of the interviewees, in fact, obtained the highest score and the 8.2 % recorded four, while almost half (48.6 %) achieved the lowest score (fig. 1).
Fig. 1 Index of acceptance of GM foods for all countries (Eurobarometer 58.0 of 2002; %; n. = 11647)
The factors influencing the opinions of Europeans on GM foods
The second part of this study analyses the mechanisms that shape opinions on GM foods. To frame the problem more precisely, it is necessary to maintain control over both the structural and “measurement” levels of the attitude to GM foods. A technique useful for this purpose is structural equation modelling (SEM) and, in particular, Lisrel (Jöreskog and Sörbom Reference Jöreskog and Sörbom1993). These tools make it possible simultaneously to consider covariations among the model's variables of the model and to include possible latent dimensions. For these reasons, the model constructed is called MIMIC, i.e. Multiple Indicator–Multiple Cause. In the specific case reported here, the model estimated the parameters of the relations between the latent dependent variable, consisting of the likelihood of purchasing genetically modified foods, and certain exogenous variables. Openness to GM foods, termed FOOD, was indicated by the following five items:
LESFATT, CHEAPER, TASTEBET, NOPESTIC AND ENVFRIENDFootnote 8.
Initially considered as independent variables were nine demographic and socio-cultural dimensions (fig. 2):
1. The gender of the interviewee (SEX)
2. Age expressed in years (AGE).
3. Education level (EDU), obtained by considering the respondent's age at the time when s/he completed formal education. For greater simplicity, this value was recoded into three levels: “low” (aged up to 15), “medium” (16-19) and “high” (20 or over)Footnote 9.
4. The interviewee's political orientation (POLITIC) on the left/right continuum (1 = Left, 10 = Right).
5. Strong support for biotechnology (BIOPRIV). This dimension concerned agreement with applications of this science in situations involving the use of private genetic information. The contexts considered were particularly delicate and connected with health (for instance, the possibility of undergoing genetic tests to detect illnesses in advance), social security and insurance. Approval of such applications meant having great faith in the possibilities offered by biotechnology. Looking with more favour on the benefits to health and well-being (more than the negative aspects) also indicated a favourable attitude towards this type of innovation. For these reasons, it likely that individuals generally more in support of biotechnology will also be more willing to purchase GM foods. The additive index used to “measure” this attitude consisted of values ranging from 0 (no support) to 6 (maximum support)Footnote 10.
6. Trust in the public and private actors involved in biotechnology research (RESEARCH). Because trust is the main basis of social relationships, it is likely that greater trust in the genetic researchers will be associated with greater openness towards GM foods. This dimension considers the role of certain specific actors: private companies, scientists, and universities. By summing the positive replies. it was possible to obtain an additive index representing the degree of trust in actors involved in biotechnology research. This index could assume values between 0 (trust in none of the research actors) and 3 (trust in all three)Footnote 11.
7. Knowledge of biotechnology (TEST). It was said in the introduction that the relation between knowledge and support for biotechnology is unclear, so that it was also of interest to verify its extent in this analysis. Knowledge was obtained by summing the exact replies given to a series of questions:
“Q. 11. For each of the following statements, please tell me whether you think it is true or false? [True, False]” and, in particular: “1. There are bacteria which live from waste water; 2. Ordinary tomatoes do not contain genes, while genetically modified tomatoes do. 3. The cloning of living things produces genetically identical copies; 4. By eating a genetically modified fruit, a person's genes could also become modified; 5. It is the mother's genes that determine whether a child is a girl (M); 6. Yeast for brewing beer consists of living organisms; 7. It is possible to find out in the first few months of pregnancy whether a child will have Down's Syndrome, Trisomy, Mongolism; 8. Genetically modified animals are always bigger than ordinary ones; 9. More than half of human genes are identical to those of a chimpanzee; 10. It is not possible to transfer animal genes into plants; 11. Criminal tendencies are mainly genetically inherited; 12. Musical abilities are mainly learned”.
