Situating moral disengagement: Motivated reasoning ...

Personality and Individual Differences 90 (2016) 353–364

Contents lists available at ScienceDirect

Personality and Individual Differences journal homepage: www.elsevier.com/locate/paid

Situating moral disengagement: Motivated reasoning in meat consumption and substitution João Graça a,⁎, Maria Manuela Calheiros a, Abílio Oliveira b a b

Instituto Universitário de Lisboa (ISCTE-IUL), CIS-IUL, Lisboa, Portugal Instituto Universitário de Lisboa (ISCTE-IUL), ISTAR-IUL, Lisboa, Portugal

a r t i c l e

i n f o

Article history: Received 13 July 2015 Received in revised form 24 September 2015 Accepted 18 November 2015 Available online xxxx Keywords: Moral disengagement Moral self-regulation Motivated moral reasoning Meat consumption Meat substitution Meat attachment

a b s t r a c t This work advances towards an increased understanding of moral disengagement and individual differences in the maintenance of widespread and cherished harmful behaviors. Drawing on meat consumption and substitution as an opportunity to study the process of moral self-regulation in situ, it presents a measure of selective deactivation of moral self-regulatory processes when considering the impact of meat consumption (i.e. the Moral Disengagement in Meat Questionnaire — MDMQ). The MDMQ developed from four sequential studies following a mixed-methods approach. Two preliminary studies (40 and 410 participants, respectively) provided input to develop the construct and initial pool of items. Two additional studies (1016 and 318 participants, respectively) allowed the assessment of item selection, factor structure, reliability, convergent and concurrent validities, predictive ability, and measurement invariance. The MDMQ was associated with a variety of individual differences concerning moral self-regulation (i.e. propensity to morally disengage; moral identity; empathy; moral emotions) and endorsement of dominance ideologies (i.e. social dominance orientation; speciesism; human supremacy beliefs). In a sequential mediation model, frequency of meat consumption affected willingness towards meat substitution indirectly via meat attachment and moral disengagement. We offer an interpretation of moral disengagement as a motivated reasoning process which is triggered by loss aversion and dissonance avoidance. © 2015 Elsevier Ltd. All rights reserved.

1. Introduction Why do good people do bad things? The theory of moral disengagement provides some insight into this timeless question. This theory explains the exercise of moral agency through self-regulatory processes (Bandura, 1991; Bandura, Barbaranelli, Caprara, & Pastorelli, 1996). Put briefly, it proposes that the process of moral self-regulation can be selectively deactivated in order to reduce dissonance, in light of the consideration of the damage associated with one's own conduct. This allows engaging in self-serving detrimental behaviors without incurring selfevaluative emotional reactions, such as guilt (Bandura, 1999, 2007). In particular, this theory provides a coherent framework tying together the array of different cognitive mechanisms that serve to deactivate moral self-regulation, and putting them under a single overarching construct of moral disengagement (Bandura et al., 1996; Detert, Treviño, & Sweitzer, 2008). Specifically, it describes how a set of eight cognitive mechanisms falling in four broader categories operate together to allow one to support or perpetrate harmful acts while maintaining a positive selfimage (Bandura, 1999; Bandura et al., 1996). The first category refers ⁎ Corresponding author. E-mail addresses: [email protected], [email protected] (J. Graça).

http://dx.doi.org/10.1016/j.paid.2015.11.042 0191-8869/© 2015 Elsevier Ltd. All rights reserved.

to the cognitive reconstrual of the conduct so it is not viewed as immoral in spite of the harm it entails. It includes moral justification (i.e. making detrimental behavior acceptable by portraying it as means to serve higher ends), euphemistic labeling (i.e. using euphemistic language so that harmful activities are downplayed and seemingly in agreement with moral standards) and advantageous comparison (i.e., comparing one's harmful conduct to worse inhumanities). The second category consists in obscuring personal responsibility to minimize one's role in causing harm. It includes diffusion (i.e. projecting responsibility into the larger group) and displacement of responsibility (i.e. framing behaviors as the result of situational pressures or other people's demands). The third category consists in misrepresenting injurious consequences that flow from one's behavior (i.e. through selective inattention, avoidance and dissociation, or minimization). Finally, the fourth category focuses on the recipients of detrimental conduct and includes dehumanization (i.e. viewing the recipients as unable to experience feelings and unworthy of moral consideration) and victim blaming (i.e. viewing victims as accountable for the harm). 1.1. Moral disengagement: stable trait vs. situated process Empirical work to date has generally supported moral disengagement theory, which has seen an increase of research interest in recent

354

J. Graça et al. / Personality and Individual Differences 90 (2016) 353–364

years (Page & Pina, 2015; Reynolds, Dang, Yam, & Leavitt, 2014). For instance, it has been found that scores of moral disengagement associate with a wide variety of aggressive and unethical behaviors in different contexts, such as counterproductive work behavior (Fida et al., 2014), workplace harassment (Claybourn, 2011), bullying (Obermann, 2011), acceptance of violence against animals (Vollum, Buffington-Vollum, & Longmire, 2004), support of military force (Aquino, Reed, Thau, & Freeman, 2007), violations of civic duties (Caprara, Fida, Vecchione, Tramontano, & Barbaranelli, 2009), unethical consumer attitudes (Egan, Hughes, & Palmer, 2015), and aggressiveness towards others (Paciello, Fida, Tramontano, Lupinetti, & Caprara, 2008). However, there have been concerns with one critical issue that clouds its contributions: moral disengagement is originally theorized as a situated cognitive process, but its measure and empirical operationalization has so far focused on an individual's propensity to morally disengage from a set of different forms of detrimental conduct, reporting at the same time to diverse contexts and interpersonal relationships. In other words, there is a sharp disconnection between theorizing a situated cognitive process while measuring a stable trait which seems to be more related to a general lack of concern for others, rather than an indicator of actual disengagement processes (Reynolds et al., 2014). In spite of the strengths of the moral disengagement framework (i.e. putting several cognitive mechanisms that deactivate moral selfregulation under an overarching construct of moral disengagement), this issue limits the theory's contributions in at least three critical ways. First, it ends up suffering from one of the main limitations it originally meant to address, which was to move beyond the study of morality in terms of abstract principles measured under decontextualized or fabricated circumstances. Second, as its measure refers to a general trait of propensity to morally disengage, it does not actually anchor the disengagement process in specific disengagement mechanisms with reference to a given behavior, making it impossible to test the core of the theory. Third, the central but vague claim that these mechanisms operate in everyday situations in which people routinely perform self-serving activities at injurious costs to others remains accepted as a truism, without ever having been clarified. To overcome these limitations, we suggest that moral disengagement theory may benefit with research which is grounded in its main conceptual propositions. To achieve this aim, opportunities to measure and observe moral disengagement as a situated process are needed. 1.2. An opportunity to study moral disengagement processes in situ Food practices are increasingly seen as providing a novel perspective from which to observe individuals' basic psychological processes associated with everyday moral action (e.g., Bastian, Loughnan, Haslam, & Radke, 2012; Bratanova et al., 2015). Meat consumption in particular can be framed as a morally significant behavior and conceptualized as a moral choice (Bastian et al., 2012). For instance, overlapping the framings of meat as food and meat as animal seems to evoke dissonance in the moral domain (see the “meat paradox”; Loughnan, Bastian, & Haslam, 2014), and it has been observed that a set of rationales resembling moral disengagement mechanisms (e.g., justifications; selfexonerations) arise when some consumers contemplate the consequences of meat production and consumption, and the possibility of changing habits (Graça, Calheiros, & Oliveira, 2014). People who choose to reduce or avoid meat consumption often portray their decision as a moral choice (Ruby, 2012), and previous studies suggest that even meat eaters see vegetarianism as a morally admirable stance, although sometimes respond defensively to the presence of vegetarians (Rothgerber, 2014; Ruby & Heine, 2011). According to moral disengagement theory, individuals will be particularly driven to employ disengagement mechanisms when adopting or maintaining harmful behaviors that are valued and desired (i.e., self-serving) (Bandura, 1999). Many studies reinforce the notion that meat occupies a central role in conventional western diets, and

