Hostname: page-component-7b7567c8c-7pwfn Total loading time: 0 Render date: 2026-11-12T11:03:40.210Z Has data issue: false hasContentIssue false
  • English
  • Français

Context Variations and Pluri-methodological Issues concerning the Expression of a Social Representation: the Example of the Gypsy Community

Published online by Cambridge University Press:  20 November 2014

Anthony Piermattéo,
Grégory Lo Monaco,
Laure Moreau,
Fabien Girandola and
Jean-Louis Tavani
Anthony Piermattéo*
Affiliation:
Aix-Marseille Université (France)
Grégory Lo Monaco
Affiliation:
Aix-Marseille Université (France)
Laure Moreau
Affiliation:
Aix-Marseille Université (France)
Fabien Girandola
Affiliation:
Aix-Marseille Université (France)
Jean-Louis Tavani
Affiliation:
Université Paris Descartes (France)
*
*Correspondence concerning this paper should be addressed to Anthony Piermattéo. Aix Marseille Université, LPS EA 849, 13621, Aix en Provence, France. E-mail: anthony.piermatteo@gmail.com
Rights & Permissions [Opens in a new window]

Abstract

Within the social representations’ field of research, the “mute zone” hypothesis considers that some objects are characterized by counternormative content that people usually do not express in standard conditions of production. Within the framework of this approach, this study aims to explore the variations in the expression about the Gypsy community following the manipulation of different contexts and the issues associated with a pluri-methodological approach of data analysis. Indeed, two methodologies have been combined. The participants were asked to express themselves in public or in private. In addition, the identity of the experimenter was also manipulated as she presented herself as a Gypsy or not. Then, through a set of analyses based on a methodological triangulation approach, we were able to observe a recurrent modulation of the participants’ answers. These analyses highlighted a greater incidence of the expression of counternormative elements when the context of expression was private and especially when the experimenter did not present herself as a Gypsy (p < .01, ηp ² = .06). These results will be discussed in terms of the contribution of the methodologies employed and their comparison within the framework of the study of counternormative content.

Keywords

social representationscounternormative contentcontexts of expressionresearch methods

Information

Type
Research Article
Information
The Spanish Journal of Psychology , Volume 17 , 2014, E85
Copyright
Copyright © Universidad Complutense de Madrid and Colegio Oficial de Psicólogos de Madrid 2014 

Since the initial work conducted by Guimelli and Deschamps (Reference Guimelli and Deschamps2000) on the mute zone of social representations (SR), the question of this social phenomenon has aroused interest and led to subsequent research. These works have focused on two areas of research: theoretical and methodological. This second area will constitute the main focus of this contribution. In this perspective, we will rely on the SR of the Gypsy community. Indeed, apart from the work carried out by Guimelli and Deschamps (Reference Guimelli and Deschamps2000), one can notice that this object has been tackled through several studies. For example, Mamontoff (Reference Mamontoff2008) conducted a study based on the question of the link between the SR of work of the Gypsy community and the practices adopted by this group. Moreover, the same author focused part of her work on the SR of the identity of this specific group, particularly through the question of the link between settlement practices and the transformation of the SR of the Gypsy identity Mamontoff (Reference Mamontoff1996). Otherwise, in line with the works conducted about the relationships between majority and ethnic minority, Perez, Moscovici, and Chulvi (Reference Pérez, Moscovici and Chulvi2007) showed that despite a positive image, the Gypsy minority is the target of an ontologization process. These results allow to consider that part of the SR of the Gypsy community could be composed of negative content. Such content could therefore be censored by individuals and groups, a phenomenon that led Guimelli and Deschamps (Reference Guimelli and Deschamps2000) to hypothesize the existence of a mute zone in the representational field.

The mute zone of SR can be defined as “a specific subset of cognitions, which, whilst being available, would not be expressed by the participants in standard conditions of production” (Guimelli & Deschamps, Reference Guimelli and Deschamps2000, p. 53). Thus, to highlight the fact that under some conditions, individuals may not express all that they think about particular objects, Guimelli and Deschamps (Reference Guimelli and Deschamps2000) suggest manipulating the contexts of expression. To this end, they use a technique called “substitution” which is directly derived from the identification paradigm (Jellison & Green, Reference Jellison and Green1981). This technique consists in varying the context of expression by creating two verbal associative task situations. In the first case, individuals have to express themselves from their own point of view (i.e., as they think), whilst in the second case, they have to express themselves as others would (e.g., as the French in general would do). The results obtained by means of this procedure show that under the substitution instruction context, the expression of counternormative elements is facilitated. Indeed, when instructed to answer “normally”, for example, Gypsies are considered as nomads and/or musicians. By contrast, when participants are asked to answer as the French in general would do, Gypsies are presented as thieves (64% in a substitution context vs. 26% in a “normal” context). Thus, with regard to various objects of studies such as “Gypsies” (Guimelli & Deschamps, Reference Guimelli and Deschamps2000), “Insecurity” (Deschamps & Guimelli, Reference Deschamps, Guimelli, Beauvois, Joule and Monteil2004), “Maghrebins” (Abric, Reference Abric and Abric2003), “Islam” (Flament, Guimelli, & Abric, Reference Flament, Guimelli and Abric2006), “the Muslim community” (Guimelli, Lo Monaco, & Deschamps, Reference Guimelli, Lo Monaco and Deschamps2010), “the Woman at work” and “Male/female work” (Flament et al., Reference Flament, Guimelli and Abric2006), “Alcohol” (Lo Monaco et al., Reference Lo Monaco, Lheureux, Chianèse, Codaccioni, Halimi-Falkowicz and Cano2009), it is under this context of substitution that some opinions, beliefs or attitudes, which are not usually expressed under the so-called “normal” conditions, appear. From a theoretical viewpoint, as suggested by Guimelli and Deschamps (2000, p. 53) this is a social desirability effect related to the salience of a tolerance norm in self-presentation which leads participants to mask their negative opinions.

Another way to reveal social desirability issues related to negative opinions has been explored by means of the normative decontextualization technique (Flament et al., Reference Flament, Guimelli and Abric2006; Lo Monaco & Guimelli, Reference Lo Monaco and Guimelli2011; Lo Monaco et al., Reference Lo Monaco, Lheureux, Chianèse, Codaccioni, Halimi-Falkowicz and Cano2009). Directly inspired by previous research on the well-known experimenter effect (e.g., Anderson, Silver, & Abramson, Reference Anderson, Silver and Abramson1988; Campbell, Reference Campbell1981; Davis, Reference Davis1997; Fazio, Jackson, Dunton, & Williams, Reference Fazio, Jackson, Dunton and Williams1995; Finkel, Guterbock, & Borg, Reference Finkel, Guterbock and Borg1991; Gilbert & Hixon, Reference Gilbert and Hixon1991; Schuman & Converse, Reference Schuman and Converse1971; Stangor, Swim, Van Allen, & Sechrist, Reference Stangor, Swim, Van allen and Sechrist2002; Sudman & Bradburn, Reference Sudman and Bradburn1974), the experimental manipulation is no longer focused on variations of the instruction but on the presentation of the experimenter. Indeed, Gilbert and Hixon (Reference Gilbert and Hixon1991) and Stangor et al. (Reference Stangor, Swim, Van allen and Sechrist2002) have already used this variable. Therefore, in relation to these previous results, one can follow Abric for whom normative decontextualization works by constructing a mediating event (interview) to facilitate the expression of counternormative positions when interviewer and respondent are characterized by the same group membership while inhibiting expression of these aspects when the respondent perceives the interviewer as dissimilar, or even as opposite. For example, concerning the SR of Islam, Flament et al. (Reference Flament, Guimelli and Abric2006) have manipulated both the presentation of the experimenter and the instruction. The manipulation of these independent variables allowed the authors to reveal a unique dimension called “masking to unmasking” by means of a Guttman effect in the principal component analysis (Flament et al., Reference Flament, Guimelli and Abric2006; Flament & Milland, Reference Flament, Milland and Abric2003, Reference Flament and Milland2005; Gaymard, Reference Gaymard2006; Lo Monaco & Guimelli, Reference Lo Monaco and Guimelli2011). This analysis showed that, on the basis of the answers to 10 Likert scales related to the SR of Islam, the four experimental conditions resulting from the combination of the two manipulated variables were hierarchized along this dimension. Then, the most favourable condition concerning masking strategies was characterized by the combination of a dissimilar interviewer with a “normal” context of expression while the most favourable condition concerning unmasking was characterized by the combination of a similar interviewer with a substitution context of expression.