Each respondent was given a score between 0 and 12 corresponding to the number of correct answersFootnote 12.
8. Trust in technology in general (TRUSTEC). This dimension was included in the model in order to verify whether, as seemed likely, a favourable view of certain technological innovations (not closely connected with biotechnology) was associated with greater acceptance of GM foods. Trust in technology was as obtained as an additive index, which was constructed by examining opinions on the effects for society of certain innovative technologies [possible answers: “Will Improve” (3); “No Effect” (2); “Will Make Things Worse” (1)]. The list was: computer and information technology; telecommunications; exploration of space, nuclear energy, mobile telephones. The sum of the scores for every question enabled a score between 0 (minimum) and 10 (maximum) to be assigned to each respondentFootnote 13.
9. Worry about various environmental problems (WORRY). GM foods may be ambivalently associated with fear about the future of the environment (Tait Reference Tait2001; Bruce Reference Bruce2002; Cooley Reference Cooley2002; Toke Reference Toke2002). On the one hand, they may diminish if they fuel hope that pesticide-free and environment-friendly foods can be developed. On the other, the introduction of genetics into the agro-food system may have unpredictable long-term consequences. Since such effects may not yet have been determined by scientists, fear of the possible risks may reduce acceptance of GM foods. Another question often raised by environmentalists concerns the legitimacy for humans to manipulate other living beings. This worry was “measured” by means of the following question:
“Q.39. At present, are you very worried, fairly worried, not very worried or not at all worried about the following topics?: 1. Destruction of the ozone layer; 2. Climate change; ; 3. Acid rain; 4. The extinction of animals and plants species; 5. The progressive elimination of tropical rain forests; 6. Hunting and shooting; 7. Natural disasters; 8. Disasters caused by industrial activities; 9. Pollution of tap water; 10. Pollution of the seas and coasts; 11. Pollution of rivers and lakes; 12. Pollution of underground water; 13. Pollution from farming; 14. Damage caused by tourism; 15. The use of genetically modified organisms; 16. The use of pesticides; 17. The use of other chemical products; 18. Air pollution; 19. Nuclear power and radioactive waste; 20. Noise; 21. Urban problems (traffic, public transport, green spaces, etc.); 22. Using up natural resources that cannot be easily replaced; 23. Environmentally friendly consumption habits; 24. Domestic and urban waste management; 25. Industrial waste management”.
Each respondent was given a score between 0 (never worried) and 25 (always worried) which, for greater simplicity, was recoded into 9 levels on the base of the distribution of frequency of the variableFootnote 14.
Fig. 2 The model of acceptance of genetically modified foods (Eurobarometer 58.0 of 2002, n. = 5028)
Since an exploratory analysis (see appendix, table A1) had shown that the attitude towards GM foods was largely unconnected with education, gender, age, and political orientation, these variables were eliminated in order to obtain a more parsimonious solutionFootnote 15.
The new model obtained after such simplification provided a fairly accurate “snapshot” of the situation. Firstly, greater openness to GM foods was associated, as hypothesized, with support for the use of personal genetic information (BIOPRIV = 0.38, fig. 2)Footnote 16. Obviously this result was not surprising, since it denotes cognitive convergence between the representations of the two fields of biotechnology application (red and green). Subjects who saw above all the usefulness for health (or for social security) of biotechnology also more easily accepted its applications in the agro-food industry.
Also trust in the actors actively involved in genetic research was positively associated with acceptance of GM foods (RESEARCH = 0.28). To this should be added that this factor was also connected with BIOPRIV (0.32Footnote 17) and can therefore be seen as an attitude of generalized trust in biotechnology.
Thirdly place, fear of environmental problems was the main obstacle against acceptance of GM foods (WORRY = -0.18). This result indicates that the greater the worry about degradation of the planet, the greater the fear that biotechnology may increase pollution. Although this hypothesis is difficult to demonstrate with the data available, and require further investigation, it may explain why campaigns by movements opposed to the development of these technologies are more successful when they emphasise the possible damage to the health of people and to that of the planet (Tait Reference Tait2001).