tends to be invested with a higher status than other food items (Fiddes, 1991; Schösler, de Boer, & Boersema, 2012). However, meat's central place in the menu is being increasingly challenged, as a global shift towards reduced meat consumption and a more plant-based diet is endorsed as a means to promote environmental sustainability, improve public health, and minimize animal suffering (Pluhar, 2010; Tilman & Clark, 2014; van Dooren, Marinussen, Blonk, Aiking, & Vellinga, 2014; Westhoek et al., 2014). Assuming that individuals will be particularly motivated to use disengagement mechanisms when adopting or maintaining harmful but cherished practices, moral disengagement may indeed play a role among some consumers when prompted to consider the impact of their eating habits, and the possibility of change. Indeed, findings from recent studies on meat consumption and substitution fit some of the original propositions advanced in moral disengagement theory. For instance, it has been found that some consumers are prone to engage in rationalizations to justify meat consumption that have a guilt-alleviating function (Piazza et al., 2015), which may be seen as a form of cognitive reconstrual of the conduct. It has also been observed that many consumers tend to dissociate meat from its origin and find it difficult to eat animal products if the consumed animal closely resembles the live animal, which parallels avoidance and dissociation of the harmful consequences (Plous, 2003). Categorization as food has been found to reduce animals' perceived capacity to suffer and restrict moral concern for animals (Bratanova, Loughnan, & Bastian, 2011), which echoes viewing the recipients as unable to experience feelings and unworthy of moral consideration. Denying farm animals certain psychological characteristics has actually been identified as form of moral disengagement among meat consumers (Bilewicz, Imhoff, & Drogosz, 2011), and previous exploratory findings suggest that consumers holding a pattern of meat attachment (i.e. a positive bond towards meat consumption which comprises hedonism, affinity, entitlement and dependence) are especially prone to morally disengage when considering the impacts of meat (Graça, Oliveira, & Calheiros, 2015). In sum, taken together, these findings indicate that the study of meat consumption and substitution may provide an opportunity for observing moral disengagement processes in situ. 1.3. Overview of aims and studies The present work draws on the potential of food practices for observing psychological processes concerning everyday moral action. Specifically, it aims to expand knowledge on the process of moral disengagement as a situated cognitive process with reference to a given behavior. Likewise, it aims to provide insights about how moral disengagement operates in everyday situations in which people routinely perform self-serving activities at injurious costs to others. As a first step in this direction, we present the development and validation of an instrument measuring the selective deactivation of moral selfregulatory processes when considering the impact of meat consumption (i.e. the Moral Disengagement in Meat Questionnaire — MDMQ). The MDMQ was developed and validated in two sequential phases following a mixed-methods approach. In the first phase, qualitative data from two preliminary studies provided input for developing the construct and initial pool of items (Graça et al., 2014; Graça, Oliveira, et al., 2015). In the second phase, two further studies provided input for validating the questionnaire and examine moral disengagement as a situated cognitive process. The current article presents a brief overview of the process of developing the construct and initial pool of items (i.e. phase one), and a more thorough report on the two studies measuring moral disengagement as a situated process (i.e. phase two). 1.4. Phase one: construct and item preliminary development To generate items for measuring moral disengagement as a situated cognitive process, we drew on qualitative data from two previous studies providing an in-depth approach to meat consumption and

J. Graça et al. / Personality and Individual Differences 90 (2016) 353–364

substitution. The first study consisted in six semi-structured focus groups in which 40 participants discussed their views about meat consumption and changing eating habits (Graça et al., 2014). Data retrieved were analyzed using thematic analysis based on moral disengagement theory (additional information about the sample and procedures is available in Supplementary Materials). The other study included a set of open-ended questions in which 410 participants shared their perceptions on the impact of meat and willingness to change eating habits (Graça, Oliveira, et al., 2015). Data retrieved were coded with thematic analysis using semantic criteria, and subjected to Multiple Correspondence Analysis to detect and represent underlying structures in the dataset (see the Supplementary Materials for additional information). In both studies, several patterns of response resembling moral disengagement mechanisms were observed among some participants (i.e., cognitive reconstrual; denial/disregard for negative consequences; diffused responsibility; withdrawal of moral consideration) (cf. Graça et al., 2014; Graça, Oliveira, et al., 2015). As input for stage two, for the purposes of the new measure, we broadly defined moral disengagement in meat as the selective deactivation of moral self-regulatory processes when considering the impact of meat consumption. Afterwards, the process of item generation followed several steps, based on two criteria: continuous crosschecking with the moral disengagement core propositions; and the salience and semantic significance of the response patterns in both previous studies. The first step was to identify the passages and phrases in the qualitative data from both studies that best captured the moral disengagement construct in the context of everyday moral action. In the second step, we selected the most vivid and representative excerpts to develop candidate items (i.e. generating a statement and/or alternative versions of a statement). We strove to avoid negations and use simple language in the statements. To favor parsimony and reduce blatant redundancy, we then followed an iterative process by deleting/combining statements, which resulted in an initial pool of 49 items for exploratory analyses. This process ensured that the preliminary item pool exhibited a broad pattern of responses (i.e., cognitive reconstrual; denial/disregard for negative consequences; diffused responsibility; withdrawal of moral consideration) which also had a good theoretical grounding. However, we did not advance any specific tentative structure a priori because psychometric criteria derived from the data (i.e. item selection and exploratory factor analysis) would determine the best-fit structure. 1.5. Phase two: measuring moral disengagement as a situated cognitive process Building on the preceding stage of item generation, the second phase developed the questionnaire and examined moral disengagement as a situated cognitive process. It comprised two studies which are presented in the current article. Specifically, in study one, 1016 participants answered the initial pool of items and other measures. These data provided information on item selection, factor structure (principal axis factoring and confirmatory factor analysis), reliability (Cronbach's alpha), and several types of validity: convergent, concurrent, and predictive ability. To assess convergent validity, we drew on previous relevant research showing associations between moral disengagement and other constructs relevant in moral regulation. For instance, it is known that high moral disengagers tend to be less troubled by feelings of guilt caused by injurious conduct (Bandura et al., 1996) and hold lower levels of empathy and moral identity internalization (Detert et al., 2008). Dominance ideologies also appear to play their part: high moral disengagers tend to show greater endorsement of power and status inequalities, and tend to believe that they are superior to others (Jackson & Gaertner, 2010; Rosenblatt, 2012). Thus, for convergent validity we measured the interrelationships between the MDMQ and propensity to morally disengage (Reynolds et al., 2014), moral identity internalization (Aquino & Reed, 2002), empathy (Goldberg, 2001), moral emotions when considering the

355

impact of meat (i.e. guilt and shame; Aquino et al., 2007), social dominance orientation (Pratto, Sidanius, Stallworth, & Malle, 1994), speciesism (Dhont, Hodson, Costello, & MacInnis, 2014) and belief in human supremacy (Dhont & Hodson, 2014). As regards concurrent validity, we considered evidence from studies on meat consumption and substitution that highlight the role of habits (Berndsen & van der Pligt, 2004), dietary identity (Fox & Ward, 2008) and gender (Rothgerber, 2013) as determinants of dietary behavior. Gender was relevant for the present study also because men tend to be higher moral disengagers (Bandura et al., 1996; McAlister, Bandura, & Owen, 2006). Finally, for predictive ability, we measured the MDMQ's additional explanatory power above and beyond the effects of attitudes towards meat and current consumption habits (Berndsen & van der Pligt, 2004) in willingness to reduce meat consumption. In short, the nomological network assessed in study one articulates an integrated set of measures from the fields of moral regulation, dominance ideologies, and our focal behavior (i.e. meat consumption and substitution) (Fig. 1). In study two, a new sample of 318 participants from a different cultural background completed the final version of the MDMQ along with other measures. These data allowed for obtaining preliminary evidence on the MDMQ's measurement invariance (confirmatory factor analysis), and providing a first examination of the role of moral disengagement as a situated cognitive process in the maintenance of a harmful but cherished behavior. Specifically, we tested a sequential mediation model with moral disengagement as a chain mediator, along with meat attachment, in the relationship between meat eating habits and willingness towards meat substitution (Fig. 2). We drew on previous exploratory evidence suggesting that meat attachment hinders willingness to change eating habits, and that more attached consumers are especially prone to morally disengage when considering the impact of meat (Graça, Oliveira, et al., 2015).

2. Study one: questionnaire assessment and validation 2.1. Participants and procedure The survey was hosted online by Qualtrics.com and participants were recruited via ads in social media. A short recruitment notice invited Portuguese users to participate in a study “exploring people's opinions on several issues related with society and different social practices, lifestyles and eating habits”. Participation was rewarded with the choice of registering in a draw to win a 7.9″ 16 GB tablet. To minimize self-selection biases, no references were made in the advertisement and cover page to the specific goals of the study. Before beginning the survey, participants were informed about the study's procedures and assured that no individual answers would be analyzed or reported. They were also informed that participation was entirely anonymous and voluntary. Participants provided their consent and were debriefed after filling the survey. The survey was available for nearly four months between July 3rd and November 5th 2014. During this period, 1278 people clicked on the cover page to participate, and 1016 (aged between 18 and 69 years, M = 26.3, SD = 9.6; 57% women) completed all the measures under analysis. Almost all participants reported eating meat at least once in a regular week (93.2%). For the purposes of this study, participants were randomly split in two samples (Table 1) to allow for conducting two separate and sequential stages of analyses (DeVellis, 1991). Sample 1 consisted of 506 participants and was used for the exploratory factor analysis. Sample 2 consisted of 516 participants and was used for the confirmatory factor analysis and gathering of further evidence on the validity of the questionnaire.

356

J. Graça et al. / Personality and Individual Differences 90 (2016) 353–364

Fig. 1. Nomological network assessed for the Moral Disengagement in Meat Questionnaire (MDMQ). Symbols between parallel lines indicate expected valence/direction of the associations.

2.2. Measures 2.2.1. Moral disengagement in meat questionnaire The initial item pool included 49 statements (e.g., “All things considered, meat is necessary to human diet”) referring to the selective deactivation of moral self-regulatory processes when considering the impact of meat consumption, using a 5 point Likert-type scale (1 = strongly disagree to 5 = strongly agree). Before answering the items, participants were prompted to consider how current patterns of meat production and consumption may impact animals (e.g., deprivation of outdoor contexts and contact with natural living environments), nature and the environment (e.g., higher environmental costs in comparison with nutritionally equivalent plantbased foods), and public health (e.g., marked increase in heart diseases).