All these studies contributed significantly to identifying conditions and objects related to social desirability issues.

However, since the work carried out by Guimelli and Deschamps (Reference Guimelli and Deschamps2000) we should highlight the fact that other research has proposed new methodologies and new ways of interpretation, sometimes competing ones. In this regard, it can be observed that a debate has arisen around the question of issues related to the use of the substitution technique. Indeed, following the studies conducted by Guimelli and Deschamps (Reference Guimelli and Deschamps2000) and Deschamps and Guimelli (Reference Deschamps, Guimelli, Beauvois, Joule and Monteil2004), an alternative interpretation of the effects of the substitution instruction in terms of social comparison has been proposed by other authors (Chokier & Moliner, Reference Chokier and Moliner2006; Chokier & Rateau, Reference Chokier, Rateau, Rateau and Moliner2009). It should be also noted that, at the same time, the interpretation of the decrease in normative pressure has been maintained (Flament et al., Reference Flament, Guimelli and Abric2006; Guimelli et al., Reference Guimelli, Lo Monaco and Deschamps2010; Lo Monaco et al., Reference Lo Monaco, Lheureux, Chianèse, Codaccioni, Halimi-Falkowicz and Cano2009; Lo Monaco & Guimelli, Reference Lo Monaco and Guimelli2011). Finally, the work conducted by Flament and Milland (Reference Flament and Milland2010) on the comparison of the two competing interpretations favoured the interpretation in line with the reduction in normative pressure.

It is important to specify that the conditions for this debate are based on the participant being led to express him/herself as members of a wider group would do and it is for this reason that the social comparison has been considered as a way of interpreting the effects of the substitution instruction. Nevertheless, the literature reveals another context manipulation that allows us to go beyond this debate, in that the participant is led to fill in the questionnaire always according to his/her own point of view and never as the French would do, as is the case for the substitution technique Footnote 3 . This context has been used for a long time outside the field of the study of SR but it has been drawn on in research topics very closely related to the questions tackled by work on the mute zone of SR. Indeed, it concerns the inhibition or, on the contrary, the facilitation of the expression of socially undesirable content. In this regard, we can point to the work carried out on the expression of prejudice towards African-Americans (Stangor et al., Reference Stangor, Swim, Van allen and Sechrist2002), or even on the fact of whether or not to report our drink-driving behaviour, (Lajunen & Summala, Reference Lajunen and Summala2003). In that perspective, the differences between the private and the public context of expression which it highlights reveal a strategy in terms of impression management (e.g., Goffman, Reference Goffman1959; Kuran, Reference Kuran1995; Lajunen & Summala, Reference Lajunen and Summala2003). Lo Monaco (Reference Lo Monaco and Guimelli2008) considered the manipulation of this context as a promising area of research within the theoretical framework of the mute zone of SR. It should be also noted that these predictions have been confirmed in the recent work conducted by Hidalgo (Reference Hidalgo2012). Thus, in accordance with Hidalgo (Reference Hidalgo2012), Lajunen and Summala (Reference Lajunen and Summala2003), and Stangor et al. (Reference Stangor, Swim, Van allen and Sechrist2002) one can expect a masking strategy of counterternormative opinions which could be more pronounced in a public context of expression than in a private one.

Considering these different aspects, the results obtained by Guimelli and Deschamps (Reference Guimelli and Deschamps2000) on the Gypsy community offer an object of study that allows to work on counter-normative opinions. Moreover, it provides base line to seize the variations in the expression of the participants following the context manipulations. Therefore, the aim of the present study is threefold. Indeed, firstly, we will focus on the manipulation of a different context of expression which is free from the debate concerning the substitution instruction. Secondly, to go further and in line with previous research, this study will also manipulate the experimenter status (Flament et al., Reference Flament, Guimelli and Abric2006; Hidalgo, Reference Hidalgo2012; Lo Monaco et al., Reference Lo Monaco, Lheureux, Chianèse, Codaccioni, Halimi-Falkowicz and Cano2009; Lo Monaco & Guimelli, Reference Lo Monaco and Guimelli2011). Thirdly, the data will be analyzed through a pluri-methodological approach (Apostolidis, Reference Apostolidis and Abric2003; Denzin, Reference Denzin1978). Indeed, such a way to proceed could allow us to highlight the reliability of the expected effects.

Method

Overview

In the present study, we used a hierarchical evocation task (Abric, Reference Abric and Abric2003) to identify the SR of the Gypsy community among 240 students.

Two independent variables were manipulated, the context (public vs. private) and the experimenter status (Gypsy experimenter vs. non-Gypsy experimenter). The results will be presented in four steps following a methodological triangulation procedure (Apostolidis, Reference Apostolidis and Abric2003; Denzin, Reference Denzin1978). Firstly, using prototypical analysis, we will focus on the frequencies and the average importance attributed to the words associated with the Gypsies from the perspective of the general sample, and at the level of the various experimental conditions (i.e., public/Gypsy experimenter; public/non-Gypsy experimenter; private/Gypsy experimenter; private/non-Gypsy experimenter).

Then, with the aim of identifying the results linked to the word association task with greater accuracy, we used a discursive corpus of sentences that participants were asked to produce in order to clarify the words or phrases which they had associated with the word “Gypsy”. These sentences, obtained on the basis of what we suggest calling “semantic contextualization”, will be analysed via a lexicometric analysis using the Alceste program (Reinert, Reference Reinert1993). As part of a third step, using correspondence factor analysis (CORR. F. A., Benzécri Reference Benzécri1976; Cibois, Reference Cibois1983; Deschamps, Reference Deschamps and Abric2003; Lo Monaco & Guimelli, Reference Lo Monaco and Guimelli2008; Lo Monaco, Piermattéo, Guimelli, & Abric, Reference Lo Monaco, Piermattéo, Guimelli and Abric2012), we will compare these experimental conditions on the basis of the frequencies of word associations (Deconchy & Deschamps, Reference Deconchy and Deschamps2006). Finally, we will focus on the attitudinal valence of the associated words with a view to highlighting the masking/unmasking phenomenon on the basis of the results of previous research (e.g., Flament et al., Reference Flament, Guimelli and Abric2006; Gaymard, Reference Gaymard2006; Lo Monaco & Guimelli, Reference Lo Monaco and Guimelli2011; Lo Monaco et al., Reference Lo Monaco, Lheureux, Chianèse, Codaccioni, Halimi-Falkowicz and Cano2009). To do this, we will carry out a contrast analysis (Rosenthal, Rosnow, & Rubin, Reference Rosenthal, Rosnow and Rubin2000).

