Collective self-fulfilling prophecies: group identification biases perceptions of environmental group norms among high identifiers

Abstract

Previous research has shown that ingroup norms influence intentions to engage in pro-environmental behavior, most notably for individuals highly identified with a group. However, intriguingly, identification may itself lead people to exaggerate descriptive pro-environmental ingroup norms to enhance positive distinctiveness of their ingroup. We investigated this possibility in two studies together with the assumption that perceived norms would mediate the effects of identification on intentions to engage in pro-environmental behavior. The results support our assumptions and show that high identifiers perceived their ingroup be more eco-friendly, which in turn increased respondents’ intentions to behave in a pro-environmental manner. We discuss the implications of this collective self-fulfilling prophecy for social identity theorizing and the prediction of pro-environmental behavior.

Keywords:

• Social identity
• norms
• pro-environmental behavior

Introduction

Social norms are powerful in influencing human behavior. Norm focus theory (Cialdini, Reno, & Kallgren, 1990) suggests that social norms that are salient in a given situation will influence individuals’ behavior and a host of research on environmental behavior supports this assumption (Cialdini et al., 1990; Fritsche, Jonas, Kayser, & Koranyi, 2010; Nolan, Schultz, Cialdini, Goldstein, & Griskevicius, 2008; Schultz, Nolan, Cialdini, Goldstein, & Griskevicius, 2007). Cialdini and Trost (1998) differentiate between descriptive (e.g., what others typically do) and injunctive (e.g., what others approve of) norms. Both types of norms have been found to be highly predictive of individual behavior (e.g., Cialdini et al., 1990; Nolan et al., 2008). The theory of planned behavior (TPB; Ajzen, 1991) also suggests that normative beliefs—what individuals think significant others expect of them—will determine individuals’ intentions to carry out a certain behavior (e.g., recycle products). Again, a large number of studies confirm the ability of TPB to predict a range of pro-environmental behaviors (Bamberg & Schmidt, 2003; Harland, Staats, & Wilke, 1999; Heath & Gifford, 2002; Kaiser & Gutscher, 2003). However, these theories are not concerned with how individuals arrive at an idea about what is normative.

Yet, an individuals’ perception of what is normative may vary as a function of individual and group-level cognitions and motivations not accounted for in these theories. Based on the tenets of social identity theory (SIT; Tajfel & Turner, 1979) we suggest that the extent to which people identify with a given social group will have profound effects on the attribution of pro-social norms to the ingroup, and in turn on their own behavior. Due to their motivation for positive ingroup distinctiveness high identifiers may exaggerate descriptive prosocial ingroup norms which then trigger own pro-sociality through a process of self-sterotyping (Hogg & Turner, 1987). To investigate this collective self-fulfilling prophecy the present research specifically tests whether ingroup identification impacts upon people’s perception of environmental ingroup norms and on their intentions to engage in pro-environmental behavior. In doing so, we also aim to provide further evidence for the feasibility of the social identity approach (i.e., SIT & Self-Categorization Theory; Reicher, Spears, & Haslam, 2010; Turner, Hogg, Oakes, Reicher, & Wetherell, 1987) for environmental research.

A social identity perspective on environmental norm perception

Recently, interest in a social identity perspective on environmental behavior has been growing, especially for research linked to social norms. From a social identity perspective, social influence operates on the basis of salient identities (i.e., group memberships; Tajfel & Turner, 1986; Turner, 1991) as a process of referent informational influence (Abrams, Wetherell, Cochrane, Hogg, & Turner, 1990). When a social identity is salient, people cognitively represent and self-stereotype themselves as ingroup members and, thus, are more likely to adhere to ingroup (vs. outgroup) norms. Several studies have found ingroup norms to predict pro-environmental behavior (Fielding, Terry, Masser, & Hogg, 2008; Masson & Fritsche, 2014; McDonald, Fielding, & Louis, 2013; Nigbur, Lyons, & Uzzell, 2010; Smith et al., 2012; Terry, Hogg, & White, 1999; White, Smith, Terry, Greenslade, & McKimmie, 2009). However, while previous research has addressed the links between environmental group norms and behavior, the question whether identification influences members’ perceptions of environmental norms has received less attention. Although research on self-categorization has revealed cognitive determinants of how people construe ingroup prototypes (Turner et al., 1987), motivational antecedents of ingroup norm perceptions have rarely been investigated.¹ However, intriguingly, motivated exaggeration of pro-environmental norms in highly identified group members may not only explain distortions of the ingroup prototype. It may also give rise to collective self-fulfilling prophecies as biased descriptive ingroup norms should in turn predict pro-environmental behavior, which means that identification would represent an indirect predictor of such behavior. Investigating motivated norm perception in groups is thus a highly relevant endeavor.