2.2.2. Propensity to morally disengage Propensity to morally disengage was measured with an eightitem scale borrowed from Reynolds et al. (2014), which is based on Bandura et al.’s (1996) original measure but includes only one item for each moral disengagement mechanism (e.g., “Taking personal credit for ideas that were not your own is no big deal.”). Participants provided their responses for each sentence using a 5 point Likerttype scale (1 = totally disagree to 5 = totally agree). In the current sample internal consistency was good (α = .76).

2.2.3. Moral identity Moral identity was measured with Aquino and Reed's (2002) Internalization subscale. This subscale captures the degree to which a person's moral identity is rooted at the core of one's being and has been found to be the more robust predictor (compared with the Symbolization subscale) of ethics-related attitudes and behavior (Aquino & Reed, 2002; Aquino et al., 2007). First, a set of nine adjectives was presented (e.g., caring, compassionate, fair, honest) along with the statement that these represent “some characteristics that might describe a person”. Participants then rated five items using a 5 point Likert-type scale (1 = strongly disagree to 5 = strongly agree) to assess the degree to which these characteristics represented an important part of their own identity. In the current sample internal consistency was good (α = .76).

2.2.4. Moral emotions We addressed two emotions – guilt and shame – borrowed from Aquino et al.’s (2007) measure of self-reported negative emotional reactions. We chose to measure guilt and shame because they are considered to be two key emotions in moral regulation (Tangney, Stuewig, & Mashek, 2007). Participants reported how strongly they felt each emotion when considering how meat consumption may impact: (a) animals, (b) nature and the environment, and (c) public health, using a scale ranging from 1 (not at all) to 5 (very much). Responses for the six items (i.e. two for each topic) were averaged to

Fig. 2. Self-serving behavior (i.e. frequency of meat consumption) associates with unwillingness to change the behavior (i.e. willingness towards meat substitution), via a positive valuation of the behavior (i.e. meat attachment) and the employment of moral disengagement mechanisms to defend and justify its maintenance (i.e., moral disengagement in meat).

J. Graça et al. / Personality and Individual Differences 90 (2016) 353–364

never, less than once per week, once or twice per week, three or four times per week, five times or more per week.

Table 1 Study one — sample characteristics. Variable

Gender Age

Education

Employment status

Category

Male Female b23 24–40 N40 Basic Secondary Higher Employed Unemployed Student Other

357

Sample 1

Sample 2

N

%

N

%

214 285 232 211 41 12 210 277 155 35 300 8

42.9 57.1 47.9 43.6 8.5 2.4 42.1 55.5 31.1 7 60.3 1.6

216 285 245 204 51 13 225 261 173 26 292 11

43.1 56.9 49 40.8 10.2 2.6 45.1 52.3 34.5 5.2 58.2 2.2

2.2.11. Dietary identity Participants were asked to indicate the extent in which they personally identity themselves as: (a) meat eater, (b) omnivore, (c) vegetarian, and (d) vegan, using a scale ranging from 1 (not at all) to 5 (very much) for each item. 2.2.12. Willingness to reduce meat consumption Participants reported their willingness to reduce meat consumption with a single item (“Please indicate your willingness to reduce meat consumption”), using a Likert-type scale ranging from 1 (not willing at all) to 5 (very willing). 2.3. Analyses and results

form a general measure of guilt (α = .88) and shame (α = .90) when considering the impacts of meat. 2.2.5. Empathy Empathy was measured with a 10-item scale from the International Personality Item Pool (IPIP; Goldberg, 2001). Although originally labeled as sympathy, it represents a general disposition of empathy and refers to an individual's willingness to take others' problems and emotions into consideration (Detert et al., 2008). Participants provided their responses for each item (e.g., “I am not interested in other people's problems”, reverse scored) using a 5 point Likert-type scale (1 = totally disagree to 5 = totally agree). In the current sample internal consistency was high (α = .86). 2.2.6. Social dominance orientation Social dominance orientation, a power motive of dominance and endorsement of group-based discrimination, was measured with Pratto et al.’s (1994) 16-item scale (e.g., “Some groups of people are just more worthy than others”) using a 5 point Likert-type scale (1 = totally disagree to 5 = totally agree). In the current sample internal consistency was high (α = .86). 2.2.7. Human supremacy Beliefs about human supremacy as a dominance ideology relevant to meat consumption and substitution were measured with a six-item scale (e.g., “Animals are inferior to humans”) taken from Dhont and Hodson (2014), using a 5 point Likert-type scale (1 = totally disagree to 5 = totally agree). In the current sample internal consistency was high (α = .87). 2.2.8. Speciesism Speciesism as power motive of dominance and support for hierarchy between humans and animals was measured with six items (e.g., “I think it is perfectly acceptable for cattle, chickens and pigs to be raised for human consumption”) taken from Dhont et al. (2014), using a 5 point Likert-type scale (1 = totally disagree to 5 = totally agree). In the current sample internal consistency was average (α = .67). 2.2.9. Attitudes Five semantic differential scales with 5-point each measured respondents' attitudes towards meat (Berndsen & van der Pligt, 2004). The five items were “bad–good”, “unpleasant–pleasant”, “against–for”, “unfavorable–favorable”, and “negative–positive”. In this sample internal consistency was high (α = .93). 2.2.10. Eating habits Participant's usual consumption of meat was measured with a single item borrowed from Hoek et al. (2011) using the following answering categories for the frequency of meat consumption in a regular week:

2.3.1. Sample 1: exploratory factor analysis and reliability An Exploratory Factor Analysis (EFA) was conducted with sample one (N = 506) on the original set of 49 items, using IBM SPSS Statistics for Windows (IBM Corp. Released, 2011). Principal axis factoring was used for its usefulness in identifying underlying dimensions and advantage of accounting for measurement error in the solution (Gorsuch, 1983). We expected the derived factors to be intercorrelated and used an oblique rotation (oblimin) (Abdi, 2003). The percentage of missing data was 0.01% and cases were deleted listwise. Absolute values of skewness ranged from to − 0.9 to 1, showing no problems of severe departure from a normal distribution. The Kaiser–Meyer–Olkin (KMO) measure of sampling adequacy yielded a value of .94 and Bartlett's test of sphericity was highly significant: x2(1176) = 13,915.93, p b .001. Parallel analysis with a 99% confidence interval was used to compare obtained eigenvalues with those generated from random data sets (Horn, 1965; O'Connor, 2000), which suggested that eight factors had eigenvalues greater than chance. Scree test and variance accounted for also pointed towards an initial break at eight factors. We thus initially considered a solution of eight factors explaining 60.8% of the variance for the 49 items. An iterative process was then followed to determine which items to retain (Bryman & Cramer, 2011; Matsunaga, 2010). First, items with a factor loading b.40 were excluded, and items with b.50 and cross-loadings N.25 were then dropped, until a solution in which all items retained had a factor loading N.5 and no significant cross-loadings was reached. The final solution comprised five factors for 20 items with 67.9% of the variance accounted for (see Table 2). On the basis of the rotated pattern matrix, the five factors could be described as means-ends justifications (five items; higher scores referring to pro-meat rationalizations and ‘yes, but’ constructions that portray meat consumption as means to serve higher ends; e.g., “All things considered, meat is necessary to human diet”), desensitization (four items; higher scores indicating increased emotional and cognitive desensitization to the death and suffering of animals used for food purposes; “If I had to kill the animals myself, I would probably stop eating meat”, reversed score), denial of negative consequences (five items; higher scores referring to the downplay of the negative impact associated with meat production and consumption; “By eating meat I engage with an industry responsible for major damages”, reverse scored), diffused responsibility (three items; higher scores indicating an increased tendency to project personal accountability into the larger group or society; e.g., “It doesn't matter if I change my habits because problems will still exist”), and reduced perceived choice (three items; higher scores indicating an increased tendency to frame alternatives to current patterns of meat consumption as impractical and/or inaccessible; e.g., “It's possible to have an adequate diet without eating meat”, reverse scored). Cronbach's Alpha suggested good consistency levels in these three-to-five item

358

J. Graça et al. / Personality and Individual Differences 90 (2016) 353–364

Table 2 Item and scale information from the exploratory factor analysis for the Moral Disengagement in Meat Questionnaire. Item

Factor loadings 1

Means-ends justifications 2. The human being has needs that include eating meat. 1. All things considered, meat is necessary to human diet. 4. Eating meat keeps the balance of the food chain. 33. Despite everything, eating meat is part of a balanced life. 3. The problems associated with meat also apply to other foods. Desensitization 35. It would be difficult for me to watch an animal being killed for food purposes.⁎ 12. If I saw an animal being killed I would have no problems eating it. 43. I would be capable of skinning, separating the organs and cutting an animal to pieces. 34. If I had to kill the animals myself, I would probably stop eating meat.⁎ Denial of negative consequences 9. People who eat meat should acknowledge the suffering in which food animals are kept.⁎ 18. It's important that people who eat meat think about the impacts to the environment.⁎ 13. Those who eat meat should be aware of its impacts on public health.⁎ 15. By eating meat I engage with an industry responsible for major damages.⁎ 7. By eating meat I'm also responsible for the problems associated to its production.⁎ Diffused responsibility 8. Even if I change my habits, I don't make a difference by myself. 27. It doesn't matter if I change my habits because problems will still exist. 45. I will consider changing my habits only if others also change theirs. Reduced perceived choice 21. Nowadays there are good alternatives to meat consumption.⁎ 24. It's possible to have an adequate diet without eating meat.⁎ 30. It's easy to have a meat-free diet.⁎ Eigenvalue Percentage of variance Cronbach's alpha