Another point concerns the fact that fears acted as an “irrational” component because they were negatively associated with the level of knowledge about the biotechnologyFootnote 18 but positively with the use of private genetic information (BIOPRIV = 0.11). This perhaps unexpected result can be explained by considering that BIOPRIV concerned personal safety (for instance health, tests on unborn babies, help for sterile couples, access to information by doctors). When the health (and often the survival) of people was at stake, the result of the judgment depended mainly on the possibility of saving lives. This highlights the difference in public opinion when biotechnology is applied in the medical or agro-food sectors. In the former case, biotechnology is seen as an additional cause for hope, while in the latter it is seen as a possible threat.
Fourthly, knowledge was moderately associated with acceptance of GM foods (TEST = 0.08). However, as already noted by other studies (Loner Reference Loner2006), the relation is modest. Although this aspect should be investigated more thoroughly, one may argue that, contrary to information deficit theory, the amount of information possessed does not greatly increase trust in biotechnology products.
Finally, support for such innovative technologies as computer technology or mobile telephony was associated, albeit modestly, with acceptance of transgenic foods (TRUSTEC = 0.07). Interestingly, this “measure” was more closely associated with approval of research actors and with support for the use of information in the medical or social fields (BIOPRIV = 0.19; RESEARCH = 0.26Footnote 19), as found by other studies (Loner, Reference Loner2006).
Although the model described above produced a first picture of the phenomenon, it was important to deepen the analysis by also introducing as an independent variable the country of the interviewee. Acceptance, or conversely rejection, of GM foods may in fact change in relation to factors connected with the national context, such as coverage by the media, policies adopted in regard to these new technology products, the culture in relation to food quality, etc.
A Lisrel (MIMIC) Multisample model was used to determine the presence of variations in the national groups: a model of this kind shows whether the impact of the factors are the same, or whether differences emerge among groups.
Since the previous model showed that the main factors associated with genetically modified foods were BIOPRIV, RESEARCH and WORRY, it was of interest to test a more parsimonious model which included only these three variables (fig. 3)Footnote 20.
Fig. 3 The multisample model of acceptance of GM foods
The results confirmed a certain amount of variability among the groups according to the country concerned (table 3). Moreover, trust in biotechnology and in its positive effects on health (BIOPRIV) was very closely associated with acceptance of GM foods in all countries. The estimate of the parameter between BIOPRIV and FOOD was not the highest only in Belgium, Denmark, Western Germany and Finland: in those countries the largest weight was again assumed by trust in research actors (RESEARCH = 0.41, 0.42, 0.41 and 0.46, respectively). Worries about the environment reduced acceptance of GM foods in Denmark, Italy and Austria, where they obtained rather higher values (WORRY = -0.26, -0.20 and 0.20, respectively). This seemingly confirms the results obtained by Gutteling et al. (Reference Gutteling, Olofsson, Fjæstad, Kohring, Goerke, Bauer, Rusanen, Bauer and Gaskell2002)Footnote 21, who found that in Denmark, Austria and Italy the media gave great prominence to articles which emphasise the risks but not the benefits of biotechnology. In regard to Italy, for instance, Mini (Reference Mini2005, p. 10) notes that “no concession is made to GMOs, unattractive, useless, poor-quality hybrids. Risks always seem superior to benefits and superior to legislative guarantees”.
Table iii Lisrel standardized estimates of the parameters of the Mimic Multisample Model of the Acceptance of GM Foods (X2 = 584.85; df = 317; p. = 0.000; RMSEA = 0.049; RMR = 0.04; GFI = 0.99; missing = listwise; WLS estimates)
Note: Significance levels of the parameters: *** = 99%; ** = 95%; * = 90%; standard error of the parameter estimated in brackets
Overall, variations in acceptance of GM foods could be attributed to BIOPRIV and RESEARCH, while the role of fears was important in only a few countries.
Since the national groups were not particularly numerous, it may also be that the different patterns recorded in the countries considered were also due, besides actual lifestyles and habits, to the structural composition of the groupsFootnote 22.