SIT suggests that individuals are motivated to sustain a positive social identity and thus aim to self-categorize as members of positively valued groups. Highly identified group members may therefore be motivated to ascribe more positive (or less negative) characteristics to their ingroup, including ingroup norms that imply socially valued behavior (e.g., pro-environmental behavior). Applying SIT terminology: The link between ingroup identification and ingroup norms ultimately concerns an individual’s perception of the ingroup prototype (i.e., the attitudes and behaviors associated with a “typical” ingroup member), the latter being a function of intergroup comparisons (Abrams, Marques, Bown, & Henson, 2000), discursive processes at the social or ingroup level as well as of individual attributions. According to SIT, people pay close attention to group members assumed to be prototypical (e.g., group leaders) and majority views as a source of reliable information about norms (Hogg, 2010). While for some groups (e.g., small sport teams) a clearly defined ingroup prototype exists, for others the prototype is more ambiguous (e.g., broad social categories). Hence, a group’s prototype, sometimes being “a fuzzy set of attributes…that capture similarities and structural relationships within the group” (Hogg, 2005, p. 138), may be open to biased perceptions on the individual level—albeit within the limits of (consensual definitions of) social reality (Ellemers, van Rijswijk, Roefs, & Simons, 1997).² Although a prototype has to be shared among the group’s members to provide a meaningful point of reference, individuals’ cognitive representation of the group may vary situationally depending on motivational demands (e.g., positive identity). Perceptions of ingroup norms thus could be employed as a strategy targeted at the satisfaction of individual and group-level motives (see below).

Perception of environmental ingroup norms: a motivational account

The basic idea of the social identity approach is that people derive a significant part of their self-concept from the social groups to which they belong. By implication, characteristics associated with these groups will affect their social identity. According to SIT, individuals strive for a positive and distinct social identity to satisfy, among others, self-esteem and distinctiveness needs. This means that group members will be motivated to see their group in a positive light and as different from other groups. Thus, highly identified group members may be “motivated tacticians” (Fiske, 1993) who due to their need for a positive and distinct social identity may have biased perceptions of their ingroup’s norms. Indeed, research has shown that highly identified group members are more likely to perceive typical ingroup traits positively (Ellemers et al., 1997) and to perceive positive traits as more typical for their group (Castano, Paladino, Coull, & Yzerbyt, 2002) than low identifiers do. Theses positivity biases among high identifiers for what is typical for the ingroup were also found in other studies (Hutchison & Abrams, 2003; Hutchison, Abrams, & De Moura, 2013; Hutchison, Abrams, Gutierrez, & Viki, 2008), albeit they mainly occurred when some level of group threat was present—for example through low ingroup status. However, according to a prominent taxonomy on the interaction of different threats and the level of ingroup commitment on perceptual responses (Ellemers, Spears, & Doosje, 2002) ingroup identification should also elevate distinctiveness and enhancement motives in the absence of threat. That is, highly identified group members should be motivated to perceive their ingroup in positive ways - thereby increasing the potential of an ingroup to function as a “resource” of positive identity (Correll & Park, 2005).

The present research

Building on our theorizing above we posit that individuals’ level of identification with an ingroup will influence their perception of ingroup norms linked to environmental behavior (i.e., a behavior that is positively valued in Germany; BMU – Bundesministerium für Umwelt und Naturschutz [Federal Environment Ministry], 2010, 2012). More specifically, we expect (experimentally manipulated) ingroup identification to positively predict perceptions of fellow ingroup members to behave in a more pro-environmental manner (Hypothesis 1). Furthermore, we expect ingroup identification to have an indirect effect on individuals’ pro-environmental intentions, mediated by perceived ingroup norms (Hypothesis 2). In this paper we present two studies testing these assumptions based on samples from Germany. In both studies, we manipulated levels of ingroup identification, albeit with different experimental designs, to investigate the effects of identification on perceptions of descriptive environmental ingroup norms. We chose environmental norms as the majority of the Germans have reported to hold pro-environmental attitudes (BMU – Bundesministerium für Umwelt und Naturschutz [Federal Environment Ministry], 2010, 2012). Furthermore, survey results indicate that a considerable share of Germans support behaviors aimed at the reduction of humans’ environmental impact. Pro-environmental behavior can thus be assumed to reflect a positive group characteristic.