M

SD

h2

2

3

4

5

.85 .85 .71 .68 .68

.00 .03 .00 .02 −.01

−.08 −.03 .09 .06 .02

−.02 −.01 .05 .05 .05

.21 .13 −.04 .18 −.14

3.51 3.54 3.17 3.44 3.58

1.11 1.14 1.09 0.97 0.97

.86 .81 .54 .64 .72

−.02 .08 −.04 .01

.78 .77 .70 .70

.14 −.08 −.08 .17

−.17 .14 .14 −.16

.08 .01 −.06 .10

2.45 2.52 2.33 2.46

1.23 1.26 1.26 1.21

.72 .65 .48 .63

.07 −.07 −.08 .13 .05

.06 −.04 −.04 .07 .02

.76 .75 .69 .67 .64

−.03 .16 .05 −.12 .00

−.02 −.02 .03 .05 .03

2.62 2.39 2.38 2.86 2.52

1.04 0.86 0.91 1.05 1.01

.61 .58 .47 .55 .46

.10 .23 −.06

.00 .04 .04

.03 .07 .04

.69 .67 .61

.03 −.02 .05

2.83 2.70 2.12

1.09 1.00 0.94

.55 .64 .39

−.08 .09 .06 6.26 31.3 .90

.07 .01 −.01 2.56 12.8 .84

.06 .06 −.03 1.93 9.6 .84

.16 .02 −.06 1.69 8.5 .76

.69 .61 .60 1.15 5.7 .70

2.45 2.52 3.17

1.00 1.03 1.14

.56 .47 .37

Notes. h2 = Item communalities. Factor loadings N|.50| are presented in bold. ⁎ Reverse-scored items.

tentative subscales, which were subject to further support in the second phase of analysis. 2.3.2. Sample 2: confirmatory factor analysis and further evidence for validation 2.3.2.1. Confirmatory factor analysis. A confirmatory factor analysis (CFA) was performed in the holdout sample (N = 516) to assess the solution obtained in the EFA, using maximum likelihood method in AMOS 20 (Arbuckle, 2011). An assessment to verify the adequacy of the data for CFA was performed for the set of 20 items, again showing no problems of severe departure from a normal distribution (i.e. absolute values of skewness ranged from − 1.05 to .74). The percentage of missing data was 0.01% and cases were deleted listwise. We tested a five-factor structure with a secondorder global dimension of moral disengagement. Criteria to assess model fit were based on different indicators. The ratio x 2 /df was used to evaluate the appropriateness of the model (with good to acceptable values referring to ≤ 5), since the model chi-square test is sensitive to sample size (Schermelleh-Engel, Moosbrugger, & Muller, 2003). Comparative fit index (CFI), Tucker Lewis index (TLI) and root-mean-square error of approximation (RMSEA) were also used as model fit indices. Criteria for good to acceptable model fit were CFI ≥ .90, TLI ≥ .90, and RMSEA ≤ .08, with higher values in CFI and TLI and lower in RMSEA referring to betterquality fit indices (Hu & Bentler, 1999; Marsh, Hau, & Wen, 2004; Vandenberg & Lance, 2000). Results revealed that the model fully met criteria for adequate model fit (x 2 /df = 2.9; CFI = .93; TLI = .92; RMSEA = .06 [.06, .07]; Fig. 3). All subscales showed modest to strong correlations with each other and the global scale (Table 3). 2.3.2.2. Convergent and concurrent validity. For convergent validity, we expected that the MDMQ would show associations with measures referring to moral self-regulation and dominance ideologies. Weaker

correlations were expected with propensity to morally disengage, moral identity, empathy and social dominance orientation, since these relate to moral self-regulation but are not specifically framed in our focal behavior. Stronger correlations were expected with feelings of guilt and shame, and endorsement of speciesism and human supremacy beliefs, since these are specifically framed in the issue of meat eating and human–animal interactions. For concurrent validity, we anticipated that scores on the measures from the MDMQ would show associations with meat eating habits, gender and dietary identity. All measures from the MDMQ showed correlations with other measures and indicators relevant to the study of moral disengagement and meat consumption. Specifically, concerning measures related to moral self-regulation, the global scale and most subscales yielded modest positive associations with propensity to morally disengage, and modest to moderate negative associations with moral identity, empathy and moral emotions (Table 3). The same pattern was observed concerning the endorsement of dominance ideologies, showing modest to moderate positive associations with social dominance orientation, speciesism and belief in human supremacy (Table 4). The results concerning concurrent validity also largely yielded the anticipated pattern of associations. Men consistently scored significantly higher in almost all dimensions (Table 5), and the subscales and global scale of the MDMQ were correlated with frequency of meat consumption and dietary identity (i.e. modest to moderate relationships with personally identifying as meat eater, weaker but still generally positive associations as omnivore, and negative correlations with self-identification scores as vegetarian and as vegan (Table 6). 2.3.2.3. Predictive ability. Concerning predictive ability, we explored whether the MDMQ provided additional explanatory variance above and beyond the effects of attitudes towards meat and eating habits in willingness to reduce meat consumption. Five hierarchical regressions were performed to examine the predictive ability of the MDMQ global scale and subscales' scores using willingness to reduce meat consumption as criterion variable. For each separate regression analysis, in Step

J. Graça et al. / Personality and Individual Differences 90 (2016) 353–364

359

Fig. 3. Confirmatory factor analysis of the MDMQ: Five-factor structure with a second-order dimension. Standardized coefficients are presented. MEJ = Means-ends justifications; DES = Desensitization; DEN = Denial of negative consequences; DIFF = Diffused responsibility; REDC = Reduced perceived choice.

1 we entered the related study variables (i.e. attitudes towards meat and eating habits), and in Step 2 the MDMQ global or subscale scores. Incremental variances of MDMQ global and subscale scores in predicting willingness to reduce meat consumption above and beyond related variables were all significant (Table 7), ranging from 3% (MDMQ Diffused responsibility) to 16% (MDMQ global scale). Table 3 Subscale and global scale reliabilities, means, standard deviations, and correlations. MDMQ scale and subscales

α

M

1. Means-ends justifications 2. Desensitization 3. Denial of negative consequences 4. Diffused responsibility 5. Reduced perceived choice 6. Global scale

.88 .81 .83 .74 .72 .89

3.50 .75 – 2.44 1.04 .27⁎ – 2.58 .72 .34⁎ .39⁎ – 2.58 .81 .42⁎ .19⁎ .34⁎ – 2.73 .84 .51⁎ .32⁎ .41⁎ .32⁎ – 2.81 .57 .74⁎ .62⁎ .76⁎ .61⁎ .70⁎

⁎ p b .001.

SD

1

2

3

4

5

No problems of multicollinearity were detected in the analyses (VIF range: 1.07 to 1.45; tolerance range: .69 to .93). We also tested the five MDMQ subscales in a single step to assess if they all predicted willingness to reduce meat consumption when controlling for each other (see Table B2 in Supplementary Materials). Most subscales showed equivalent predictive capacity (β ranging from −.17 to −.23), with the exception of Diffused Responsibility, which showed lower but still significant values (β = −.12). 3. Study two: moral disengagement as a situated cognitive process 3.1. Participants and procedure For the second study participants were recruited through Amazon Mechanical Turk (MTurk-http://www.mturk.com/mturk/). A short recruitment notice advertised to U.S. based participants presented the study as exploring “people's opinions about food and different eating

360

J. Graça et al. / Personality and Individual Differences 90 (2016) 353–364

Table 4 Correlations with measures referring to moral self-regulation and dominance ideologies. Moral self-regulation

Dominance ideologies

MDMQ scale and subscales

Propensity to morally disengage

Moral identity (internalization)

Empathy

Moral emotions (guilt)

Moral emotions (shame)

SDO

Speciesism

Human supremacy

1. Means-ends justifications 2. Desensitization 3. Denial of negative consequences 4. Diffused responsibility 5. Reduced perceived choice 6. Global scale

.02 .15⁎⁎ .10* .33⁎⁎

.01 −.25⁎⁎ −.26⁎⁎ −.22⁎⁎

.09 .18⁎⁎

−.08 −.22⁎⁎

−.06 −.29⁎⁎ −.19⁎⁎ −.13⁎⁎ −.16⁎⁎ −.23⁎⁎

−.39⁎⁎ −.48⁎⁎ −.51⁎⁎ −.23⁎⁎ −.35⁎⁎ −.58⁎⁎

−.39⁎⁎ −.41⁎⁎ −.42⁎⁎ −.21⁎⁎ −.31⁎⁎ −.52⁎⁎

.10⁎ .20⁎⁎ .18⁎⁎ .30⁎⁎ .18⁎⁎ .26⁎⁎

.40⁎⁎ .55⁎⁎ .49⁎⁎ .35⁎⁎ .36⁎⁎ .62⁎⁎

.29⁎⁎ .43⁎⁎ .41⁎⁎ .34⁎⁎ .28⁎⁎ .51⁎⁎

⁎ p b .05. ⁎⁎ p b .01.

habits”, along with a link to the Qualtrics website hosting the survey. MTurk was chosen in light of evidence that participants tend to be more demographically diverse than standard internet samples, realistic compensation rates do not affect data quality, and the data obtained are at least as reliable as those obtained via traditional methods (Buhrmester, Kwang, & Gosling, 2011). Participants were offered $.75 for their participation. Before beginning the survey, participants were informed about the study's procedures and assured that no individual answers would be analyzed or reported. They were also informed that participation was entirely anonymous and voluntary, and provided their consent as participants. The survey was accessible in March 2nd 2015. Three hundred and eighteen persons (aged between 18 and 72 years, M = 36.3, SD = 11.2) participated in the study. One hundred and eighty five were male (58.2%) and 133 were female (41.8%). Most participants had completed higher education (204; 64.4%), followed by secondary (89; 28.1%) and primary (24; 7.6%). As regards their employment status, around two thirds were employed (227; 71.4%), 37 were unemployed (11.6%), 23 were students (7.2%) and 31 were retired or held a different status (9.1%).