Also to be stressed is the role of trust. This aspect, as hypothesised, was associated with a positive view of the subject to whom it was directed. In confirmation of this, in the model developed, the estimate of the parameter between acceptance of GM foods and trust in the actors involved in genetic research (RESEARCH) was greater than that of the obstacle raised by fear in each national group (table 3).
The attitude towards GM foods was therefore closely connected with public trust in the social actors involved in the development of such products and hopes that they would improve the life (and the health) of citizens. The fear of negative consequences instead acted as a brake; but, as said, this effect was significant in only a few countries. In this regard, it is useful to bear in mind the role performed by the mass media in heightening fears. The main hypothesis on this matter (“media coverage hypothesis”) was formulated some ten years ago by Alan Mazur (Reference Mazur1975, Reference Mazur1981). It states that, regardless of the type of news reported, as media coverage increases, so does public distrust in technology (thereby inducing an evaluation error). Research results do not agree on the impact of the media on conflicts concerning biotechnology: Gutteling (Reference Gutteling2005) and Neresini (Reference Neresini2005), for instance, have not corroborated Mazur's thesis.
Moreover, comparative analyses have emphasised the importance of distinguishing among general evaluations, which are often value-loaded, and attitudes toward specific products, which are evaluated mainly in terms of their usefulness (on Spain see Lujàn and Todt Reference Lujàn and Todt2000).
The causes of the distrust by Europeans in GM foods are therefore manifold, and as Gaskell et al. have written:
In conclusion, no single explanation accounts for the greater resistance to food biotechnology in Europe. Media intensity, knowledge of biotechnology and trust in the regulatory process are implicated and interrelate. Different histories of media coverage and regulation go together with different patterns of public perceptions; and these in turn reflect deeper cultural sensitivities, not only towards food and novel food technologies but also towards agriculture and the environment.
(Gaskell et al. Reference Gaskell, Thompson, Allum, Bauer and Gaskell2002, p. 373-374)Subsequent studies should therefore clarify the effect of media coverage, especially in relation to the fears that may be aroused by the application of scientific innovations in the absence of clear rules on their use, and on who should reap the benefits (economic and for health) of new discoveries.
The analysis reported in the first part of this study has identified the factors associated with the phenomenon under examination. However, it is incomplete, because it is unable to shed full light on the mechanisms involved in acceptance of GM foods. The variance in acceptance of GM foods explained by its association with RESEARCH, BIOPRIV and WORRY ranges, in fact, from 36 % (Italy) to 44 % (Ireland and Portugal), and it is likely that at the moment of evaluation the subjects were considering other factors as well (table 3).
A variation on the theme: those who think differently from the majority
In order to obtain more detailed knowledge about the mechanisms that shape opinions on GM foods, the perspective of analysis was changed by focusing on those subjects who “diverged” from the general pattern. Although, as said, the analysis that follows may appear at odds with the rest of the study, it seeks to combine reflections of a substantial and methodological kind. From the substantial point of view, it is important to take account of the attitudes of those subjects who constituted the “minority” who were in favour of GM foods. These respondents may in fact have represented the vanguard of a new tendency likely to exert great influence on the rest of public opinionFootnote 23. It was therefore important to examine those respondents favourable towards GM foods: did they possess characteristics very similar to the other respondents? If they did not, how did they differ from the others? From the methodological point of view, it was interesting to “reverse” the principles underlying the technique used to construct the index of openness to GM foods – i.e. Mokken's (Reference Mokken1971) scale analysis procedureFootnote 24. To understand the type of analysis which follows, it is useful to remember that the Mokken Scale Analysis is based on analysis of deviations from the perfect cumulative scale devised by Guttman (Reference Guttman, Stouffer, Guttman, Suchman, Lazarsfeld, Star and Clausen1950). According to Mokken, if the scale is sufficiently “strong”, it is possible to consider “deviations” from the “perfect” scale as due to chance (Mokken Reference Mokken1971, pp. 41-48)Footnote 25. Such deviations are called “Guttman errors”. To clarify this concept, it is sufficient to point out that Guttman's model entails the dominance of each (more difficult) item over the easier ones, and no exceptions are admitted once the order of difficulty of the questions has been identified. In this way, in order to establish a subject's position on the continuum (or rather on the straight line along which the subjects and the items are ordered), it is only necessary to know the highest step (i.e. question, item) to which the subject has replied positively. An “error”, or violation of the model, is thus defined according to the order that a subject gives to two items. More precisely, an “error” occurs when the subject gives a positive reply to the more difficult of the two items, not to the easier one (Van Schuur Reference Van Schuur2003)Footnote 26. Although the term “error” has a negative connotation in everyday language, in this context it refers only to the interviewee's choice of a pattern of replies at odds with those of the majority, i.e. with the “perfect” scale.