Study 1

In our first study, we tested the effects of ingroup identification (manipulated) on perceived ingroup norms and behavior intentions for an index of six pro-environmental behaviors. We also controlled for the effects of consumer efficacy and personal environmental attitudes, two well-established predictors of pro-environmental behavior.

Method

Sample

One hundred and seventy-seven students were approached at a German university campus and agreed to participate in a paper-and-pencil survey on environmental attitudes and behavior. We excluded one participant due to missing data. The final sample contained ninety-nine women and seventy-seven men (M_age = 25.14, SD_age = 3.82).

Procedure and measures

Table 1 presents the means, standard deviations, Cronbach’s alpha coefficients, and inter-scale correlations for each of the variables.

Table 1. Means, standard deviations, Cronbach’s alpha coefficients (provided in parentheses), and inter-scale correlations between variables in Study 1.

Variable	M	SD	1.	2.	3.	4.	5.	6.
1. Intention	4.53	1.10	(.68)	.27**	.40**	.52**	.04	.11
2. Perceived ingroup norm	3.56	1.06		(.80)	.05	.08	.32**	.39**
3. Consumer efficacy beliefs	5.20	0.95			(.70)	.35**	.07	−.05
4. Environmental attitudes	4.49	4.67				(.84)	.02	.00
5. Ingroup identification (measured)	3,88	1.32					(.90)	.47**
6. Ingroup identification (manipulated)	0.06	1.00						^a

^a Cronbach's alpha not computed (single item measure).

** p < .01.

Prior to the manipulation, respondents were asked to indicate their consumer efficacy beliefs for pro-environmental behaviors on six seven-point scale items (from 1 = “does not help much” to 7 = “helps very much”).³ The items were “How much does it help to fight climate change if the majority of the consumers [use green electricity; reduce household waste; switch off electric appliances when not in use; boycott companies harmful to the environment; purchase local/regional products; reduce meat consumption or keep a vegetarian diet]?” For the final scale we excluded one item on reducing meat consumption because it had a low corrected item-total correlation of r(176) = .21. The statistics provided in Table 1 refer to the final scale.

Participants were then randomly assigned to one of the two experimental conditions: high vs. low ingroup identification. In the high (low) identification condition, participants were asked to think about a group they enjoy belonging to and which they strongly identify with (a group they belong to but only weakly identify with). Participants were given several examples of possible ingroups including general social categories (e.g., sex, age), university and vocational groups (e.g., students of a university, co-workers), recreational groups (e.g., members of a sport team), local/regional groups (inhabitant of a city/region), religious groups, and opinion-based groups (e.g., globalization critics). Respondents were asked to either choose one of the groups provided or to name another ingroup. In the high identification condition, 53.7% of the participants selected university or vocational groups, 12.7% recreational groups, 8.4% local/regional groups, 6.3% group of mothers or fathers, 5.3% opinion-based groups, 2.1% religious groups, 11.5% other groups. In the low identification condition, 34.1% of the respondents chose a local/regional group, 31.7% university-related groups, 14.6% social categories (age, sex), 14.7% religious groups, 4.9% other groups.

After the manipulation, participants completed a seven-item measure of ingroup identification based on Leach et al. (2008), serving as a manipulation check (e.g., “I am glad to be an ingroup member.”; seven-point scale ranging from 1 = “strongly disagree” to 7 = “strongly agree”). Next, we assessed perceived ingroup norms linked to pro-environmental behavior using six items (“What is the proportion of ingroup members that [use green electricity; switch off electronic appliances when not in use; reduce household waste; purchase regional products; reduce meat consumption or keep a vegetarian diet; boycott companies harmful to the environment]?”; seven-point scale ranging from 1 = “very small” to 7 = “very large”).