Population

240 students (166 women, M age = 22.30; SD = 3.60) from a French university were asked to fill in a questionnaire. Subjects were randomly divided amongst four experimental conditions defined by a factorial combination of the context (private vs. public) and the experimenter status (non-Gypsy experimenter vs. Gypsy experimenter).

Procedure

Verbal association task

Participants were asked to fill in a hierarchal evocation questionnaire (Abric, Reference Abric and Abric2003). This method is based on free association which, according to Abric (1994, p. 66) “is a fundamental technique for collecting the content of social representations”. This remark is also supported by the large number of studies conducted on different types of objects (e.g., de Rosa, Reference de Rosa1988, Reference de Rosa1993; Deschamps & Guimelli, Reference Deschamps and Guimelli2002, Reference Deschamps, Guimelli, Beauvois, Joule and Monteil2004; Deschamps, Paez, & Pennebaker, Reference Deschamps, Paez and Pennebaker2001; Di Giacomo, 1981, Reference Di Giacomo, Doise and Palmonari1986; Guimelli & Deschamps, Reference Guimelli and Deschamps2000; Le Bouedec, Reference Le Bouedec1984; Lo Monaco et al., Reference Lo Monaco, Lheureux, Chianèse, Codaccioni, Halimi-Falkowicz and Cano2009; Lo Monaco & Guimelli, Reference Lo Monaco and Guimelli2008; Marková et al., Reference Marková, Moodie, Farr, Drozda-Senkowska, Erös, Plichtová and Mullerová1998; Moscovici, Reference Moscovici1976; Piermattéo, Lo Monaco, Guimelli, & Brel, Reference Lo Monaco, Piermattéo, Guimelli and Abric2012; Wagner, Valencia, & Elejabarrieta, Reference Wagner, Valencia and Elejabarrieta1996).

Participants were asked to associate four words or phrases with the object of SR under study (i.e., the Gypsies). Then, they were asked to rank the four words or phrases from the most important (1) to the least (4). As a consequence, it was possible for each word or phrase associated with the inductor to be characterized by a value corresponding to its average importance in the representational field. Thus, it was possible to cross the average mean with its frequency to be able to formulate hypotheses about the central or peripheral status of the words or phrases (Abric, Reference Abric and Abric2003; Ernst, Delouvée, & Roland-Lévy, Reference Ernst-Vintila, Delouvée and Roland-Levy2011; Gaymard, Reference Gaymard2006). Following this second step, we asked participants to estimate the positive versus negative character for each one of the answers associated with the inductor. This assessment was conducted using a 7-point Likert scale from 1 (absolutely negative) to 7 (absolutely positive). Proceeding in this way presents the advantage of highlighting potential contradictions in participants’ responses. Indeed, such contradictions are often able to reveal the specificities of the internal organization of the representational field and can be very useful for understanding the logic guiding of this organization (Guimelli & Deschamps, Reference Guimelli and Deschamps2000; Lo Monaco & Guimelli, Reference Lo Monaco and Guimelli2008; Tavani, Reference Tavani2012).

In order to improve the contextualization of the words or phrases associated by the participants, they were asked to formulate four sentences on the basis of the four words or phrases associated with the inductor. This way of proceeding is directly inspired by the previous work conducted by Guimelli (Reference Guimelli1996, Reference Guimelli and Abric2003) within the framework of the basic cognitive schemes model. However, in this study we decided to go further by asking the participants to specify the meaning of their association within a specific semantic context in order to faciliate the thematic analysis following all verbal association tasks. In our opinion, this technique, that we suggested calling “semantic contextualization”, would improve the objectivity of the thematic analysis and provide a corpus of sentences for a lexicometric analysis of the data.

Context manipulation

In the present study, the context variable refers to two specifc contexts related to the answers given by the participants, a private context and a public one. The private context consists of assuring participants that their answers will remain anonymous. They are also asked to place their completed questionnaire into a closed envelope before giving it back to the experimenter in order to ensure real confidentiality. In contrast, in the direct public context participants do not complete the questionnaire by themselves. Effectively, as in the previous studies by Blanchard, Crandall, Brigham, and Vaughn (Reference Blanchard, Crandall, Brigham and Vaughn1994), Blanchard, Lily, and Vaughn (Reference Blanchard, Lilly and Vaughn1991), Monteith, Deenen, and Tooman (Reference Monteith, Deneen and Tooman1996), Plant and Devine (Reference Plant and Devine1998), Sechrist, Swim, and Stangor (Reference Sechrist, Swim and Stangor2004), and Stangor et al. (Reference Stangor, Swim, Van allen and Sechrist2002), the experimenter reads the questions and response procedures to the participants who have to give their answer orally to the former who completes the questionnaire herself. This direct public context (as opposed to the anticipated public context described by Lambert et al., Reference Lambert, Payne, Jacoby, Shaffer, Chasteen and Khan2003) also required that participants revealed their first name, last name and their student ID number as a way of increasing the visibility of their answers.

Manipulation of the experimenter status

This variable, linked to the presentation of the experimenter, includes 2 modalities: same ethnic origin (Parisian, i.e., non-Gypsy) and different ethnic origin (Gypsy).

In the experiment conducted by Flament et al. (Reference Flament, Guimelli and Abric2006), this aspect was manipulated by means of the physical appearance related to the ethnic membership of the experimenter. Indeed, in the works conducted on SR of Islam, Flament et al. (Reference Flament, Guimelli and Abric2006) asked the experimenter, who declared her name to be Yamina in the specific experimental conditions, to wear a headscarf. In the present study, appearance could not be manipulated, or only with great difficulty, so we focused on manipulating the information given to the participants.

Therefore, before presenting the questionnaire to the participants, in the “non-Gypsy” condition, the experimenter pretended to be a Parisian named Laure Durand, conducting research on the social image of Gypsies as part of her social science studies. In the “Gypsy” condition, the same experimenter pretended to be a Gypsy named Carmen Flores conducting research on the social image of Gypsies as part of her social science studies.

Results

Hierarchical Evocations

On the basis of the associated and ranked words, we proceeded to perform a prototypical analysis using a cross-table which included the frequency and average importance criteria of induced words (cf. Table 1).

Table 1.

General results in terms of frequency and average importance associated with the categories of word reported by the participants

Firstly, concerning the analysis of the words produced by the participants, two sub-categories seem to be distinguishable among the elements characterized by both a high frequency and importance. The first sub-category seems to refer to the different terms characterizing this group and its lifestyle (e.g., “Nomad”, “Community”, “Travellers”), whereas the second one seems to describe the way in which this group is perceived by society (e.g., “Social exclusion”, “Stigmatization”). These elements which potentially constitute the central core of the SR of Gypsies (Abric, Reference Abric and Abric2003) refer to a representational content focused on descriptive aspects. Nevertheless, although these elements seem to be rather “neutral”, those composing the first periphery (i.e., elements characterized by both high frequency and low importance) are more negative (e.g., “poverty”, “nuisance” and “thieves”).

In the second periphery (i.e., elements characterized by both low frequency and low importance), some negative elements are also associated (e.g. “Mendicancy”, “Communitarianism”, “Dirty”).