According to an alternative interpretation, “deviations” are not due to pure chance, but depend on the presence of one or more additional dimensions connected to the phenomenon (see e.g. Bart and Krus, Dalton and MacReady, Shye et al., cited in Van Schuur Reference Van Schuur2003, p. 141). This hypothesis can be verified by examining the attitude of subjects who think differently. As Van Schuur (Reference Van Schuur2003, p. 150) points out: “subjects’ numbers of Guttman errors can be regarded as their values on a new variable”. It is thus possible to use this information for further analysis. In the case of the scale of acceptance of GM foods, this meant analysing the “Guttman errors” in relation to the index of openness to GM foods. The analysis now described sought to identify the presence of systematic patterns in the replies of respondents who did not match the “perfect” model (i.e. the Guttman scale). The analysis concentrated on those subjects who ordered the items differently and examined their attitudes towards GM foods. Specifically, this part of the study seeks to show that violations of the “perfect” model were not random but denoted specific and particular patterns of thought.
In this case, to determine whether these were systematic patterns, it was sufficient to observe the variations from the scale of acceptance of GM foods previously identified. The latter, as described earlier, was composed, in decreasing order of difficulty, of: LESFATT->CHEAPER->TASTEBET->NOPESTIC->ENVFRIEND.
Guttman errors can be verified in two steps. The first consists of identifying the respondents who order the items in a different way, while the second concentrates on these individuals in order to determine their distinctive characteristics.
As regards the distribution of the Guttman errors, more than four-fifths of the interviewees did not commit even one “error” in the ordering. This obviously indicated that the great majority matched the “perfect” scale (80.3 %, table 4). It was therefore sufficient to know the scores obtained by these individuals to establish the replies that they gave to all the questions. Since they perfectly matched the deterministic scale, it was possible to know the replies given to each question. That “errors” were relatively rare is also shown by the fact that, conversely, only seven out of one hundred respondents committed more than two “errors” (table 4)Footnote 27.
Table iv Number of Guttman errors in the whole sample (Eurobarometer 58.0 of 2002; row %; n. = 11643)
Those respondents that matched the perfect scale (no “errors”) on average obtained lower scores (on the index of acceptance of GM foods) than did the other respondents: indeed, six out of ten scored zero (60.3 %, table 5). This pattern once against highlights the great resistance of the majority of respondents against the introduction of GM foods. Moreover, fewer than three out of ten of them exceeded two points (29.8 % summing those who obtained three, four or five points). On examining those who committed at least one “error”, one notes that none of them obtained the maximum score (as predictable, since this was a cumulative scale). Instead, on considering a score equal to or above three – which indicates a percentage of subjects favourable to GM foods well above the average – one finds that in all groups with at least one “error” the score is higher than that of subjects who matched the perfect scale.
Table v Acceptance of GM foods according to the number of Guttman errors (Row %; N. = 11643; Eurobarometer 58.0 of 2002)
In four groups, therefore, over half the respondents would purchase GM foods: these were the groups with 1, 3, 4 and 6 “errors” (last column of Table 5) in which the share of interviewees who reached or exceeded three points was well over 50 %. It should be pointed out, however, that these groups all together made up less than 20 % of respondents.