3.2. Measures 3.2.1. Moral disengagement in meat questionnaire Moral disengagement in meat was measured with the 20-item final version of the MDMQ that was developed and validated in the first study. Participants read a short introductory sentence (“In recent times, meat consumption is being increasingly debated on the grounds of environmental sustainability, health and safety concerns, and animal rights/welfare arguments. Please tell us your personal opinion about meat consumption indicating the extent to which you agree with the following sentences”) and then provided their responses for each sentence using a 5 point Likert-type scale (1 = totally disagree to 5 = totally agree). We conducted a CFA with AMOS 20 using maximum likelihood as the estimation method to assess measurement invariance. The model comprised of five factors with a second order global dimension of moral disengagement fully met criteria for adequate fit to the Table 5 Mean differences between men (N = 216) and women (N = 285) on the MDMQ scale and subscales. MDMQ scale and subscales

1. Means-ends justifications 2. Desensitization 3. Denial of negative consequences 4. Diffused responsibility 5. Reduced perceived choice 6. Global scale ⁎ p b .05. ⁎⁎ p b .01. ⁎⁎⁎ p b .001.

Men

Women

M

SD

M

SD

3.49 2.89 2.65 2.71 2.85 2.96

.76 .97 .69 .78 .83 .56

3.50 2.01 2.52 2.47 2.64 2.72

.76 .95 .75 .82 .84 .56

F(1499)

Cohen's d

.01 86.46⁎⁎⁎ 4.09⁎ 10.95⁎⁎ 7.61⁎⁎ 21.38⁎⁎⁎

.01 .92 .18 .30 .25 .43

data (x2/df = 2.8; TLI = .91; CFI = .92; RMSEA = .08 [.07, .08]). In the current sample internal consistency was high (α = .90). 3.2.2. Meat attachment questionnaire Meat attachment was measured with a 16-item questionnaire (e.g., “If I was forced to stop eating meat I would feel sad”) referring to a positive bond towards meat consumption (Graça, Calheiros, et al., 2015). Participants provided their responses for each sentence using a 5 point Likert-type scale (1 = totally disagree to 5 = totally agree). In the current sample internal consistency was high (α = .95). 3.2.3. Eating habits Participant's usual consumption of red and white meat was measured using the following answer categories for the frequency of meat consumption in a regular week: never, less than once per week, once or twice per week, three or four times per week, five times or more per week. An aggregate measure of meat eating habits was computed by averaging participants' responses on red and white meat consumption. 3.2.4. Willingness towards meat substitution Participants reported their willingness to (i) reduce meat consumption, (ii) avoid eating meat, and (iii) follow a plant-based diet, using a 5-point Likert-type scale (ranging from 1 — very unwilling to 5 — very willing and 1 — surely not to 5 — surely yes, respectively). Responses for the three items (i.e. one for each focal behavior) were averaged to form a general measure of willingness towards meat substitution (α = .90). 3.3. Analyses and results To explore the role of moral disengagement as a situated cognitive process in the maintenance of harmful but cherished behaviors, we tested a three step sequential mediation model in IBM SPSS Statistics for Windows (IBM Corp., 2010) using the PROCESS macro by Hayes (2012, 2013). Specifically, willingness towards meat substitution was regressed on frequency of meat consumption via a chain of two sequential mediations (i.e. meat attachment and moral disengagement in meat). No missing data were observed. Descriptive statistics and bivariate correlations are presented in Table 8. In spite of the high intercorrelations, a preliminary regression analysis with willingness as criterion variable and the other three as predictors revealed no problems of multicollinearity (VIF range: 1.54 to 3.78; tolerance range: .26 to .65). Path coefficients for the sequential mediation model were estimated by using ordinary least squares regressions, and 95% confidence intervals were determined by bootstrapping (based on 10,000 draws). There were positive significant associations from meat consumption to meat attachment (β = .47, SE = .04, p b .001) and from meat attachment to moral disengagement (β = .62, SE = .03, p b .001). In turn, there were negative significant associations from meat attachment to meat substitution (β = −.75, SE = .09, p b .001), and from moral disengagement to meat substitution (β = −.48, SE = .11, p b .001). In this

J. Graça et al. / Personality and Individual Differences 90 (2016) 353–364

indirectly via meat attachment and moral disengagement (as illustrated in Fig. 2).

Table 6 Correlations with eating habits and dietary identity. MDMQ scale and subscales

Eating habits

Dietary identity Meat eater

Omnivore Vegetarian Vegan

1. Means-ends justifications 2. Desensitization 3. Denial of negative consequences 4. Diffused responsibility 5. Reduced perceived choice 6. Global scale

.48⁎⁎

.50⁎⁎

.33⁎⁎

−.36⁎⁎

−.42⁎⁎

.27⁎⁎ .28⁎⁎

.24⁎⁎ .27⁎⁎

.20⁎⁎ .09⁎

−.20⁎⁎ −.27⁎⁎

−.19⁎⁎ −.27⁎⁎

.34⁎⁎ .43⁎⁎

.32⁎⁎ .43⁎⁎

.10⁎ .23⁎⁎

−.26⁎⁎ −.39⁎⁎

−.19⁎⁎ −.33⁎⁎

.51⁎⁎

.50⁎⁎

.28⁎⁎

−.42⁎⁎

−.41⁎⁎

⁎ p b .05. ⁎⁎ p b .001.

model, the path between meat consumption and moral disengagement was not significant (β = .03, SE = .02, p = .15). The indirect effect of meat consumption via meat attachment in meat substitution was significant, β = −.35, SE = .06, 95% CI [−.47, −.25]; as was the indirect effect of meat consumption via meat attachment and moral disengagement in meat substitution, β = −.14, SE = .04, 95% CI [−.21, −.07]. The indirect effect of meat consumption in meat substitution via moral disengagement was not significant, β = −.02, SE = .01, 95% CI [−.04, .00]). Finally, the direct effect of meat consumption in meat substitution (β = −.52, SE = .05, p b .001) was no longer significant (β = −.01, SE = .05, p = .83), indicating full mediation by the model. In short, in a sequential mediation model, frequency of meat consumption affected willingness towards meat substitution

Table 7 Hierarchical regressions for incremental validity in predicting willingness to reduce meat consumption above and beyond attitudes and eating habits. Variable Step 1 Attitudes Eating habits Step 2 — means-ends justifications (MDMQ MEJ) Attitudes Eating habits MDMQ MEJ Step 2 — desensitization (MDMQ DES) Attitudes Eating habits MDMQ DES Step 2 — denial of negative consequences (MDMQ DEN) Attitudes Eating habits MDMQ DEN Step 2 — diffused responsibility (MDMQ DIF) Attitudes Eating habits MDMQ DIF Step 2 — reduced perceived choice (MDMQ RED) Attitudes Eating habits MDMQ RED Step 2 — MDMQ Global scale Attitudes Eating habits MDMQ Global scale ⁎ p b .05. ⁎⁎ p b .01. ⁎⁎⁎ p b .001.