Finally, the contrasted elements (i.e., elements characterized by low frequency and high importance which refer to the possible existence of sub-groups among the population) include words related to children’s issues and to the distinctive lifestyle which characterizes the Gypsies and also the theme of solidarity.

To conclude this first analysis, we note that the majority of the induced words collected are more or less in line with the representational content observed by Guimelli and Deschamps (Reference Guimelli and Deschamps2000). However, some differences can be noted concerning in particular the presence of elements underlining the stigmatization of this population among the potentially central elements of this SR. Nevertheless, it seems difficult to take this comparison with Guimelli and Deschamps' work further as our study draws on an experimental design involving two independent variables. Despite everything, the term “Thefts”, which was one of the most salient and focusing elements in the work of Guimelli and Deschamps (Reference Guimelli and Deschamps2000), is highlighted once again within the framework of this study under the form “Thieves”. To be brief, and given the illustrative character of this element, we propose to present the different experimental conditions by focusing on it (cf. Table 2).

Table 2.

Frequencies and average importance associated with the word “thief” according to the experimental condition

Note: The first number refers to the frequency while the number in parentheses refers to the average importance.

Lexicometric Analysis

On the basis of the sentences written by the participants to explain the words related to the word association task, we conducted an automatic content analysis using the Alceste program (Reinert, Reference Reinert1993). As part of this analysis this discursive content was organized in accordance with the four experimental conditions that compose this study. The results highlight two stable lexical classes related to our inductive word: “Gypsies”.

As we can see in Table 3, the first class represents 53% of the total content produced by the participants and refers to the culture and the folklore associated with the gypsy community. The content composing this class is significantly related to the public condition, χ2(1) = 7.00, p < .01, and is therefore, more specifically mentioned when participants give their answers in this experimental condition. Effectively, in such a public context of expression, the participants, who are asked to express themselves on the theme of the Gypsies, seem to produce answers that can be considered as descriptive or even normative, referring to the practices, cultural habits and community functioning that characterize this group.

Table 3.

Themes and words associated with the two classes composing the sentences related to the word association task

This first class is composed of three sub-themes. The first one concerns community activities and relates to terms that compose a content widely based on the Gypsies’ artistic values and their festal activities. The two other categories refer respectively to the culture and to the Spanish origins of the Gypsies and also to the artistic folklore of this group.

The second class represents 47% of the total content and refers principally to the social context in which Gypsies live and to the way in which they are perceived by society. Unlike the first class, this one is significantly linked to private conditions of expression, χ2(1) = 7.00, p < .01, and seems to principally relate to negative aspects characterizing the Gypsies. We can also establish a link between these results and the categories observed in connection with the analysis of the word association task. More precisely, this class is characterized by terms such as “thieves”, χ2(1) = 6.00, p < .05, “stealing”, χ2(1) = 8.00, p < .01, and “poor”, χ2(1) = 12.00, p < .001, terms that, for the majority, compose the peripheral system of the SR of Gypsies for the whole sample of participants.

This analysis emphasises the impact of the manipulation of the context of expression on the discourse of the participants. It thus appears that the private context favours the expression of the negative aspects of the Gypsies. Moreover, these results confirm the thematic analysis of the free associations and highlight the value of contextualizing the participants’ answers.

Correspondence Factor Analysis (CORR. F. A.)

All the associations were submitted to a CORR. F. A. (Benzécri, Reference Benzécri1976; Deschamps, Reference Deschamps and Abric2003; Guimelli & Deschamps, Reference Guimelli and Deschamps2000; Lo Monaco & Guimelli, Reference Lo Monaco and Guimelli2008; Mouret, Lo Monaco, Urdapilleta, & Parr, 2013). This factorial analysis highlights the differences in terms of frequencies of association relative to the independent variables. It gives access to a summary of the data by revealing a particular structure. It also enables the identification of the most significant factorial axes (Deschamps, Reference Deschamps and Abric2003; Doise, Clémence, & Lorenzi-Cioldi, Reference Doise, Clemence and Lorenzi-Cioldi1993). The axes are composed of the different modalities of the independent variables. Finally, it emphasizes the correspondences between the modalities of the independent variables and the words or phrases associated by the participants.

Our analysis was based on word categories (N = 22) created a posteriori according to a principle of semantic proximity (these word categories were used in the framework of the hierarchical associations analysis).

We then performed the CORR. F. A. using as independent variables the different experimental conditions (i.e., public-non-Gypsy; private-non-Gypsy; public-Gypsy; private-Gypsy; Deconchy & Deschamps, Reference Deconchy and Deschamps2006). In order to decide whether an observation (i.e., a word or phrase) contributed to the formation of a factor, we considered the recommendations formulated by Deschamps (Reference Deschamps and Abric2003) and used in previous and recent works on SR (Lo Monaco & Guimelli, Reference Lo Monaco and Guimelli2008; Lo Monaco et al., Reference Lo Monaco, Piermattéo, Guimelli and Abric2012; Mouret et al., Reference Mouret, Lo Monaco, Urdapilleta and Parr2013) in termes of contribution by factor (CF) Footnote 1 . The two first factors represented 86.20% of inertia (i.e., Factor 1 = 59.62%; Factor 2 = 26.58%). The variables which contribute to the formation of the first factor are the “private-non-Gypsy” and the “public-Gypsy” conditions: CF (private-non-Gypsy) = .46 + CF (public-Gypsy) = .33, thus a total contribution of 79% for the formation of factor 1. Factor 2 is formed by the contribution of the “private-non-Gypsy” and the “private-Gypsy” conditions: CF (private-non-Gypsy) = .27 + CF (private-Gypsy) = .52, thus a total contribution of 78% for the formation of factor 2. We observe (cf. Figure 1) Footnote 2 specific correspondences between the “private-non-Gypsy” experimental condition and the following observations: “social exclusion”, “thieves”, “dirty” and “Gypsy-characters”.

Figure 1.

Graphical representation of the results obtained by means of the CORR. F. A. concerning factors 1 and 2.

Note: Grayed blocks refer to the experimental conditions. “ Experimental conditions” contribute to the formation of Factor 1; “Experimental conditions” refer to the experimental conditions which contribute to the formation of Factor 2; “ Experimental condition ” refer to the experimental conditions which contribute to the formation of both Factors 1 and 2. “Observations” refer to the observations which contribute to the formation of Factor 1; “Observations” refer to the observations which contribute to the formation of Factor 2; “ Observations ” refer to the observations which contribute to the formation of both Factors 1 and 2.

The assignment of observations on the two factors shows the impact of the manipulated variables on the production of the participants. One can observe a contradiction in particular between the “public-Gypsy” and “private-non-Gypsy” conditions. These two conditions are characterized by divergent content related to associative logic consistent with our hypotheses. Thus, in a public context and faced with a Gypsy experimenter, participants provided a stereotypical view of this group with descriptive aspects. In this case, Gypsies are presented as travellers, of Spanish origin and characterized by a culture, an artistic folklore and a specific appearance (i.e., “general prototype”). On the other hand, when participants answered in a private context and faced with a non Gypsy experimenter, they express terms related to nuisances caused by Gypsies (e.g., perceived danger, problems caused, and the fact that they take advantage of the system) and the fact they are victims of stigmatization. It appears that participants were more able to express counternormative content when they were placed in conditions which aimed at promoting the expression of such points of view. Indeed, in the “private-non-Gypsy” condition, participants mentioned “thieves” significantly more frequently (37.93%) than in the “public-Gypsy” condition (24.14%), χ2(1) = 3.95, p < .05. Moreover, participants mentioned the element “dirty” significantly more frequently in the “private-non-Gypsy” condition (54.55%) than in the “public-Gypsy” condition (9.09%), χ2(1) = 46.53, p < .001.