Following the norm items, six items measured respondents’ intention to engage in pro-environmental behavior in the future (“How often do you plan to [reduce household waste; switch off electric appliances not in use; reduce meat consumption or keep a vegetarian diet; purchase local/regional products] within the next 2 months?”; seven-point scale from 1 = “never” to 7 = “always”; ‘I intend to [boycott polluting companies; use green electricity] in the future / in the following 12 months.]; seven-point scale from 1 = “strongly disagree”, 7 = “strongly agree”). Finally, three items were used to assess environmental attitudes (e.g., “To behave eco-friendly is important for me”; seven-point scale from 1 = “strongly disagree” to 7 = “strongly agree”).

Results and discussion

Manipulation check

Ingroup identification was greater in the high identification condition (M = 4.46, SD = 1.08) compared to the low identification condition (M = 3.22, SD = 1.25; F(1, 174) = 49.9, p < .001, η_p² = .22), indicating a successful manipulation of ingroup identification.

Ingroup identification, perceived ingroup norm and pro-environmental intentions

As shown in Table 1, manipulated ingroup identification was positively correlated with perceived ingroup norms, r(176) = .40, p < .01, indicating that respondents in the high identification condition perceived their ingroup to behave in a more pro-environmental manner (M_{ingroup norm} = 3.94, SD_{ingroup norm} = 1.00) than those in the low identification condition (M_{ingroup norm} = 3.11, SD_{ingroup norm} = 0.96; F(1, 174) = 30.98, p < .001, η_p² = .15). In contrast, intention to engage in pro-environmental behavior did not differ between the high identification condition (M_intention = 4.65, SD_intention = 1.10) and the low identification condition (M_intention = 4.39, SD_intention = 1.09; F(1, 174) = 2.30, p = .13).⁴

To test whether the relation between ingroup identification and intention to engage in pro-environmental behavior was mediated by perceived ingroup norm, we conducted a simple mediation analysis, employing the PROCESS Macro for SPSS (Hayes, 2013; Model 4). We entered manipulated ingroup identification as an independent variable, perceived ingroup norms as mediator, and intention to engage in pro-environmental behavior as dependent variable. We also included environmental attitudes (b = .36, p < .01) and perceived consumer efficacy (b = .28, p < .01), two well-established predictors of pro-environmental behavior, in the mediation analysis (as covariates) to test the stability of our results. The mediation model is depicted in Figure 1.⁵ The total and direct effects of ingroup identification on intention were .14, p < .05 and .05, p = .50 respectively. As hypothesized, there was a significant indirect effect of identification on intention through ingroup norm (IE = .09, SE = .04), with a bias corrected (BC) 95% confidence interval (CI) of .03 to .19. That is, greater ingroup identification led to perceptions of the ingroup as more eco-friendly, which in turn led to higher intentions to behave in a pro-environmental manner.

Figure 1. Simple mediation of the group identification-intention relationship through perceived ingroup norm in Study 1(unstandardized regression coefficients).

Note: Environmental attitudes and consumer efficacy beliefs (covariates) are omitted. *p < .05; **p < .01.

Consistent with Hypothesis 1, Study 1 shows that high (vs. low) identifiers perceived their fellow ingroup members to behave in a more pro-environmental manner (i.e., pro-environmental ingroup norm). Furthermore, mediation analysis indicated that ingroup norms were positively associated with pro-environmental intentions and that ingroup identification indirectly influenced respondents’ intentions to engage in eco-friendly action, thus supporting Hypothesis 2. Finally, inclusion of environmental attitudes and consumer efficacy beliefs (as covariates) in the mediation analysis did not change the significance of our results.

However, the interpretation of our findings might be restricted as the participants were allowed to freely choose a group to manipulate ingroup identification. That is, asking people to name groups they highly identify within the context of a study on environmental attitudes may have led to the a priori selection of pro-environmental groups. This would compromise our account of biased-norm perception in high identifiers. We conducted a second study to rule out this alternative explanation.

Study 2

In Study 2, we aimed to replicate the results of Study 1 and to exclude a potential alternative explanation. Respondents were thus not allowed to freely choose an ingroup, but we manipulated identification with two laboratory groups (cognitive styles) to show that high (vs. low) levels of ingroup identification lead to ascription of more pro-environmental behavior among ingroup members (i.e., biased perception of the ingroup prototype).