B

SE

361

β

−.44 .05 −.39⁎⁎⁎ −.19 .06 −.15⁎⁎⁎ −.35 .05 −.31⁎⁎⁎ −.17 .06 −.13⁎⁎ −.36 .07 −.23⁎⁎⁎ −.36 .05 −.32⁎⁎⁎ −.16 .06 −.13⁎⁎ −.28 .04 −.28⁎⁎⁎ −.35 .05 −.31⁎⁎⁎ −.19 .06 −.14⁎⁎ −.45 .06 −.30⁎⁎⁎ −.39 .05 −.35⁎⁎⁎ −.17 .06 −.13⁎⁎ −.24 .05 −.19⁎⁎⁎ −.37 .05 −.33⁎⁎⁎ −.12 .06 −.09⁎ −.37 .05 −.29⁎⁎⁎ −.22 .05 −.20⁎⁎⁎ −.13 .05 −.10⁎ −.93 .08 −.46⁎⁎⁎

ΔR2

ΔF

dfs

.23⁎⁎⁎

71.62 2487

.05⁎⁎⁎

60.57 1486

.07⁎⁎⁎

68.97 1486

.08⁎⁎⁎

73.24 1486

.03⁎⁎⁎

56.79 1486

.07⁎⁎⁎

69.22 1486

4. Discussion To our knowledge, this work is the first to measure and investigate moral disengagement as a situated cognitive process occurring with reference to a particular harmful but cherished and widespread practice. It offers a first step towards an increased understanding of how moral disengagement mechanisms operate in situ to deactivate moral selfregulation in self-serving behaviors that have detrimental consequences for others and the environment. Overall, the findings suggest that the proposed measure of moral disengagement is a valid and reliable way to assess the selective deactivation of moral self-regulatory processes when considering the impact of meat consumption. One key contribution this measure provides is that it addresses a critical limitation that so far has clouded moral disengagement's explanatory and predictive capacity. In particular, a sharp disconnection has been noted between theorizing a situated cognitive process while measuring a stable trait which arguably refers to a general lack of concern for others (Reynolds et al., 2014). Thus, this study provides a means to test the core of the theory by anchoring the disengagement process in specific mechanisms with reference to a given behavior. In the current work, exploratory and confirmatory factor analysis revealed five dimensions within the construct, namely means-ends justifications (i.e. promeat rationalizations and ‘yes, but’ constructions that portray meat consumption as means to serve higher ends), desensitization (i.e. becoming emotionally and cognitively desensitized to the death and suffering of animals used for food purposes), denial of negative consequences (i.e. downplay the negative impact associated with meat production and consumption), diffused responsibility (i.e. projecting personal accountability into the larger group or society as a whole), and reduced perceived choice (i.e. frame alternatives to current patterns of meat consumption as inaccessible and impractical). Conceptually, these dimensions resonate with the theory's original core propositions regarding the various paths through which different disengagement mechanisms operate (Bandura, 1999; Bandura et al., 1996): meansends justifications may be seen as referring to the cognitive reconstrual of the conduct, so it is not viewed as immoral in spite of the harm it entails; desensitization suggests a disregard for the victims' suffering and withdrawal of moral consideration; denial of negative consequences refers to misrepresenting injurious costs that flow from one's behavior; diffused responsibility and reduced perceived choice may be seen as self-exonerations that obscure personal agency in causing harm. All dimensions showed modest to moderate correlations with each other and strong correlations with a global measure of moral disengagement. Across samples, a five-factor model with 20 items and a second-order global dimension of moral disengagement fully met criteria for model fit. As for the nomological network of interrelationships observed between our measure of moral disengagement and other relevant constructs (Fig. 1), the anticipated pattern of associations was largely found with measures referring to moral self-regulation, such as propensity to morally disengage (Reynolds et al., 2014), moral identity (Aquino & Reed, 2002), empathy (Goldberg, 2001), and moral emotions (i.e. guilt and shame; Aquino et al., 2007); and measures referring to dominance ideologies, such as social dominance orientation (Pratto et al.,

Table 8 Means, standard-deviations and correlations of variables for study two. .16⁎⁎⁎ 102.94 1486 1. Moral disengagement 2. Meat attachment 3. Meat consumption 4. Meat substitution ⁎ p b .001.

M

SD

1

2

3

2.96 3.62 4.39 2.76

.73 .94 1.19 1.29

– .84⁎ .53⁎ .74⁎

– .59⁎ .79⁎

– .49⁎

362

J. Graça et al. / Personality and Individual Differences 90 (2016) 353–364

1994), speciesism (Dhont et al., 2014) and human supremacy beliefs (Dhont & Hodson, 2014). These findings indicated convergent validity. In turn, associations with eating habits (Berndsen & van der Pligt, 2004), dietary identity (Fox & Ward, 2008) and gender (Bandura et al., 1996; Rothgerber, 2013) indicated concurrent validity. Specifically, individuals who showed higher levels of moral disengagement with reference to meat consumption and substitution also tended to display higher general propensity to morally disengage and to attribute less importance to moral traits in how they view themselves. These individuals also tended to show lower levels of empathy, experience less selfevaluative emotional reactions (i.e. guilt and shame) when considering the impact of meat consumption, endorse group-based discrimination within humans, and display power motives of dominance and support of hierarchy of humans over other species. They were also more likely to eat meat more often and identify more strongly as meat eaters and omnivores, and less as vegetarians or vegans. Men tended to score higher in our measure of moral disengagement than women. As for predictive ability, the MDMQ provided additional explanatory power above and beyond the effects of attitudes towards meat and current consumption habits in willingness to reduce meat consumption. In sum, the MDMQ seems to offer a valid, reliable, self-standing and parsimonious measure of moral disengagement as a situated cognitive process regarding meat consumption and substitution. One additional contribution provided by this work is that it helped to gain insights about how moral disengagement operates in everyday situations in which people routinely perform self-serving activities at injurious costs to others. Previous research has applied moral disengagement theory mainly to high moral intensity scenarios and behaviors such as interpersonal aggression (e.g., Paciello et al., 2008), stealing or deception (e.g., Chugh, Kern, Zhu, & Lee, 2014), and armed conflicts (e.g., McAlister et al., 2006), which Bandura (1999) termed “extraordinary circumstances” (p. 205). But the role of moral disengagement in normative, widespread and culturally cherished practices has remained virtually ignored. To shed light on this issue, we tested a sequential mediation model in which engagement in a self-serving behavior associates with unwillingness to change the behavior, via a positive valuation of the behavior and the employment of moral disengagement mechanisms to defend and justify it. The findings provided support for a full mediation in which frequency of meat consumption affected willingness towards meat substitution indirectly via meat attachment and moral disengagement (Fig. 2). It is known that while highly routinized actions (as is the case of meat consumption) hinder perceptions of moral relevance, individuals often face a motivational conflict when the damage of their behavior becomes salient (Kelman & Hamilton, 1989; Tsang, 2002). They may be willing to try to repair and minimize the damage or take steps to avoid engaging in the behavior, but they may also be motivated to justify and defend it, as letting the behavior go can imply considerable perceived costs for the self. In this later scenario, when approaching the situation with a strong initial preference towards a particular solution, a process of motivated reasoning may follow which drives one's cognitions in the direction of the desired resolution (Tsang, 2002). Applying these propositions to the current findings, one tentative interpretation is that meat consumption fosters feelings of meat attachment, which trigger moral disengagement as a loss-aversion mechanism when one is prompted to consider the impact of meat. Moral disengagement then acts as a motivated reasoning process which defends and justifies the behavior in spite of the damages it may imply, thus reducing or avoiding feelings of dissonance that are eventually caused by considering the detrimental consequences of ones' habits. While this interpretation will require further empirical substantiation in the near future, there is indeed evidence that when confronted with food-related considerations, individuals with a strong initial stance towards the issue tend to disregard opposing information and seek out congruent information to support their pre-existing inclinations (Ong, Frewer, & Chan, 2015).

Looking forward, the view that loss aversion and dissonance reduction trigger moral disengagement as a motivated reasoning process may have important theoretical and practical implications. As for theory development, it highlights that a broader understanding of moral agency requires an integrated perspective in which self-regulation is bounded by competing motivators affecting moral awareness and moral cognitions. Moving beyond a strictly ‘rationalized’ approach to moral reasoning, understanding how these motivators operate on a deeper level may be critical to explain why “large-scale inhumanities are often perpetrated by people who can be considerate and compassionate in other areas of their lives” (Bandura, 2007, p. 9). It may also open a window for integrating socio-cognitive and intuitionist approaches to moral behavior (Bandura et al., 1996; Haidt, 2001). As for practical implications, they raise a call for interventions to address the underlying core motivations that trigger the moral disengagement process on a deeper level, instead of simply focusing on situational manipulations aimed at making the contexts less conducive to moral disengagement. Staying with meat consumption as an example, much of our relationship with food and food choice occurs at a non-conscious level within deep-rooted patterns of habit and behavior (Köster, 2009), which very likely hinder willingness to change and perceptions of moral relevance. Thus, policies and interventions designed to limit meat consumption and encourage a shift towards a more plant-based diet (for either sustainability, health, or ethical reasons) will probably boost their efficacy if they are able to capture and bypass these competing motivations (e.g., meat attachment; Graça, Calheiros, et al., 2015). Another potentially promising direction for future research is to explore individual or group (cluster) differences in the employment of distinct disengagement mechanisms based on the situation, or in response to a given cue. For instance, Rothgerber (2014) noted that different individuals may engage in distinct patterns of justification when exposed to reminders of the aversive consequences of meat consumption (e.g. acting more ambivalent and dissociative vs. openly embracing the practice of meat consumption through various justifications). The MDMQ may assist in discriminating such patterns of response under the framework of moral disengagement theory. On this note, it may be worth exploring if different mechanisms will show distinct patterns or strength of association with variables such as moral awareness (Reynolds & Miller, 2015). One may question for example if individuals with higher (vs. lower) levels of moral awareness concerning environmental sustainability or animal suffering will differ in terms of their propensity to employ specific disengagement mechanisms, when prompted to consider the impact of meat consumption. While one would hardly find suitable answers for such questions drawing only on a trait-like measure of moral disengagement, it is expected that the MDMQ global scale and subscales will facilitate more fine-grained analyses and interpretations. This will allow the examination of how a range of variables in terms of personality, values, behavior and context may help eliciting or avoiding the different paths for moral disengagement. In effect, when assessing predictive ability, all subscales provided relatively independent contributions to explain our focal variable, even if structurally anchored in an overarching construct of moral disengagement. Also, these subscales yielded slightly different patterns of association with the measures referring to moral regulation and dominance ideologies which the current study assessed. Thus, taken as a whole, these findings and propositions suggest that moral disengagement comprises an intricate and relatively sophisticated system of motivated reasoning. Drawing on food practices to observe psychological processes associated with everyday moral action may open a window for further research to unravel the underpinnings of such a system. In spite of these possibilities, the present work is narrowed by its scope and cross-sectional nature. Being the first to measure and investigate moral disengagement situated in a normative, widespread and culturally cherished practice, it offers a starting point for further research. It is necessary to subject the propositions we advanced to experimental manipulation. Thus, one priority for future studies is to substantiate