Although the pattern resulted from the analytical procedure used, and was therefore largely predictable (gives the model's conformity with the requirements established by Mokken), one may wonder whether the “deviations” from the majority did not also derive from the specific views and attitudes of these respondents. This hypothesis can be verified by observing the differences in the average scores of the index of acceptance of GM foods according to the number of Guttman errors (fig. 4).
Fig. 4 Index of acceptance of GM foods. Mean scores of respondents according to the number of Guttman errors(n. = 11643, Eurobarometer 58.0)
This analysis supports the hypothesis that acceptance of GM foods involved a different attitude towards them: for those subjects more in favour of such foods, the order of preference among the five items changed with respect to the majority. Those who committed at least one “error” obtained a score higher by around one point on average than the score of those who did not commit even one errorFootnote 28.
The general pattern can be interpreted by considering that, for some respondents, the order of importance of the items differed from the average pattern. Obviously, this should have occurred with greater frequency for the pairs of items “closest” in the scale, i.e. those with similar mean scores: CHEAPER-LESSFAT and ENVFRIEND-NOPESTICFootnote 29.
In order to determine whether “deviations” from the “perfect” scale concealed one or more alternative attitudes towards GM foods, it is necessary to answer a further question. This concerns the aspects that distinguish the “perception” of GM foods by these respondents. To address the question it was sufficient to examine the distribution of the replies that respondents who committed at least one “error” gave to the individual items (fig. 5 and 6).
Fig. 5 Cheaper (CHEAPER), Fewer fats (LESSFAT) and Quality and better taste (TASTEBET) Positive responses according to the number of Guttman Errors (% of positive responses; n. = 11643; Eurobarometer 58.0 of 2002)
Fig. 6 Environment-friendly (ENVFRIEND) and Fewer pesticides (NOPESTIC). Positive responses according to the number of Guttman errors (% of positive responses; n. = 11643; Eurobarometer 58.0 of 2002)
For those respondents who committed one Guttman error, the most convincing aspect of biotechnology concerned its ability to solve environmental problems: almost nine out of ten would purchase biotechnology products if they were produced in an environment-friendly manner (ENVFRIEND = 86.6 %, fig. 6). More than two-thirds, moreover, would make the same choice if such produces contained fewer pesticides – that is, were less harmful to the planet, but also to health – considering that residues of the fertilizers and pesticides used in agriculture also contain substances dangerous to humans (NOPESTIC = 67.6 %). Because these respondents committed just one “error”, it is likely that in the majority of cases the violation was due to an inversion (in the order of preferences) between ENVFRIEND and NOPESTIC, or between CHEAPER and LESSFAT, which as said, had very similar mean scores (see table 2). This is particularly important because it concerns the second largest group (7.9 % of respondents) and indicates an attitude that links environmentalism with personal health.
The second pattern was exhibited by respondents who committed two “errors”, i. e. 5 % of the interviewees. This group had characteristics similar to the previous category and consisted of subjects who appreciated the absence of pesticides (NOPESTIC = 53.1 %, fig. 6), eco-sustainable farming (ENVFRIEND = 60.3 %), and the good taste of food (TASTEBET = 54.8 %, fig. 5). For these individuals, therefore, ecological reasons mixed with a concern for the quality of food.
Whilst the members of the group with three “errors” also appreciated the absence of pesticides (NOPESTIC = 54.6 %, fig. 6), they placed particular importance on the taste of food, on price and, to a greater extent, on a lower fat content (LESSFAT = 64.3 %, fig. 5). Unlike the previous groups, therefore, the choices of these respondents were determined less by environmental worries than by concerns for personal well-being and the quality of food.