The CORR. F. A. thus highlights the results observed by means of the qualitative analysis of the sentences. However, more relevant to our study is the fact that by means of the CORR. F. A. it was possible to reveal the cross-effect of the independent variables. In this perspective, the CORR. F. A. allows us to confirm the differences observed between the experimental conditions within the framework of the prototypical analyses by highlighting the significant differences in terms of frequencies of association.

Attitudinal Valence of the Beliefs Associated With Gyspsies

In order to highlight the results obtained by means of the CORR. F. A. concerning the experimental conditions, we focused on the average attitudinal scores related to the associations produced by the participants (cf. Table 4). As a reminder, the participants were asked to assess their associations by means of 7-point Likert scales (from 1 = very negative to 7 = very positive). Such a measure enables us to clarify the meaning of the associations by asking the participants to take a position on their own production. On the basis of such assessments, the different experimental conditions were submitted to a contrast analysis (Rosenthal et al., Reference Rosenthal, Rosnow and Rubin2000). Consequently, on the basis of previous works including a similar design (Flament et al., Reference Flament, Guimelli and Abric2006; Gaymard, Reference Gaymard2006; Lo Monaco & Guimelli, Reference Lo Monaco and Guimelli2011; Lo Monaco et al., Reference Lo Monaco, Lheureux, Chianèse, Codaccioni, Halimi-Falkowicz and Cano2009) the different experimental conditions were ordered as follows: private-non-Gypsy; private-Gypsy; public-non-Gypsy; public-Gypsy. We hypothesized a linear sequence ranging from the experimental condition which was the most favourable for the expression of counternormative aspects to the experimental condition which was the most favourable for the expression of normative content. According to Rosenthal et al. (Reference Rosenthal, Rosnow and Rubin2000), in order to show a linear effect, we must observe a significant linear contrast and not significant quadratic and cubic ones.

Table 4.

Average importance and standard deviation related to attitudes regarding associated words according to experimental conditions

Note: The first number refers to the mean while the number in parentheses refers to standard deviation.

The analyses revealed a significant effect of the general model concerning the ordering of the experimental conditions, F(3, 235) = 5.19, p < .01, η p ² = .06. In this regard we obtained a significant linear effect, t(238) = 3.69, p < .001, d = 0.48. Otherwise, neither the quadratic contrast, t(238) = 1.19, p = .23, d = 0.15, nor the cubic one, t(238) = 0.62, p = .53, d = 0.08, reached the conventional threshold of significance. Considering these results and the experimental conditions, it seems possible to conclude a linear order of the experimental conditions as follows: private-non-Gypsy > private-Gypsy > public-non-Gypsy > public-Gypsy.

Nevertheless, despite the observed linear effect, one cannot ignore the proximity between the means relative to the “private-Gypsy” and “private-non-Gypsy” conditions. Thus, it seems that in a private context, the evaluation of the associated beliefs did not depend on the status of the experimenter. Otherwise, the linear effect indicates that the associated beliefs are more positively evaluated in a public context of expression especially when the experimenter is presented as a Gypsy.

Discussion

On the basis of these results, the emergence of cross-cutting trends can be observed across all the analyses we conducted. Effectively, we observe a rather regular presence of positive and negative elements concerning the Gypsies among the participants' answers. Thus, beyond a great deal of descriptive and normative content such as the categories “nomadic”, “community”, “solidarity” or “conviviality”, other categories like “thieves”, “dirty” and “nuisances” could be noticed. These different categories of content appear within the framework of the prototypical analysis but also at the level of the qualitative analysis of the sentences supplied by the participants in order to contextualize their answers, thus providing us with clues which would help improve the objectivity of the facilitating the categorization of the associated words. This qualitative analysis allowed us to distinguish two classes. The first one is related to content that could be considered as descriptive and/or normative while the second one is composed of a more negative content. This distinction confirms the results of the prototypical analysis and the relevance of the categorization carried out on the basis of the words associated by the participants.

Here we can see that more than 13 years after the study conducted by Guimelli and Deschamps (Reference Guimelli and Deschamps2000), the SR of the Gypsies is still characterized by the presence of normative and counternormative content. Therefore Gypsies could still be considered as a “sensitive object”. As these authors have highlighted, the presence of such counternormative content entails the use of specific methodologies with a view to collecting those elements which are less likely to be found in standard conditions of questioning. Thus, the methodologies that we employed, and more specifically, the variation of the context of expression, led us to update this content. We thus note, as we have mentioned above in the text, the presence of the terms “thieves” and “dirty”, terms that were also observed by Guimelli and Deschamps (Reference Guimelli and Deschamps2000). However, despite these similarities that stress the relative stability of the SR (Guimelli, Reference Guimelli1988; Moscovici, Reference Moscovici1976), the presence of counternormative elements is far less noticeable within the framework of the present study, at least in relation to the frequencies of associated words. Although the hypothesis of an evolution of the social context in a way that strengthens a non-discrimination norm could hardly be ruled out, it seems more plausible to consider that such a difference between the results of these studies could be the consequence of the use of different methodologies in order to gather counternormative content. However this last hypothesis does not necessarily imply that the manipulations of the context of expression or the normative decontextualization are less effective than the substitution technique. Effectively, these two techniques are characterized by a fundamental difference that consists of putting the participant in a situation where he has to answer in his own name, as opposed to the substitution technique. It seems therefore, reasonable to think that, the normative pressure leads participants to inhibit the expression of counternormative content, these techniques are more threatening as they increase the involvement of the participant with regard to his own expression. The consequence of this difference may therefore be the reduced presence of counternormative content. However, these techniques have the advantage of gathering content that conveys what participants have in mind more accurately than the substitution technique which could lead to the collection of a potentially wrong appraisal of the social norms that characterize another group of people as mentioned in the literature on pluralistic ignorance (see Miller, Monin, & Prentice, Reference Miller, Monin, Prentice, Terry and Hogg2000; Prentice & Miller, Reference Prentice and Miller1993).

Still, we also notice that these two methodologies are not equivalent in terms of their efficacy and it was mostly the manipulation of the context of expression that led us to observe differences in the answers given by the participants. In this respect, the pluri-methodological stance adopted in the present study allowed us to show the impact of this variable in the course of the many analyses we performed. Indeed, we observe variations in the mention of the word “thieves” in terms of frequency (i.e., prototypical analyses), or in terms of the differentiation of the two classes (i.e., lexicometric analysis) or in terms of a stronger correspondence with specific experimental conditions (i.e., CORR. F. A.) and finally, in terms of a contradiction between attitudinal valences on the linear effect (i.e., contrasts analyses). However, the second variable we manipulated did not achieve the same success. This limited effect may be explained by the operationalization of this variable. Effectively, the experimenter has to present herself as a Parisian girl or as a Gypsy according to the experimental conditions. It is therefore possible that some observable clues (e.g., her way of expression, dressing, etc.) could have contributed to lowering the credibility of this manipulation. Otherwise, unlike the experimental induction conducted by Flament et al. (Reference Flament, Guimelli and Abric2006) which was based on the ethnic membership of the experimenter characterized by several physical aspects, within the framework of this study it was quite difficult to proceed in such a way. Thus, we cannot be as sure as these authors were about the categorization of the experimenter as a Parisian or a Gypsy student.