J. Graça et al. / Personality and Individual Differences 90 (2016) 353–364

and improve the proposed sequential mediation model. Our rationale for the ordering of the variables was drawn theoretically and in accordance with the conclusions provided in previous research in moral disengagement, but could not yet follow from conclusive empirical evidence. According to our findings and interpretation, we expect that priming the salience of meat attachment will increase levels of moral disengagement when considering the impact of meat. Likewise, we predict that manipulations of the different disengagement mechanisms will decrease self-evaluative emotional reactions (i.e. guilt and shame) when individuals are exposed to information on the impact of meat, and influence willingness to change eating habits. Besides providing experimental support for our model, these studies may also help to shed light on the nature and temporal aspects of moral disengagement. Indeed, it still remains an open question if moral disengagement should be understood as an anticipatory cognitive exercise with a causal role in a given behavior, a set of post-hoc rationalizations of the behavior, or a mix between the two (i.e. a self-reinforcing cycle in which post hoc rationalizations for an act feed pre-act moral disengagement in future periods) (Detert et al., 2008; Reynolds et al., 2014).

363

8. If I saw an animal being killed I would have no problems eating it. 9. Those who eat meat should be aware of its impacts on public health. 10. By eating meat I engage with an industry responsible for major damages. 11. It's important that people who eat meat think about the impacts to the environment. 12. Nowadays there are good alternatives to meat consumption. 13. It's possible to have an adequate diet without eating meat. 14. It doesn't matter if I change my habits because problems will still exist. 15. It's easy to have a meat-free diet. 16. Despite everything, eating meat is part of a balanced life. 17. If I had to kill the animals myself, I would probably stop eating meat. 18. It would be difficult for me to watch an animal being killed for food purposes. 19. I would be capable of skinning, separating the organs and cutting an animal to pieces. 20. I will consider changing my habits only if others also change theirs. A.3. Response scale

4.1. Conclusion In The descent of man, and selection in relation to sex, Charles Darwin (1871) argued that of all differences between humans and other animals, “the moral sense or conscience is by far the most important” (p. 67). However, the paradox of why good people do bad things certainly complicates our views of what it really means to be invested with a “moral sense or conscience”. While contributions provided by moral disengagement theory have helped illuminating this qualm, they are limited in that they do not yet address how self-regulatory processes operate with reference to harmful but cherished and widespread practices. The issue of self-regulation in such practices may be critical not only for the sake of knowledge and raising self-awareness, but also for the thriving and survival of humanity. Generations to come must tackle unsustainable practices that on the current scale may eventually violate environmental thresholds that support human development. This work offers a first exploration of the role of moral disengagement as a cognitive process in the maintenance of such practices. Additionally, it suggests that there is much to be learned with a broadened perspective that integrates competing (self-serving) motivations in our understanding of moral awareness and moral selfregulation. Appendix A. MDMQ instructions, items, response scale and scoring

Response scale ranges from 1 (strongly disagree) to 5 (strongly agree). A.4. Scoring - Items 5, 7, 9, 10, 11, 12, 13, 15, 17, and 18 are reverse scored; - to calculate the scores for each dimension and the global scale: a. Means-ends justifications: MEAN (items: 1, 2, 3, 4, 16) b. Desensitization: MEAN (items: 8, 17, 18, 19) c. Denial of negative consequences: MEAN (items: 5, 7, 9, 10, 11) d. Diffused responsibility: MEAN (items: 6, 14, 20) e. Reduced perceived choice: MEAN (items: 12, 13, 15) f. Moral disengagement in meat (global scale): MEAN (items: 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20); and - final values for each dimension and global scale range from 1 to 5. Higher values indicate higher levels of moral disengagement when considering the impact of meat consumption.

Appendix B. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.paid.2015.11.042.

A.1. Participant instructions References In recent times, meat consumption is being increasingly debated on the grounds of environmental sustainability, health and safety concerns, and animal rights/welfare arguments. Please tell us your personal opinion about meat consumption indicating the extent to which you agree with the following sentences. Try to answer as spontaneously as possible, without giving much thought to the answers. A.2. Items 1. 2. 3. 4. 5.

All things considered, meat is necessary to human diet. The human being has needs that include eating meat. The problems associated with meat also apply to other foods. Eating meat keeps the balance of the food chain. By eating meat I'm also responsible for the problems associated to its production. 6. Even if I change my habits, I don't make a difference by myself. 7. People who eat meat should acknowledge the suffering in which food animals are kept.

Abdi, H. (2003). Factor rotations in factor analyses. In M. Lewis-Beck, A. Bryman, & T. Futin (Eds.), Encyclopedia for research methods for the social sciences (pp. 978–982). Thousand Oaks, CA: Sage. Aquino, K., & Reed, A. (2002). The self-importance of moral identity. Journal of Personality and Social Psychology, 83(6), 1423–1440. Aquino, K., Reed, A., Thau, S., & Freeman, D. (2007). A grotesque and dark beauty: How moral identity and mechanisms of moral disengagement influence cognitive and emotional reactions to war. Journal of Experimental Social Psychology, 43(3), 385–392. Arbuckle, J.L. (2011). Amos 20 User's Guide. Armonk, NY: IBM Corporation. Bandura, A. (1991). Social cognitive theory of moral thought and action. In W.M. Kurtines, & J.L. Gewirtz (Eds.), Handbook of moral behavior and development: Theory, research and applications, Vol. 1. (pp. 71–129). Hillsdale, NJ: Erlbaum. Bandura, A. (1999). Moral disengagement in the perpetration of inhumanities. Personality and Social Psychology Review, 3(3), 193–209. Bandura, A. (2007). Impeding ecological sustainability through selective moral disengagement. International Journal of Innovation and Sustainable Development, 2(1), 8–35. Bandura, A., Barbaranelli, C., Caprara, G.V., & Pastorelli, C. (1996). Mechanisms of moral disengagement in the exercise of moral agency. Journal of Personality and Social Psychology, 71, 364–374. Bastian, B., Loughnan, S., Haslam, N., & Radke, H.R.M. (2012). Don't mind meat? The denial of mind to animals used for human consumption. Personality and Social Psychology Bulletin, 38(2), 247–256. Berndsen, M., & van der Pligt, J. (2004). Ambivalence towards meat. Appetite, 42(1), 71–78.