Method

Sample

One hundred and twenty-one students were approached at a German university campus and agreed to participate in a paper-and-pencil survey. As a cover story, participants were told that the survey aimed to investigate potential links between cognitive styles, personality and climate change-relevant attitudes and behaviors. We excluded three participants due to missing data on the two dependent variables. The final sample contained seventy women and fourty-eight men (M_age = 22.61, SD_age = 3.37).

Procedure and measures

Table 2 presents the means, standard deviations, Cronbach’s alpha coefficients, and inter-scale correlations for each of the variables. Nine-point scales were employed for all measures to increase variance in our measures.

Table 2. Means, standard deviations, Cronbach's alpha coefficients (provided in parentheses), and inter-scale correlations between variables in Study 2.

Variable	M	SD	1.	2.	3.	4.	5.	6.
1. Intention	3.25	1.06	(.76)	.54**	.63**	.10	.23*	−.05
2. Perceived ingroup norm	2.82	0.94		(.82)	.38**	.19*	.22*	−.06
3. Environmental attitudes	3.83	1.24			(.90)	.21*	.06	.06
4. Ingroup identification (measured)	3.61	1.06				(.92)	.43**	−.03
5. Ingroup identification (manipulated)	0.19	0.98					^a	−.06
6. Self-esteem	0.19	0.98						(.90)

^a Cronbach's alpha not computed (single item measure).

* p < .05;

** p < .01.

Participants were randomly assigned to one of the two experimental conditions designed to manipulate respondents’ identification with a laboratory group (high vs. low ingroup identification). Participants were presented a series of two ambiguous or reversible figures, which are optical illusion images that allow for two distinct interpretations of what they represent (Wimmer & Doherty, 2011). They were then asked to indicate verbally what they saw on each of the pictures. The reversible figures showed ambiguous stimuli that could be perceived either as an old or a young woman and as a duck or a rabbit . Following that, participants received the questionnaire, ostensibly based on their previous interpretation of the ambiguous stimuli. The first page of the questionnaire contained a half-page text about two (fictional) cognitive styles denoting two groups: the deep or the broad approach. The text stated that members of both groups (broad vs. deep) differ with regard to perception, processing and storage / retrieving of information.

In the high identification condition, participants were told that, based on their interpretation of the reversible figures, they belonged to the group deep / broad approach.⁶ Participants then read that members of their group, among others, can perceive and process information more quickly and possess better analytical abilities than members of the other group. Participants then answered four items to repeat the information provided by the text (e.g., “Members of the [cognitive style group] possess good analytical abilities”; “agree” or “disagree”), Finally, participants were asked to indicate a social identity (i.e., group membership) unimportant for their self. We thereby aimed to further increase the situational importance of the cognitive style group. Respondents were given several examples of possible ingroups (e.g., sex, age, national group) and were asked to either choose one of the groups provided or to name another ingroup unimportant for their self.

Similarly, in the low identification condition, respondents were told that members of their group perceived and processed information slower and stored information in a less systematic manner than people with a different cognitive style. Again, participants then answered four items to repeat the information provided by the text (e.g., “Members of the [cognitive style group] store information in a less systematic manner”; “agree” or “disagree”). Finally, respondents were asked to indicate or name an ingroup important for their self.

Following the manipulation, identification with the cognitive style group (broad or deep approach) was measured using five items, serving as a manipulation check (e.g., “I am glad to be a member of [cognitive style group].”, “I identify with [cognitive style group]”; nine-point scale ranging from 1 = “strongly disagree” to 9 = “strongly agree”).

We used slightly different measures of perceived ingroup norm and intention to engage in pro-environmental behavior that included eight behaviors (instead of six, Study 1): use green electricity; reduce household waste; switch off electric appliances when not in use; boycott companies harmful to the environment; purchase local/regional products; reduce meat consumption or keep a vegetarian diet to fight climate change, avoid going by plane, donate money to environmental organizations. Environmental attitudes were measured using the same items as in Study 1. Finally, we included a German adaption of Rosenberg’s self-esteem scale (Ferring & Filipp, 1996) to account for possible effects of the manipulation of identification on personal self-esteem. After returning the questionnaire, all participants were thanked, fully debriefed, and dismissed.

