INTERPERSONAL RELATIONS AND GROUP PROCESSES Bodies Obliged and Unbound:Differentiated Response Tendencies for Injunctive and Descriptive Social Norms The d de cffortful sclf-rceul for Four expe in the t ?r ty t D.. Aevwords:social norms.social influcnce.conformity.self-regulation Initially within soc psychology.the concept of such as recycling (Schulz999).littering (Cialdini,Kallgren,& ocial norms has been rede ned in recent decades b tion (Goldstein,Cialdini,&Grisk Cialdini,GoldsteinGriskevicius,07)(see Bor .A ).and student gambling (LarimerNeighbors.2003) of Denver:R B.Cialdin Deparmen of Mar part.by ow at the Department of Org onal Studics Cialdini et).In brief,this theory emphasizes the need to n.Krista Ranby and A Reid for th reful en human goals (.g making accurate/efficient decisions and gain ning s his te ing/maintai of this the see Cialdini in NM 7131-00MC05 090.Univer type ar 43
INTERPERSONAL RELATIONS AND GROUP PROCESSES Bodies Obliged and Unbound: Differentiated Response Tendencies for Injunctive and Descriptive Social Norms Ryan P. Jacobson Arizona State University Chad R. Mortensen Metropolitan State College of Denver Robert B. Cialdini Arizona State University The authors suggest that injunctive and descriptive social norms engage different psychological response tendencies when made selectively salient. On the basis of suggestions derived from the focus theory of normative conduct and from consideration of the norms’ functions in social life, the authors hypothesized that the 2 norms would be cognitively associated with different goals, would lead individuals to focus on different aspects of self, and would stimulate different levels of conflict over conformity decisions. Additionally, a unique role for effortful self-regulation was hypothesized for each type of norm— used as a means to resist conformity to descriptive norms but as a means to facilitate conformity for injunctive norms. Four experiments supported these hypotheses. Experiment 1 demonstrated differences in the norms’ associations to the goals of making accurate/efficient decisions and gaining/maintaining social approval. Experiment 2 provided evidence that injunctive norms lead to a more interpersonally oriented form of self-awareness and to a greater feeling of conflict about conformity decisions than descriptive norms. In the final 2 experiments, conducted in the lab (Experiment 3) and in a naturalistic environment (Experiment 4), self-regulatory depletion decreased conformity to an injunctive norm (Experiments 3 and 4) and increased conformity to a descriptive norm (Experiment 4)— even though the norms advocated identical behaviors. By illustrating differentiated response tendencies for each type of social norm, this research provides new and converging support for the focus theory of normative conduct. Keywords: social norms, social influence, conformity, self-regulation Initially controversial within social psychology, the concept of social norms has been redeemed in recent decades by research demonstrating its power to explain and predict an array of different behaviors (Rivis & Sheeran, 2003; J. R. Smith & Louis, 2009), such as recycling (Schultz, 1999), littering (Cialdini, Kallgren, & Reno, 1991; Cialdini, Reno, & Kallgren, 1990; Kallgren, Reno, & Cialdini, 2000; Reno, Cialdini, & Kallgren, 1993), energy conservation (Goldstein, Cialdini, & Griskevicius, 2008; Schultz, Nolan, Cialdini, Goldstein, & Griskevicius, 2007), alcohol use (see Borsari & Carey, 2003, for a meta-analytic review), speech volume and table manners (Aarts & Dijksterhuis, 2003), tax evasion (Wenzel, 2004), and student gambling (Larimer & Neighbors, 2003). The resurgence of this concept has been stimulated, in part, by a set of key conceptual refinements comprised within Cialdini and colleagues’ focus theory of normative conduct (Cialdini, in press; Cialdini et al., 1990). In brief, this theory emphasizes the need to distinguish between descriptive and injunctive forms of normative information, which are differentially relevant to two fundamental human goals (e.g., making accurate/efficient decisions and gaining/maintaining social approval) and can have unique effects on behavior when made selectively focal. Although the evidence for the main tenets of this theory is compelling (see Cialdini, in press, for a review), the intervening psychological processes (e.g., cognitive, affective, and self-regulatory processes) involved in responding to each norm type are not fully understood. This article was published Online First December 20, 2010. Ryan P. Jacobson, Department of Psychology, Arizona State University; Chad R. Mortensen, Department of Psychology, Metropolitan State College of Denver; Robert B. Cialdini, Departments of Psychology and Marketing, Arizona State University. Ryan P. Jacobson is now at the Department of Organizational Studies, Anderson School of Management, University of New Mexico. We thank Kathryn Jacobson, Krista Ranby, and Allecia Reid for their helpful comments on an earlier version of this article, Amanda Jacobson and Tressa Mortensen for their careful editorial review of the document, Lynda Mae for permission to collect data in her classes, Uriah Anderson for his technical assistance in video formatting, and the dedicated team of undergraduate research assistants who made collection of these data possible. Correspondence concerning this article should be addressed to Ryan P. Jacobson, Anderson School of Management, MSC05 3090, University of New Mexico, Albuquerque, NM 87131-001. E-mail: rjacobso@mgt.unm.edu Journal of Personality and Social Psychology, 2011, Vol. 100, No. 3, 433– 448 © 2010 American Psychological Association 0022-3514/10/$12.00 DOI: 10.1037/a0021470 433 This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly
434 JACOBSON.MORTENSEN.AND CIALDINI egulation (i.e.,reduction in "will es the effective ndicating a high d f social in Cialdini et al. with the propose toward intraper ng) al effects of n duct and from functi ations of the aturalistic s,large body of fied-basedre h ha o m The Focus Theory of Normative Conduc f-regulatory (as well as bch vioral) ed the first typ nation the om research anc and the the injuncti vidence for f-reguionin other forms of Effortful Self-Regulation ularly relev Within the past o cades,a has nersonal goal of beha (see Bau neister Vohs,2004) 1998).The dich tion (1.e pprov blefordicer and a variety en ology of fo nflu ce also applie s to socia ms and nces u and predict thei he second pro on of focus theory that desc nts),these have e tha scular strength has tode cribe the nature of thi ted b nat:m sage fram purpose to that can b Ad ive nomms are often simulta alable given behav For example despite et al..2007).This provides evidence that the notion of effortfu er t (as element of the self's role in controlling behavior. ing both on the particular behavior and the Self-Regulation in Persuasion and Compliance penen que effec individuals'resp ses to persuasive me pliance techniques.Consistently,this research has supported the
In the present research, we advance such understanding by demonstrating that an impaired capacity for effortful selfregulation (i.e., reduction in “willpower”) increases the effectiveness of social influence attempts that highlight descriptive norms but decreases the effectiveness of attempts that highlight injunctive norms. Additionally, and consistent with the proposed conceptual rationale for this distinction, we also demonstrate differences (a) in the norms’ effects on the cognitive accessibility of goal-related concepts, (b) on attention directed toward intrapersonal versus interpersonal aspects of self, and (c) on the experience of conflict over decisions to follow or disregard the norm. Our theoretical argument for these differences, discussed subsequently, is derived both from implications of the focus theory of normative conduct and from functional considerations of the roles of effortful self-regulation and normative influence in social life. The Focus Theory of Normative Conduct An initial observation underlying focus theory was that the general term norm, both in common and scientific usage, has often been used to refer to two different types of social information (Schaffer, 1983). Cialdini and colleagues (Cialdini et al., 1991, 1990) labeled the first type of information the descriptive norm, which refers to the perceived prevalence or typicality of a given behavior, and the second type the injunctive norm, which refers to the perceived degree of social approval/disapproval for the behavior. According to the first of two main propositions of the theory, distinguishing between these two types of information is crucial because each is highly relevant for a different fundamental human goal. Whereas the descriptive norm provides information that is particularly relevant for the intrapersonal goal of behaving effectively or accurately (i.e., choosing correctly), the injunctive norm is particularly relevant for the interpersonal goal of building and maintaining social relationships (i.e., social approval; Cialdini & Trost, 1998). The distinction between these two goals as uniquely responsible for different forms of social influence has a relatively long tradition in social psychology (e.g., Deutsch & Gerard, 1955). A key insight of focus theory is that this general principle of goal-directed influence also applies to social norms and can enhance the ability of researchers to understand and predict their effects on behavior. The second proposition of focus theory is that descriptive or injunctive norms will affect an individual’s behavior only to the extent that each form of information is currently salient—a state that can be affected by both situational (e.g., message framing) and dispositional factors (e.g., strong personal endorsement of a particular norm). This addresses a key challenge for understanding the unique effects of these two types of normative information—the fact that both descriptive and injunctive norms are often simultaneously available for any given behavior. For example, despite general social disapproval for littering, the descriptive evidence in a given setting can indicate that this behavior is actually quite common. Furthermore, these norms can either be incongruent (as in the prior example) or congruent (e.g., littering