Nevertheless, all the research conducted in the area of the study of the mute zone of SR and particularly that which has drawn on this kind of methodology has highlighted a more significant effect of the context of expression (e.g., substitution technique) compared to normative decontextualization once the two variables are used together. Nevertheless, despite this limited impact of the “normative decontextualization” variable, the congruity with the initial results of Guimelli and Deschamps (Reference Guimelli and Deschamps2000) and the strong regularity of the results observed in the multiple analyses we conducted, emphasise the dual value of manipulating the context of expression as part of the collection of content in relation to sensitive objects and the updating of the results related to these objects.

From these positive aspects stems one of the main limitations of this study. Effectively, by conducting research on the same object studied by Guimelli and Deschamps (Reference Guimelli and Deschamps2000) 13 years earlier, and by using rival techniques to the substitution technique, it would be possible to consider that a direct comparison of these techniques within the framework of the same study would have been more relevant. However, as the use of the same object by Guimelli and Deschamps (Reference Guimelli and Deschamps2000) seemed more appropriate for demonstrating the efficacy of the variables that were manipulated, the aim of this study is not to compare these techniques with the substitution technique.

Indeed, as indicated above, these techniques are fundamentally different from the substitution technique in that they don't involve individuals in the same way and that they favour collections of a different nature, more related to what individuals have in their own mind than to their perception of the norms of how the object is perceived. So, these techniques are characterized by their own advantages and flaws, and are to our point of view, not directly comparable.

Nevertheless it would be interesting to combine these collection techniques. Indeed, this study opens the way to other experiments which would allow us to support these reflections by combining, for example, manipulation of the context of expression with the substitution technique. Such developments will help to further knowledge on techniques that aim to collect counternormative content. Moreover, such techniques are not limited to the field of SR. Indeed, numerous sensitive objects are studied outside of the conceptual and methodological framework of this field, as in the case, for example, of opinion surveys. Within such a framework, the use of manipulation of the context of expression could lead to a better knowledge of the way individuals regard a sensitive object and to a better prediction of their standpoint.

Footnotes

1. The CF of an observation or of a modality of an independent variable must be higher than the average CF to be considered as contributing to the formation of the factor (thresholds are the following: .25 for the variables and .04 for the observations).

2. The “public-non-Gypsy” condition is absent from Figure 1 because its contribution was lower than .25.

3. We can notice the existence of another debate linked to the substitution technique. Indeed, one can consider that the expression of counternormative contents associated to the substitution technique could be attributed to the awareness of the representation shared by the group in the name of whom the participants are asked to express themselves (for more details about this alternative explanation see Flament et al., Reference Flament, Guimelli and Abric2006).