Results and discussion

Manipulation check

Identification with the cognitive style group was greater in the high condition (M = 4.54, SD = 2.16) compared to the low condition (M = 2.76, SD = 1.58; F(1, 116) = 26.31, p < .001, η_p² = .18), indicating a successful manipulation of ingroup identification. Furthermore, personal self-esteem was not greater in the high identification condition (M = 6.55, SD = 1.57) compared to the low identification condition (M = 6.73, SD = 1.53; F(1, 116) = 0.39, p = .54), indicating that our manipulation of ingroup identification did not affect personal self-esteem.

Ingroup identification, perceived ingroup norm and pro-environmental intentions

As shown in Table 2, manipulated ingroup identification was positively correlated with perceived ingroup norms, r(118) = .22, p = .01. Consistent with Study 1, respondents in the high identification condition perceived their ingroup to behave in a more pro-environmental manner (M_{ingroup norm} = 5.10, SD_{ingroup norm} = 1.09) than those in the low identification condition (M_{ingroup norm} = 4.55, SD_{ingroup norm} = 1.32; F(1, 116) = 5.99, p = .02, η_p² = .05). Similarly, intention to engage in pro-environmental behavior was higher in the high identification condition (M_intention = 5.84, SD_intention = 1.15) than in the low identification condition (M_intention = 5.23, SD_intention = 1.47; F(1, 116) = 6.43, p = .01, η_p² = .05)

As in Study 1, we conducted a mediation analysis to test whether the relation between (manipulated) ingroup identification and intention to engage in pro-environmental behavior was mediated by the perceived ingroup norm. We also included environmental attitudes (b = .36, p < .01) in the analysis (as covariate). The mediation model is depicted in Figure 2. The total and direct effects of ingroup identification on intention were .27, p < .01 and .18, p = .05 respectively. Consistent with Study 1, there was a significant indirect effect of identification on intention through the perceived ingroup norm (IE = .10, SE = .00, 95% CI = 02. to .22).⁷ That is, greater ingroup identification led to perceptions of the ingroup as more eco-friendly, which in turn led to higher intentions to engage in pro-environmental behavior. Our results thus support Hypotheses 1 and 2.

Figure 2. Simple mediation of the group identification-intention relationship through perceived ingroup norm in Study 2 (unstandardized regression coefficients).

Note: Environmental attitudes (covariate) is omitted. *p < .05; **p < .01.

Replicating the results of Study 1, Study 2 found that high identifiers perceived their fellow ingroup members to engage in more pro-environmental behavior (i.e., pro-environmental ingroup norm) than those less identified with their cognitive style group (i.e., support for Hypothesis 1). Mediation analysis also indicated that ingroup norms were positively associated with pro-environmental intentions and that ingroup identification (indirectly) influenced respondents’ intentions to behave in a more eco-friendly manner, thus further corroborating Hypothesis 2.

General discussion

The aim of the present research was to examine the links between the level of individuals’ identification with an ingroup and their perception of fellow ingroup members’ pro-environmental behavior (i.e., perceived environmental ingroup norms). As predicted, those people who were highly (vs. less) identified with a group perceived their fellow ingroup members to engage in more pro-environmental behavior. As stated by social identity theory, people tend to derive a significant part of their self-concept from the groups they belong to. Thus, individuals should be motivated to (re)present groups central for their self in a positive light, be it to cast a positive light on themselves or to strengthen the group as a “valuable” identity resource. Pro-environmental behavior can be considered as a form of socially valued conduct. Ascribing higher levels of environmentalism to an ingroup, therefore, should increase the group’s potential to provide a positive identity and should thus be more attractive for high identifiers—an assumption supported by the present research.

Furthermore, levels of ingroup identification also had an indirect impact upon a person’s pro-environmental intentions in the sense of a collective self-fulfilling prophecy: Highly identifying with an ingroup increased people’s perception of the ingroup as acting in a pro-environmental manner (descriptive ingroup norm) which in turn seemed to drive their personal pro-environmental intentions. Obviously, identification was not only linked to the distortion of ingroup perception but also affected, albeit indirectly, decisions to engage in pro-environmental behavior in the future. The indirect effects of identification on intention remained stable after the inclusion of other well-established predictors (i.e., environmental attitudes, perceived consumer efficacy), indicating that ingroup identification can influence pro-environmental intentions via biased perceptions of the ingroup as a distinct causal process and even in the face of barriers (i.e., low perceived efficacy).