disapproved and uncommon)— depending both on the particular behavior and the particular context in question. Despite this complexity, it remains possible to examine the unique effects of each norm type because these are determined by the norm’s subjective salience rather than by its objective availability in the situation. As an illustration, in an experiment conducted in a heavily littered parking garage (i.e., descriptive norm indicating a high prevalence of littering), Cialdini and colleagues (Cialdini et al., 1990, Study 4) demonstrated that the presence of a competing injunctive norm against littering (e.g., the litter was swept into large piles) tended to reduce littering behavior— but only when that norm was made salient by the presence of a confederate littering directly in front of the swept litter piles. Thus, in some situations, procedures that selectively heighten the salience of a descriptive or an injunctive norm can lead to distinctly different behavioral responses (i.e., increased vs. decreased littering). Focusing primarily on the behavioral effects of norm salience in naturalistic environments, a large body of field-based research has supported these two main propositions of the theory (see Cialdini, in press, for a review). However, implications of the theory for differences in the intervening psychological responses to the two norm types have remained largely unexplored. The present investigation targets this gap—testing predicted differences in the cognitive, affective, and self-regulatory (as well as behavioral) responses to descriptive and injunctive social norms in lab-based and naturalistic environments. In addition to the implications of the focus theory, our predictions were derived from research and theory focused on self-regulation as an effortful capacity, prior evidence for self-regulation in other forms of social influence and persuasion, and from functional considerations of the roles of the two types of normative information in social life. Effortful Self-Regulation Within the past two decades, a sizable literature has focused on the manner in which the self exerts control over inner states and overt behaviors (see Baumeister & Vohs, 2004). Evidence suggests that effortful self-regulation is involved in (among other processes and behaviors) impulse restraint, decision making, emotion regulation, thought suppression, and a variety of different kinds of physical and mental task performances (Baumeister, Bratslavsky, Muraven, & Tice, 1998; Eisenberg, Smith, Sadovsky, & Spinrad, 2004; Vohs et al., 2008). Despite clear differences in the processes/behaviors involving effortful self-regulation (e.g., differences in cognitive requirements, role of affect, physical requirements), these have been shown to rely on a general resource that is limited but renewable (Baumeister et al., 1998). A metaphor of muscular strength has often been used to describe the nature of this self-regulatory capacity—as a general purpose tool that can be temporarily “fatigued” and then regenerated through inactivity or “rest” (Muraven & Baumeister, 2000; Schmeichel & Baumeister, 2004). Additionally, the consumption of glucose has been shown to attenuate the depleting effects of self-regulatory tasks (Gailliot et al., 2007). This provides evidence that the notion of effortful self-regulation as akin to strength has a tangible, physiological basis—as well as having illustrative value in describing a key element of the self’s role in controlling behavior. Self-Regulation in Persuasion and Compliance Effortful self-regulation has also been shown to play a role in individuals’ responses to persuasive messages and specific compliance techniques. Consistently, this research has supported the 434 JACOBSON, MORTENSEN, AND CIALDINI This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly
RESPONSES TO INJUNCTIVE AND DESCRIPTIVE NORMS 435 des or b 20070Wh self-regulatory depletion on respons tory effort directed toward resisting the powerful cious.This led them to conclude tha Heathe 1Q9R1 20081 similar result for the effect of depletion on lation often involves ific fomm of that pits the e unrelated task. ister.2005 Baumeister.DeWall.Ci con ing compliance with a request: footin-thoor an esigned to indu liking for the requestor.As with a numbe self-regulatory consequences of social exclusion ing on individuals 1 life eed to strated a reduced capacity for self-regulation follo talsaeeofaootin eofthesesciimpuesndmakccctanohe self-r had be who wer disr ow in the city for the basi of this ev hors self- depl hich inc the he rait of self-m itoring which focuses on the tend to contro ar forms of soc This evidence for the role of effortful self-regulation draw on the capacity fo vior.If soci so pitalize on heuristi However.there o ons d lete al.. 06 same ng mechani nd.thu osocial standards,being rejected or excluded by vn to redu asks that we actually dence links i impaired f-regu on to a variety ofne ent to hh the s (e.g erspective on the role of self-regulatory processes in social life A Complementary Role for Injunctive Norms Functional Perspective on Self-Regulation o promote re this c cating social standards for dilemmas (Baumeister Vohs.2007).For example.one's immediate bargain of social life.As theorists have argued (e.g..Bicchieri