Although the last three groups comprised a very small number of respondents, they warrant close attention because they largely reversed the order of items made by the majority. The preferences of these respondents were dominated by economic aspects, which were indicated by 84.4 % in group four (CHEAPER, fig. 5) and by all (100.0 %) of those who committed five or six “errors”. However, each group had distinctive features. In the one with six “errors”, all respondents (100.0 %) also gave a positive response to the absence of fats (LESSFAT = 100.0 %). Those belonging to group five also gave great importance to better taste (TASTEBET = 67.2 %), while those subjects who committed four “errors” favourably evaluated the absence of fats, taste and, to some extent, environmental qualities. This pattern may have been due to the fact that lower-income earners attach greater importance to price, although the survey data unfortunately cannot be used to verify this hypothesis.
The group with four “errors” had interesting characteristics because its members ordered the items in the exact reverse of the “perfect” scale: for these subjects, as reported earlier, CHEAPER was the most popular item, and NOPESTIC was the most difficult. In other words, price mattered a great deal for these individuals, while the absence of pesticides was of little importance to them. At cognitive level, the “yardstick” which they use to evaluate GM foods was the reverse of that used by the majority. These were probably the most materialistic interviewees, because they were little concerned about health, but very much so about price.
Looking at the general pattern, it is possible to add that the number of positive responses for CHEAPER grow as the Guttman errors increased (from 20.3 % in the group with no “errors’ to 100 % in those with five or six: fig. 5), while it decreases for ENVFRIEND and NOPESTIC (fig. 6). As regards ecological and health aspects (ENVFRIEND and NOPESTIC), the highest number of positive responses are instead made by subjects who committed one “error”, while as “violations” increase, the weight of these factors decreases until it disappears in the groups with five or six violations (fig. 6).
Conclusions
As predictable, a positive judgement on transgenic foods is associated with trust in biotechnologies and in the actors involved in their development. Conversely, worries about the environment constitute a barrier against the introduction of GM foods. However, the analysis has shown that the deterrent represented by fears is strong in only some countries. The effect of the fears can be regarded as an “irrational” component, but, as other studies have argued, it also involves the management of decisions concerning applications of these technologies and the way in which information about them is transmitted to the public (Bucchi et al. Reference Bucchi, Neresini and Pellegrini2003; Bucchi and Neresini Reference Bucchi and Neresini2004; Pellizzoni Reference Pellizzoni2006).
Attitudes towards GM foods vary considerably according to the context. When interviewees were unable to identify a clear advantage (above all for health), the response was often negative. Not coincidentally, the reasons of those opposed to the innovations centred on uncertainty and on the dangers – i.e. fears – deriving from the new technologies. Only when the stakes were higher (that is, when people's lives and survival were at risk) was there a change in the conditions in which the choice was made and greater openness to medical applications. This highlights the usefulness of concentrating analysis on the mechanisms involved in interpretative frames.
The study has shown that both attitudes (openness or closure) were based on processes perceived as rational by the interviewees. According to the analytical perspective adopted, respondents who accepted GM foods could be regarded as “deviant”. As said, this term should be taken, not in the “negative” sense, but relatively to the attitude shared by the great majority of respondents.
The method of analysis assumed that the minority of supporters of GM foods choose according to what they consider to be good reasons for responding positively. These individuals, in fact, found strong justification and support for their judgement in two aspects which, in some respects, were the opposite of each other. On the one hand, they hoped that products containing fewer pesticide residues would create a cleaner world and healthier foods. On the other, their acceptance of GM foods depended on economic factors and on opportunities to spend less. One may presume that these latter individuals were more materialistic, or had lower incomes. In the former case, they may also have been people more worried about, or anyway more sensitive to, the environment. From the point of view of the rational reasons for supporting GM foods, anchorage in economic factors was comparable with that for worries about health. In both situations, the decision was conditional on external conditions over which the respondents had – and perceived that they had – little control (in the latter case, for instance, it may have been a serious illness, in the former economic hardship). For respondents in these circumstances, GM foods may have represented an additional resource. Both when life is at stake, and when poverty is a serious threat, any possible salvation is accepted. These actors perceived GM foods as an additional rationally-founded hope.