364

J. Graça et al. / Personality and Individual Differences 90 (2016) 353–364

Bilewicz, M., Imhoff, R., & Drogosz, M. (2011). The humanity of what we eat: Conceptions of human uniqueness among vegetarians and omnivores. European Journal of Social Psychology, 41(2), 201–209. Bratanova, B., Loughnan, S., & Bastian, B. (2011). The effect of categorization as food on the perceived moral standing of animals. Appetite, 57(1), 193–196. Bratanova, B., Vauclair, C.M., Kervyn, N., Schumann, S., Wood, R., & Klein, O. (2015). Savouring morality. Moral satisfaction renders food of ethical origin subjectively tastier. Appetite, 91, 137–149. Bryman, A., & Cramer, D. (2011). Quantitative data analysis with IBM SPSS 17, 18 and 19: A guide for social scientists. Hove: Routledge. Buhrmester, M., Kwang, T., & Gosling, S.D. (2011). Amazon's Mechanical Turk a new source of inexpensive, yet high-quality, data? Perspectives on Psychological Science, 6(1), 3–5. Caprara, G.V., Fida, R., Vecchione, M., Tramontano, C., & Barbaranelli, C. (2009). Assessing civic moral disengagement: Dimensionality and construct validity. Personality and Individual Differences, 47(5), 504–509. Chugh, D., Kern, M.C., Zhu, Z., & Lee, S. (2014). Withstanding moral disengagement: Attachment security as an ethical intervention. Journal of Experimental Social Psychology, 51, 88–93. Claybourn, M. (2011). Relationships between moral disengagement, work characteristics and workplace harassment. Journal of Business Ethics, 100(2), 283–301. Darwin, C. (1871). The descent of man, and selection in relation to sex. London: John Murray. Detert, J.R., Treviño, L.K., & Sweitzer, V.L. (2008). Moral disengagement in ethical decision making: A study of antecedents and outcomes. The Journal of Applied Psychology, 93(2), 374–391. DeVellis, R. F. (1991). Scale development: Theory and applications. Newbury Park: Sage Publications. Dhont, K., & Hodson, G. (2014). Why do right-wing adherents engage in more animal exploitation and meat consumption? Personality and Individual Differences, 64, 12–17. Dhont, K., Hodson, G., Costello, K., & MacInnis, C.C. (2014). Social dominance orientation connects prejudicial human–human and human–animal relations. Personality and Individual Differences, 61, 105–108. Egan, V., Hughes, N., & Palmer, E.J. (2015). Moral disengagement, the dark triad, and unethical consumer attitudes. Personality and Individual Differences, 76, 123–128. Fida, R., Paciello, M., Tramontano, C., Fontaine, R.G., Barbaranelli, C., & Farnese, M.L. (2014). An integrative approach to understanding counterproductive work behavior: The roles of stressors, negative emotions, and moral disengagement. Journal of Business Ethics, 1-14. Fiddes, N. (1991). Meat: A natural symbol. New York: Routledge. Fox, N., & Ward, K. (2008). Health, ethics and environment: a qualitative study of vegetarian motivations. Appetite, 50(2), 422–429. IBM Corp. (2011). IBM SPSS Statistics for Windows, Version 20.0. Armonk, NY: IBM. Goldberg, L.R. (2001). International Personality Item Pool: A scientific collaboratory for the development of advanced measures of personality traits and other individual differences. Retrieved from http://ipip.ori.org Gorsuch, R.L. (1983). Factor analysis (2nd ed.). Hillsdale, NJ: Erlbaum. Graça, J., Calheiros, M.M., & Oliveira, A. (2015a). Attached to meat? (Un)Willingness and intentions to adopt a more plant-based diet. Appetite. http://dx.doi.org/10.1016/j. appet.2015.06.024 (Advance online publication). Graça, J., Oliveira, A., & Calheiros, M.M. (2015b). Meat, beyond the plate. Data-driven hypotheses for understanding consumer willingness to adopt a more plant-based diet. Appetite, 90, 80–90. Graça, J., Calheiros, M.M., & Oliveira, A. (2014). Moral disengagement in harmful but cherished food practices? An exploration into the case of meat. Journal of Agricultural and Environmental Ethics, 27(5), 749–765. Haidt, J. (2001). The emotional dog and its rational tail: A social intuitionist approach to moral judgment. Psychological Review, 108(4), 814–834. Hayes, A.F. (2012). Process: A versatile computational tool for observed variable mediation, moderation, and conditional process modeling. Retrieved from http://www. afhayes.com/public/process2012.pdf Hayes, A.F. (2013). Introduction to mediation, moderation, and conditional process analysis: A regression-based approach. New York, NY: The Guilford Press. Hoek, A.C., Luning, P.A., Weijzen, P., Engels, W., Kok, F.J., & de Graaf, C. (2011). Replacement of meat by meat substitutes. A survey on person- and product-related factors in consumer acceptance. Appetite, 56(3), 662–673. Horn, J.L. (1965). A rationale and test for the number of factors in factor analysis. Psychometrika, 30, 179–185. Hu, L., & Bentler, P. (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modelling, 6(1), 1–55. IBM Corp (2011). IBM SPSS statistics for Windows, version 20.0. Armonk, NY: IBM. Jackson, L.E., & Gaertner, L. (2010). Mechanisms of moral disengagement and their differential use by right‐wing authoritarianism and social dominance orientation in support of war. Aggressive Behavior, 36(4), 238–250.

Kelman, H.C., & Hamilton, V.L. (1989). Crimes of obedience: Toward a social psychology of authority and responsibility. New Haven, CT: Yale University Press. Köster, E.P. (2009). Diversity in the determinants of food choice: A psychological perspective. Food Quality and Preference, 20(2), 70–82. Loughnan, S., Bastian, B., & Haslam, N. (2014). The psychology of eating animals. Current Directions in Psychological Science, 23(2), 104–108. Marsh, H.W., Hau, K.T., & Wen, Z. (2004). In search of golden rules: Comment on hypothesis-testing approaches to setting cutoff values for fit indexes and dangers in overgeneralizing Hu and Bentler's (1999) findings. Structural Equation Modeling, 11(3), 320–341. Matsunaga, M. (2010). How to factor-analyze your data right: Do's, don'ts, and how-to's. International Journal of Psychological Research, 3(1), 97–110. McAlister, A.L., Bandura, A., & Owen, S.V. (2006). Mechanisms of moral disengagement in support of military force: The impact of Sept. 11. Journal of Social and Clinical Psychology, 25(2), 141–165. Obermann, M.L. (2011). Moral disengagement in self-reported and peer-nominated school bullying. Aggressive Behavior, 37(2), 133–144. O'Connor, B.P. (2000). SPSS and SAS programs for determining the number of components using parallel analysis and Velicer's MAP test. Behavior Research Methods, Instruments, & Computers, 32(3), 396–402. Ong, A.S., Frewer, L., & Chan, M. (2015). Cognitive dissonance in food and nutrition — A review. Critical Reviews in Food Science and Nutrition. http://dx.doi.org/10.1080/ 10408398.2015.1013622. Paciello, M., Fida, R., Tramontano, C., Lupinetti, C., & Caprara, G.V. (2008). Stability and change of moral disengagement and its impact on aggression and violence in late adolescence. Child Development, 79(5), 1288–1309. Page, T.E., & Pina, A. (2015). Moral disengagement as a self-regulatory process in sexual harassment perpetration at work: A preliminary conceptualization. Aggression and Violent Behavior, 21, 73–84. Piazza, J., Ruby, M.B., Loughnan, S., Luong, M., Kulik, J., Watkins, H.M., & Seigerman, M. (2015). Rationalizing meat consumption. The 4Ns. Appetite, 91, 114–128. Plous, S. (2003). Is there such a thing as prejudice toward animals? In S. Plous (Ed.), Understanding prejudice and discrimination (pp. 509–528). NewYork: McGraw-Hill. Pluhar, E.B. (2010). Meat and morality: Alternatives to factory farming. Journal of Agricultural and Environmental Ethics, 23(5), 455–468. Pratto, F., Sidanius, J., Stallworth, L.M., & Malle, B.F. (1994). Social dominance orientation: A personality variable predicting social and political attitudes. Journal of Personality and Social Psychology, 67(4), 741–763. Reynolds, S.J., & Miller, J.A. (2015). The recognition of moral issues: Moral awareness, moral sensitivity and moral attentiveness. Current Opinion in Psychology, 6, 114–117. Reynolds, S.J., Dang, C.T., Yam, K.C., & Leavitt, K. (2014). The role of moral knowledge in everyday immorality: What does it matter if I know what is right? Organizational Behavior and Human Decision Processes, 123(2), 124–137. Rosenblatt, V. (2012). Hierarchies, power inequalities, and organizational corruption. Journal of Business Ethics, 111(2), 237–251. Rothgerber, H. (2013). Real men don't eat (vegetable) quiche: Masculinity and the justification of meat consumption. Psychology of Men & Masculinity, 14(4), 363–375. Rothgerber, H. (2014). Efforts to overcome vegetarian-induced dissonance among meat eaters. Appetite, 79, 32–41. Ruby, M.B. (2012). Vegetarianism. A blossoming field of study. Appetite, 58(1), 141–150. Ruby, M.B., & Heine, S.J. (2011). Meat, morals, and masculinity. Appetite, 56(2), 447–450. Schermelleh-Engel, K., Moosbrugger, H., & Muller, H. (2003). Evaluating the fit of structural equation models: Tests of significance and descriptive goodness-of-fit measures. Methods of Psychological Research Online, 8(2), 23–74. Schösler, H., de Boer, J., & Boersema, J.J. (2012). Can we cut out the meat of the dish? Constructing consumer-oriented pathways towards meat substitution. Appetite, 58(1), 39–47. Tangney, J.P., Stuewig, J., & Mashek, D.J. (2007). Moral emotions and moral behavior. Annual Review of Psychology, 58, 345–372. Tilman, D., & Clark, M. (2014). Global diets link environmental sustainability and human health. Nature, 515(7528), 518–522. Tsang, J.A. (2002). Moral rationalization and the integration of situational factors and psychological processes in immoral behavior. Review of General Psychology, 6(1), 25–50. van Dooren, C., Marinussen, M., Blonk, H., Aiking, H., & Vellinga, P. (2014). Exploring dietary guidelines based on ecological and nutritional values: A comparison of six dietary patterns. Food Policy, 44, 36–46. Vandenberg, R.J., & Lance, C.E. (2000). A review and synthesis of the measurement invariance literature: Suggestions, practices, and recommendations for organizational research. Organizational Research Methods, 3(1), 4–70. Vollum, S., Buffington-Vollum, J., & Longmire, D.R. (2004). Moral disengagement and attitudes about violence toward animals. Society and Animals, 12(3), 209–235. Westhoek, H., Lesschen, J.P., Rood, T., Wagner, S., De Marco, A., Murphy-Bokern, D., ... Oenema, O. (2014). Food choices, health and environment: Effects of cutting Europe's meat and dairy intake. Global Environmental Change, 26, 196–205.