notion that self-regulatory resources are involved in resisting attempts to change attitudes or behavior. Wheeler and colleagues (Wheeler, Brin˜ol, & Hermann, 2007) investigated the effects of self-regulatory depletion on responses to counterattitudinal messages, demonstrating increased persuasion— especially when message arguments were specious. This led them to conclude that depletion had reduced participants’ abilities to generate counterarguments to the message, thus increasing persuasion. Burkley (2008) reported similar results for the effect of depletion on persuasion while also demonstrating that exposure to a persuasive message impaired subsequent self-regulation performance on an unrelated task. Fennis, Janssen, and Vohs (2009) investigated the role of selfregulation in four social influence techniques geared toward inducing compliance with a request: foot-in-the-door (Freedman & Fraser, 1966), door-in-the-face (Cialdini et al., 1975), low-ball (Cialdini, Cacioppo, Bassett, & Miller, 1978), and an approach designed to induce liking for the requestor. As with a number of other tactics for inducing compliance, these operate in part by capitalizing on individuals’ tendencies to rely on decision-making heuristics (e.g., consistency, reciprocity, and liking; Cialdini & Goldstein, 2004). In their research, Fennis and colleagues demonstrated a reduced capacity for self-regulation following exposure to the initial stage of a foot-in-the-door technique, increased compliance with door-in-the-face and liking-based techniques among participants whose self-regulatory capacities had been depleted, and higher compliance with a low-ball technique among participants who were dispositionally low in the capacity for effortful self-regulation. On the basis of this evidence, the authors concluded that these four compliance techniques affect behavior by inducing self-regulatory depletion—which increases the heuristic responding tendencies that underlie these particular forms of social influence. This evidence for the role of effortful self-regulation in persuasion and compliance provides a clear reason to expect that impairment of this capacity should increase the degree to which social norms affect behavior. If social norms also capitalize on heuristic processing tendencies, then self-regulatory depletion should amplify their effects by decreasing the ability to carefully consider behavioral alternatives. However, there is reason to expect that descriptive and injunctive norms do not both influence behavior via this same underlying mechanism, and, thus, self-regulatory depletion should not have the same effects on the two forms of normative social influence. Although we do expect self-regulatory depletion to increase conformity to descriptive norms, we actually expect it to have the opposite effect for injunctive norms— decreasing the extent to which they influence behavior. This expectation is derived both from a closer consideration of the goaldirected nature of normative social influence and from a functional perspective on the role of self-regulatory processes in social life. Functional Perspective on Self-Regulation The fundamental function of self-regulation is to alter behavior, and this often requires the suppression of impulses that threaten to undermine desired long-term goals (Baumeister, Schmeichel, & Vohs, 2007; Baumeister & Vohs, 2007). Thus, conflicts between alternative motivations serve as a central feature underlying most self-regulatory dilemmas (Baumeister & Vohs, 2007). For example, one’s immediate impulse is to devour the leftover slice of decadently frosted double chocolate birthday cake, but this motive conflicts with a long-term desire to appear slim and attractive. Thus, maintaining one’s diet requires self-regulatory effort directed toward resisting the powerful and immediate cake craving. As a number of researchers have asserted (e.g., Heatherton & Vohs, 1998; Polivy, 1998), social standards (explicit social pressures or internalized social standards) play a key role in stimulating the need for self-regulation of behavior. Thus, self-regulation often involves a specific form of conflict that pits the desire for social approval against self-interested motives that are more independent and immediate in nature. In fact, Baumeister and colleagues (Baumeister, 2005; Baumeister, DeWall, Ciarocco, & Twenge, 2005) have argued that resolving this form of conflict serves as the fundamental role for self-regulation in social life—as a capacity that allows restraint of immediate self-interests that could interfere with the long-term benefits of cooperative social behaviors. As these authors (Baumeister et al., 2005) explained in their research on the self-regulatory consequences of social exclusion: The broader context for this investigation is the view that human social life contains an implicit bargain. People have a strong need to belong, and they survive, flourish, and reproduce by means of inclusion; however, they also have naturally selfish impulses that create conflict