The results support previous studies that found high identifiers to perceive positive traits as more typical for their group (e.g., Hutchison et al., 2013). Most of these studies, however, detected biased ingroup perceptions when some level of group threat was present. In contrast, our results indicate that biased perception may also occur in the absence of such threat, but as a result of motivational processes linked to positive identity and distinctiveness. In addition, Rabinovich et al. (2012) found that perceptions of the ingroup prototype varied as function of the inter-group comparison context (more vs. less eco-friendly outgroup), thus providing only limited evidence for a “glorification” of the ingroup. Perceiving ingroups as more or less eco-friendly could reflect an identity management strategy to ensure (or at least: increase the possibility of) positive “social” comparisons (Tajfel & Turner, 1979). Our results also hint at possible limitations that measures of perceived (group or social) norms imply. Accordingly, self-reported norm measures that are frequently applied in environmental research, for example an individual’s perception of what significant others do (or think the individual should do), might be subject to biases due to motivational processes (e.g., need for positive distinctiveness). Often such biases will work outside conscious awareness, and thus constitute a part of the individual’s social reality.

Despite consistent support for the hypothesized effects across two studies, future research may focus on different types of behavior and norms to examine further the conditions of motivated group-based social influence. Concerning behavior, future studies could build on the differentiation between high-cost and low-cost behaviors to test the reliability of the present results. Previous research has found ingroup norms to be more predictive for high-cost (vs. low-cost) behavior (Masson & Fritsche, 2014). Accordingly, perceptions of ingroup norms may be more affected by the level of ingroup identification for high-cost norms, as the latter are more indicative of the ingroup’s potential as valuable social identity. Concerning norms, future studies could test whether biased norm perception is also present for injunctive norms. Furthermore, experimental designs could be employed to test the effects of ambiguous norm information on perceived ingroup norms. Ambiguous norm information may strengthen the observed effects of group identification by increasing differences in norm perceptions among high and low identifiers.

Conclusion

The present research shows that the level of ingroup identification influences an individual’s perception of fellow ingroup members’ behavior. Two studies indicate that high identifiers ascribe higher levels of pro-environmental behavior to their ingroup than low identifiers. Additionally, results showed that identification also exhibited an indirect effect on respondents’ intention to behave in a pro-environmental manner, mediated by perceived ingroup norms. Applying the social identity approach, the research contributes to new perspectives for both pro-environmental behavior research and research on social norms. Specifically, it offers insight into the beneficial (indirect) effects of social identity: People who have groups they highly identify with tend to see the group, and thus themselves, in a positive light, leading them to manifest pro-sociality in their own conduct. They fulfill a collective prophecy.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1. Antecedents of personal (or moral) norms have been proposed, among others, by the Norm Activation Model (Schwartz & Howard, 1981) and the Value-Belief-Norm Theory (Stern, Dietz, Abel, Guagno, & Kalof, 1999).

2. Furthermore, what is perceived as prototypical for the ingroup will vary as a function of the (intergroup) comparison context (Turner, 1999). For example, business students described the ingroup’s prototype as being rational and carrier-minded when psychology students were the salient outgroup, but not when comparing themselves to experienced business managers (Ferguson, Branscombe, & Reynolds, 2011; Rabinovich, Morton, Postmes, & Verplanken, 2012).

3. We also measured perceived severity of climate change and responsibilities linked to combat climate change prior to the manipulation. However, inclusion of these scales in the analyses did not change the significance of the results reported below. They were thus omitted from the paper.

4. Across both studies, respondents' age was neither correlated with perceived ingroup norm nor with intentions to behave in a pro-environmental manner. Furthermore, women (vs. men) did not report higher levels of perceived ingroup norm and but intended to engage in pro-environmental behavior more often. However, inclusion of sex in the analysis did not change the significance of the results reported below.

5. If not otherwise specified, we report unstandardized regression coefficients.

6. We counterbalanced the labeling of the groups (broad vs. deep) across the two experimental conditions (high vs. low identification) to preclude mere label effects.

7. We also conducted a mediation analysis with the identical pro-environmental behavior index as Study 1 (six pro-environmental behaviors). Results were similar, indicating that the indirect effect of identification on intention remained stable.