For a minority of respondents, therefore, their attitude towards GM foods was connected with factors different from those on which the opinion of the large majority was based. It was therefore useful to integrate “measurement” of the phenomenon – though effective in explaining the reasons underlying the choices of a large part of respondents – by seeking to understand the reasons for and the mechanism shaping the opinions of those who thought differently from the majority.
On the basis of the findings set out above, it is possible to group Europeans into five categories representing the same number of attitudes towards GM foods (table 6).
Table vi Division of Europeans according to their attitudes towards GM foods (Eurobarometer 58.0 data, N. = 11643)
The first group (no Guttman errors) comprises the great majority of respondents (over 80 %). These individuals were distinguished by a closure determined by fear, and by the uncertainties surrounding possible negative effects of GM foods. As Verbekes et al. point out:
in a typical food consumer decision-making process, safety is usually a non-negotiable product attribute. There is evidence to suggest that consumers expect all food to be intrinsically safe and a well-informed and rational consumer would never knowingly purchase or consume unsafe food.
(Verbekes et al. Reference Verbeke, Frewer, Scholderer and De Brabander2007, p. 2)Partial openness is apparent in the remaining categories, but the reasons for acceptance depend on different factors. These motivations to some extent override the fears raised by the potential risks of GM foods. For almost 8 % of Europeans (group with one Guttman error), a valid reason for approving these products is their benefit for the health of the planet or of people. Future research could analyse the characteristics of these individuals in order to determine whether they belong to organizations for defence of the environment. The attitude of environmentalists, in fact, lies at the centre of the cleavage between the hopes and fears raised by scientific discoveries and progress. As various commentators have pointed out, biotechnologies may be both a threat and an opportunity to defend the planet (albeit from different points of view: see Buttel Reference Buttel, Hardy and Baker Segelken1998; Hoban Reference Hoban, Hardy and Segelken1998, Reference Hoban, Serageldin and Persley2003; Roller Reference Roller2001; Loner Reference Loner2006 or, for the opinion of agriculturists: Loop Reference Loop, Hardy and Baker Segelken1998).
In the second group (two “errors”), which comprises around 5 % of respondents, health reasons overlap with those concerning the quality of food. For the remaining categories (three “errors” and more), composed of a small minority of respondents, openness largely depends on the hope of spending less.
According to Scholderer (Reference Scholderer2005), only direct experience of the benefits can break down the great resistance of Europeans to the revolution in dietary habits represented by GM foods. More than on information and education campaigns, changes in their attitudes will depend closely on practical aspects: the quality of products, beneficial consequences for health and the environment, and competitive prices. As Bruce (Reference Bruce2002) notes, the success of biotechnology will depend on attaining a shared public vision able to restore trust in the bio-industries, rules, and governments involved in recent technology-related crises (of which BSE has been the most striking example). Future research should seek to establish whether the categories comprising supporters of GM foods are still a small minority, or whether they reflect a growing attitude among Europeans.
From the methodological point of view, the analysis has demonstrated the importance of considering the presence of other dimensions which may be associated with the phenomenon. Specifically, observation of “errors” in ordering has revealed the reasoning of those respondents who gave priority to materialistic aspects (such as price) over more idealistic ones (such as protecting the environment), or the opposition between those who considered health benefits in terms of the absence of pesticide residues and those who instead considered lower fat content. The results have shown that study of human reality – often fleeting and complex – cannot ignore the opinion of the minority of citizens who think differently. Observation of the characteristics of “deviants” (in the non-negative sense with which the term has been used here) yields more profound knowledge on the mechanisms underlying social phenomena.
Table a1
Lisrel standardized estimates of the parameters of the Mimic Multisample Model of the acceptance of GM foods
Significance levels of the parameters: *** = 99 %; ** = 95 %, * = 90 %, standard error of the parameter estimated in brackets
FIT:
Model 1 : Chi-Square = 17.25 ; df = 14; P = 0.25 ; RMSEA = 0.007 ; RMR = 0.007
Model 2 : Chi-Square = 38.56 ; df = 28; P = 0.09 ; RMSEA = 0.010 ; RMR = 0.001