References

Abric, J.-C. (1994). Méthodologie de recueil des représentations sociales [social representations gathering methodology]. In Abric, J.-C. (Ed.), Pratiques sociales et représentations [Social practices and representations] (pp. 59–82). Ramonville Saint-Agne, France: Erès.Google Scholar
Abric, J.-C. (2003). La recherche du noyau central et de la zone muette des représentations sociales [The searching for the central core and the dumb zone of social representations]. In Abric, J.-C. (Ed.), Méthodes d’étude des représentations sociales [Methods of study of social representations] (pp. 59–80). Ramonville Saint-Agne, France: Erès.Google Scholar
Anderson, B. A., Silver, B. D., & Abramson, P. R. (1988). The effects of the race of interviewer on race-related attitudes of Blacks in SRC/CPS National Election Studies. Public Opinion Quarterly, 52, 289–324. http://dx.doi.org/10.1086/269108 Google Scholar
Apostolidis, T. (2003). Représentations sociales et triangulation: Enjeux théorico-méthodologiques [Social representations and triangulation: Theoretical and methodological issues]. In Abric, J. C. (Ed.), Méthodes d’étude des représentations sociales (pp. 13–35). Ramonville Saint-Agne, France: Erès.Google Scholar
Benzécri, J. P. (1976). L’analyse des correspondances [The correspondence analysis] . Paris, France: Dunod.Google Scholar
Blanchard, F. A., Crandall, C. S., Brigham, J. C., & Vaughn, L. A. (1994). Condemning and condoning racism: A social context approach to interracial settings. Journal of Applied Psychology, 79, 993–997. http://dx.doi.org/10.1037/0021-9010.79.6.993 Google Scholar
Blanchard, F. A., Lilly, T., & Vaughn, L. A. (1991). Reducing the expression of racial prejudice. Psychological Science, 2, 101–105. http://dx.doi.org/10.1111/j.1467-9280.1991.tb00108.x Google Scholar
Campbell, B. (1981). Race-of-interviewer effects among southern adolescents. Public Opinion Quarterly, 45, 231–244. http://dx.doi.org/10.1086/268654 Google Scholar
Chokier, N., & Moliner, P. (2006). La « zone muette » des représentations sociales, pression normative et/ou comparaison sociale [The “mute zone” of social representations, normative pressure and/or social comparison] Bulletin de psychologie, 59, 281–286. http://dx.doi.org/10.3917/bupsy.483.0281 Google Scholar
Chokier, N., & Rateau, P. (2009). Représentations d’objets « sensibles » et processus de comparaison soi/autrui [Representations of “sensitive” objects and self-other comparison process]. In Rateau, P. & Moliner, P. (Eds.), Représentations sociales et processus sociocognitifs [Social representations and sociocognitive processes] . (pp. 31–44). Rennes, France: Presses Universitaires de Rennes.Google Scholar
Cibois, P. (1983). L’analyse factorielle [The factorial analysis] . Paris, France: Presses Universitaires de France.Google Scholar
Davis, D. W. (1997). Nonrandom measurement error and race of interviewer effects among African Americans. Public Opinion Quarterly, 61, 183–207. http://dx.doi.org/10.1086/297792 Google Scholar
de Rosa, A. S. (1988). Sur l’usage des associations libres dans l’étude des représentations sociales de la maladie mentale [On the use of free-associations in the study of the social representations of mental illness]. Connexions, 51, 27–50.Google Scholar
de Rosa, A. S. (1993). Social representations and attitudes: problems of coherence between the theoretical definition and procedure of research. Papers on Social Representations, 2, 1–15.Google Scholar
Deconchy, J.-P., & Deschamps, J.-C. (2006). Croyances et controle. D’un discours “forcé” à un discours “spontané” [Beliefs and control. From a “forced” discourse to a “spontaneous” one]. Revue Internationale de Psychologie Sociale, 19, 93–124.Google Scholar
Denzin, N. K. (1978). The research act: A theoretical introduction to sociological methods. New York, NY: McGraw-Hill.Google Scholar
Deschamps, J. C. (2003). Analyse des correspondances et variations des contenus des représentations sociales [Correspondence analysis and variations in the content of social representations]. In Abric, J. C. (Ed.), Méthodes d’étude des représentations sociales (pp. 179–200). Ramonville Saint-Agne, France: Erès.Google Scholar
Deschamps, J.-C., & Guimelli, C. (2002). La composante émotionnelle des représentations sociales: Émotions rapportées et tendances à l'action dans une étude comparative des représentations sociales de l'insécurité en France et en Suisse [The emotional component of social representations: Reported emotions and action tendency in the framework of a comparative study of social representations of insecurity in France and in Switzerland]. Nouvelle Revue de Psychologie Sociale, 1, 78–84.Google Scholar
Deschamps, J.-C., & Guimelli, C. (2004). L’organisation interne des représentations sociales de la sécurité/insécurité et l’hypothèse de la « zone muette » [The inner organization of social representations of security/insecurity and the “mute zone” hypothesis]. In Beauvois, J.-L., Joule, R.-V. & Monteil, J.-M. (Eds.), Perspectives cognitives et conduites sociales [Cognitive perspectives and social behaviors] , (Vol. 9., pp. 105–130). Rennes, France: Presses Universitaires de Rennes.Google Scholar
Deschamps, J.-C., Paez, D., & Pennebaker, J. (2001). Mémoire collective des évènements socio-politiques et culturels: Représentation sociale du passé à la fin du millénium [Collective memory of socio-political and cultural events: Social representation of the past at the end of the millennium]. Psychologie et Société, 2, 26–53.Google Scholar
Di Giacomo, J.-P. (1981). Aspects méthodologiques de l’analyse des représentations [Methodological aspects of the analysis of representations]. Cahiers de Psychologie Cognitive, 1, 397–422.Google Scholar
Di Giacomo, J. P. (1986) Alliance et rejets intergroupes au sein d'un mouvement de revendication [Intergroup alliances and rejections at the heart of a protestation movement]. In Doise, W. & Palmonari, A. (Eds.), L'ètude des reprèsentations sociales (pp. 118–138). Neuchâtel, Suiza: Delachaux et Niestlé.Google Scholar
Doise, W., Clemence, A., & Lorenzi-Cioldi, F. (1993). The quantitative analysis of social representations. Hemel Hempstead, UK: Harvester Wheatsheaf.Google Scholar
Ernst-Vintila, A., Delouvée, S., & Roland-Levy, C. (2011). Under threat. Lay thinking about terrorism and the three-dimensional model of personal involvement. A social psychological analysis. Journal of Risk Research, 14, 297–324. http://dx.doi.org/10.1080/13669877.2010.533468 Google Scholar
Fazio, R. H., Jackson, J. R., Dunton, B. C., & Williams, C. J. (1995). Variability in automatic activation as an unobstrusive measure of racial attitudes: A bona fide pipeline? Journal of Personality and Social Psychology, 69, 1013–1027. http://dx.doi.org/10.1037//0022-3514.69.6.1013 Google Scholar
Finkel, S. E., Guterbock, T. M., & Borg, M. J. (1991). Race-of-interviewer effects in a presidential poll: Virginia 1989. Public Opinion Quarterly, 55, 313–330.Google Scholar
Flament, C., Guimelli, C., & Abric, J.-C. (2006). Effets de masquage dans l’expression d’une représentation sociale [Masking effects in the expression of a social representation]. Les Cahiers Internationaux de Psychologie Sociale, 69, 15–31.Google Scholar
Flament, C., & Milland, L. (2003). Un effet Guttman dans l’analyse des représentations sociales [A Guttman effect in the analysis of social representations]. In Abric, J.-C. (Ed.), Méthodes d’étude des représentations sociales (pp. 201–220). Ramonville Saint-Agne, France: Erès.Google Scholar
Flament, C., & Milland, L. (2005). Un effet Guttman en ACP [A Guttman effect in PCA]. Mathématiques et Sciences Humaines, 71, 25–49. http://dx.doi.org/10.4000/msh.2946 Google Scholar
Flament, C., & Milland, L. (2010). La substitution dans les études de représentations sociales: quel processus impliqué ? [The substitution technique in social representations studies: Which process involved ?]. Psychologie française, 55, 195–210. http://dx.doi.org/10.1016/j.psfr.2010.05.003 Google Scholar
Gaymard, S. (2006). The representation of old people: Comparison between the professionals and students. International Review of Social Psychology, 19, 69–92.Google Scholar
Gilbert, D. T., & Hixon, J. G. (1991). The trouble of thinking: Activation and application of stereotyping beliefs. Journal of Personality and Social Psychology, 60, 509–517. http://dx.doi.org/10.1037//0022-3514.60.4.509 Google Scholar
Goffman, E. (1959). The presentation of self in everyday life. New York, NY: Doubleday.Google Scholar
Guimelli, C. (1996). Valence et structure des représentations sociales [Valence and structure of social representations]. Bulletin de Psychologie, 49, 58–72.CrossRefGoogle Scholar
Guimelli, C. (1998). Chasse et nature en Languedoc. Étude de la dynamique d’une représentation sociale chez des chasseurs languedociens languedociens [Hunting and nature in Languedoc. The study of the dynamic of a social representation among Languedocian hunters] . Paris, France: L’Harmattan.Google Scholar
Guimelli, C. (2003). Le modèle des schèmes cognitifs de base (SCB). Méthode et applications [The basic cognitive schemes model (BCS). Methodology and applications]. In Abric, J.-C. (Ed.), Méthodes d’étude des représentations sociales (pp. 119–146). Ramonville Saint-Agne, France: Erès.Google Scholar
Guimelli, C., & Deschamps, J.-C. (2000). Effets de contexte sur la production d’associations verbales. Le cas de la représentation sociale des Gitans [Contextual effects on the production of word associations. The case of the social representation of Gypsies]. Les Cahiers Internationaux de Psychologie Sociale, 47, 44–54.Google Scholar
Guimelli, C., Lo Monaco, G., & Deschamps, J.-C. (2010). The lawsuit against “Charlie Hebdo” and its effects on the social representations of the Muslim Community. International Review of Social Psychology, 23, 5–36.Google Scholar
Hidalgo, M. (2012). Représentations sociales et contextes: Études autour de l’expression et des comportements en lien avec les éléments masqués [Social representations and contexts: Studies about the expression and behaviors related to masked elements] . Aix-en-Provence, France: Thèse de Doctorat de l’Université d’Aix-Marseille.Google Scholar
Jellison, J. M., & Green, J. (1981). A self-presentation approach to the fundamental attribution error: The norm of internality. Journal of Personality and Social Psychology, 40, 643–649. http://dx.doi.org/10.1037//0022-3514.40.4.643 Google Scholar
Kuran, T. (1995). Private truths, public lies. The social consequences of preference falsification. Cambridge, UK: Harvard University Press.Google Scholar
Lajunen, T., & Summala, H. (2003). Can we trust self-reports of driving? Effects of impression management on driver behaviour questionnaire responses. Transportation Research Part F: Traffic Psychology and Behaviour, 6, 97–107. http://dx.doi.org/10.1016/S1369-8478(03)00008-1 Google Scholar
Lambert, A. J., Payne, B. K., Jacoby, L. L., Shaffer, L. M., Chasteen, A. L., & Khan, S. R. (2003). Stereotypes as dominant responses: On the “social facilitation” of prejudice in anticipated public contexts. Journal of Personality and Social Psychology, 84, 277–295. http://dx.doi.org/10.1037//0022-3514.84.2.277 Google Scholar
Le Bouedec, G. (1984). Contribution à la méthodologie d’étude des représentations sociales [Contribution to the methodology of study of social representations]. Cahiers de Psychologie Cognitive, 4, 245–272.Google Scholar
Lo Monaco, G. (2008). Les représentations sociales du vin. Etude du rôle de quelques facteurs intervenant dans la construction sociale de la réalité [Social representations of wine. A Study of the role of some factors intervening in the social construction of reality] . Aix-en-Provence, France: Thèse de Doctorat de l’Université de Provence.Google Scholar
Lo Monaco, G., & Guimelli, C. (2008). Représentations sociales, pratique de consommation et niveau de connaissance: Le cas du Vin [Social representations, consumption practices and level of knowledge: The case of wine]. Les Cahiers Internationaux de Psychologie Sociale, 78, 35–50.Google Scholar
Lo Monaco, G., & Guimelli, C. (2011). Hegemonic and polemical beliefs: Culture and consumption in the social representation of wine. The Spanish Journal of Psychology, 14, 232–245. http://dx.doi.org/10.5209/rev_SJOP.2011.v14.n1.21 Google Scholar
Lo Monaco, G., Lheureux, F., Chianèse, L., Codaccioni, C., Halimi-Falkowicz, S., & Cano, P. (2009). Contexte d’expression, statut social des intervenants de santé et production d’un discours normatif: Le cas de l’alcool et des jeunes [Contexts of expression, social status of health actors and production of a normative discourse: The case of alcohol and young people]. Pratiques psychologiques, 15, 367–386. http://dx.doi.org/10.1016/j.prps.2008.02.002 Google Scholar
Lo Monaco, G., Piermattéo, A., Guimelli, C., & Abric, J.-C. (2012). Questionnaire of characterization and correspondence factor analysis: A methodological contribution in the field of social representations. The Spanish Journal of Psychology. 15, 1233–1243.Google Scholar
Mamontoff, A. M. (1996). Transformation de la représentation sociale de l’identité et schèmes étranges: Le cas des Gitans [Transformation of the social representation of identity and strange schemes: The case of Gypsies]. Les Cahiers Internationaux de Psychologie Sociale, 29, 64–77.Google Scholar
Mamontoff, A. M. (2008). Dynamique de la fonction identitaire des représentations sociales dans le cas d’une rencontre entre deux cultures [Dynamic of the identity function of social representations in the framework of the encounter between two cultures]. Anuario de Psicología, 39, 249–268.Google Scholar
Marková, I., Moodie, I., Farr, R., Drozda-Senkowska, E., Erös, F., Plichtová, J., ... Mullerová, O. (1998). Social representations of the individual: A post-communist perspective. European Journal of Social Psychology, 28, 797–829. http://dx.doi.org/10.1002/(SICI)1099-0992(199809/10)28:5<797::AID-EJSP896>3.0.CO;2-6 Google Scholar
Miller, D. T., Monin, B., & Prentice, D. A. (2000). Pluralistic Ignorance and inconsistency between private attitudes and public behaviors. In Terry, D. J. & Hogg, M. A. (Eds.), Attitudes, behavior and social context. The role of norms and group membership (pp. 95–113). London, UK: Lawrence Erlbaum Associates.Google Scholar
Monteith, M. J, Deneen, N. E., & Tooman, G. D. (1996). The effect of social norm activation on the expression of opinions concerning Gay men and Blacks. Basic and Applied Social Psychology, 18, 267–288. http://dx.doi.org/10.1207/s15324834basp1803_2 Google Scholar
Moscovici, S. (1976). La psychanalyse, son image et son public [Psychoanalysis: Its image and its public] . Paris, France: Presses Universitaires de France.Google Scholar
Mouret, M., Lo Monaco, G., Urdapilleta, I., & Parr, W. (2013). Social representations of wine and culture: A comparison between France and New Zealand. Food Quality and Preference, 30, 102–107. http://dx.doi.org/10.1016/j.foodqual.2013.04.014 Google Scholar
Pérez, J. A., Moscovici, S., & Chulvi, B. (2007). The taboo against group contact: Hypothesis of Gypsy ontologization. British Journal of Social Psychology, 46, 249–272. http://dx.doi.org/10.1348/014466606X111301 CrossRefGoogle ScholarPubMed
Piermattéo, A., Lo Monaco, G., Guimelli, C., & Brel, P. (2012). Représentations sociales et applications dans le champ du marketing du vin [Social representations and applications in the field of wine marketing]. Psihologia Sociala, 29, 53–70.Google Scholar
Plant, E. A., & Devine, P. G. (1998). Internal and external motivation to respond without prejudice. Journal of Personality and Social Psychology, 75, 811–832. http://dx.doi.org/10.1037//0022-3514.75.3.811 CrossRefGoogle Scholar
Prentice, D. A., & Miller, D. T. (1993). Pluralistic ignorance and alcohol use on campus: Some consequences of misperceiving the social norm. Journal of Personality and Social Psychology, 64, 243–256. http://dx.doi.org/10.1037//0022-3514.64.2.243 Google Scholar
Reinert, M. (1993). Les “mondes lexicaux” et leur “logique” à travers l'analyse statistique d'un corpus de récits de cauchemars [The “lexical worlds” and their “logic” through the statistical analysis of a body of nightmare stories]. Langage et société, 66, 5–39.Google Scholar
Rosenthal, R., Rosnow, R. L., & Rubin, D. B. (2000). Contrasts and effect sizes in behavioral research: A correlational approach. Cambridge, UK: Cambridge University Press.Google Scholar
Schuman, H., & Converse, J. M. (1971). The effect of black and white interviewers on Black responses. Public Opinion Quarterly, 35, 44–68. http://dx.doi.org/10.1086/267866 CrossRefGoogle Scholar
Sechrist, G. B., Swim, J. K., & Stangor, C. (2004). When do the stigmatized make attributions to discrimination occurring to the self and others? The roles of self-presentation and need for control. Journal of Personality and Social Psychology, 87, 111–122. http://dx.doi.org/10.1037/0022-3514.87.1.111 Google Scholar
Stangor, C., Swim, J. K., Van allen, K. L., & Sechrist, G. B. (2002). Reporting discrimination in public and private contexts. Journal of Personality and Social Psychology, 82, 69–74. http://dx.doi.org/10.1037//0022-3514.82.1.69 Google Scholar
Sudman, S., & Bradburn, N. M. (1974). Response effects in surveys: A review and synthesis. Chicago, IL: Aldine.Google Scholar
Tavani, J. L. (2012). Mémoire sociale et pensée sociale. Etudes empiriques de leurs influences croisées [Social memory and social thinking. Empirical studies of their crossed influences] . Paris, France: Thèse de Doctorat de l’Université Paris Descartes.Google Scholar
Wagner, W., Valencia, J., & Elejabarrieta, F. (1996). Relevance, discourse and the “hot” stable core of social representation. A structural analysis of word associations. British Journal of Social Psychology, 35, 331–351. http://dx.doi.org/10.1111/j.2044-8309.1996.tb01101.x Google Scholar
Figure 0

Table 1.General results in terms of frequency and average importance associated with the categories of word reported by the participants

Figure 1

Table 2.Frequencies and average importance associated with the word “thief” according to the experimental condition

Figure 2

Table 3.Themes and words associated with the two classes composing the sentences related to the word association task

Figure 3

Figure 1.Graphical representation of the results obtained by means of the CORR. F. A. concerning factors 1 and 2.Note: Grayed blocks refer to the experimental conditions. “Experimental conditions” contribute to the formation of Factor 1; “Experimental conditions” refer to the experimental conditions which contribute to the formation of Factor 2; “Experimental condition” refer to the experimental conditions which contribute to the formation of both Factors 1 and 2. “Observations” refer to the observations which contribute to the formation of Factor 1; “Observations” refer to the observations which contribute to the formation of Factor 2; “Observations” refer to the observations which contribute to the formation of both Factors 1 and 2.

Figure 4

Table 4.Average importance and standard deviation related to attitudes regarding associated words according to experimental conditions