and friction among neighbors. If people are to live together, they must curb some of these selfish impulses and make certain other sacrifices. (p. 590) A wide range of research is consistent with the notion that self-regulation is essential in helping the individual negotiate conflicts between an interpersonal drive for social approval and more intrapersonally driven (i.e., “selfish”) motives. For example, the trait of self-monitoring, which focuses on the tendency to control one’s behavior in social settings, is positively related to selfregulatory performance (Seeley & Gardner, 2003). Additionally, self-presentational behaviors draw on the general capacity for effortful self-regulation (Vohs, Baumeister, & Ciarocco, 2005), and self-regulatory depletion reduces the likelihood of costly forms of helping that require the suppression of immediate selfinterests (DeWall, Baumeister, Gailliot, & Maner, 2008). Furthermore, engaging in difficult or “high maintenance” social interactions depletes self-regulatory capacity (Finkel et al., 2006). Perhaps because effortful self-regulation is so centrally linked to adhering to social standards, being rejected or excluded by a group has also been shown to reduce performance on tasks that require self-regulation (Baumeister et al., 2005). Finally, a wealth of evidence links impaired self-regulation to a variety of negative social outcomes (e.g., destructive relationship behavior, expressions of prejudice, psychopathology; Finkel & Campbell, 2001; Monteith, 1993; Tangney, Baumeister, & Boone, 2004). A Complementary Role for Injunctive Norms An understanding of this important function of self-regulation—to promote resolution of conflicts between interests of the independent self and the social collective—suggests that this capacity is likely to be involved in facilitating the influence of injunctive social norms. In communicating social standards for behavior, injunctive norms are also clearly relevant to Baumeister and colleagues’ (Baumeister et al., 2005) notion of the implicit bargain of social life. As theorists have argued (e.g., Bicchieri, RESPONSES TO INJUNCTIVE AND DESCRIPTIVE NORMS 435 This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly
436 JACOBSON.MORTENSEN.AND CIALDINI side well-kn age from the nove T regulation in responding to descriptive norms is likely to be similar d h Aunt Poll puckish om of his tic cu for behavior ance using deserintive nomms can at times lead demo towar orite band must co ial inf the mind-numbingly unp to provide the f self-regulatory effort towa dering th onl goalsin the g the hen and to straint is necessary.In this m may not s omple ents in address chal engesof the ended Fridav after degree of conformity to descriptive social norms Implications for Cognition and Affect contrary to what may be Consideration of these different roles for injunctive and descrip Self-Regulation and Descriptive Norms racy/and the goaof social approval be rast,the inforr ion communicated by descriptivenorms eleva ct b what is typi ontrast t motives)For n about the es to the intrapersona sts that the choice of that favor hasa high-calori also might be expected to have different conse onal aspects of self.To the extent that injunctive nomms en socially beneficial be r,they shoul ie. ntrapersonal (i.e.,ma ng acd ate/effi )rather than mula on mo the ice icius Gold use of descriptiveso cial informat tion as the decision-making heu ristic of"social proof,"suggesting that it serves as a time-and s“injunctive norm:
2006; Lahti & Weinstein, 2005),1 responding to injunctive norms often involves the experience of dual (and dueling) motives. As an illustration, consider a well-known passage from the novel, The Adventures of Tom Sawyer. In the passage, puckish Tom is assigned to whitewash Aunt Polly’s front fence as a penance for one of his many transgressions. On the sun-drenched Saturday morning in which he is to serve his sentence, Tom must decide between behavior prescribed by a salient injunctive norm (i.e., he “should” respect his guardian Aunt Polly and dutifully whitewash the fence) and what, at the moment, he might rather do (e.g., play pirates on an abandoned, ramshackle riverboat). More mundane examples demonstrating the relevance of this conflict to injunctive norms are pervasive in daily social life: Just available front row tickets to our favorite band must compete with a spouse’s reminder that we are obligated to dine with our ill-tempered in-laws. Although we might love to spend a beautiful Friday afternoon at the beach or the golf course, a colleague’s e-mail message reminds us that we really ought to endure the mind-numbingly unproductive, multihour department meeting. Although self-regulatory capacity appears to provide the “strength” necessary to restrain drives toward immediate intrapersonal goals in these cases, a critical role for injunctive norms is to indicate when and to what extent such restraint is necessary. In this manner, injunctive norms and effortful self-regulation can often serve as functional complements in addressing challenges of the implicit bargain of social life—increasing the likelihood of uncomfortable dinners with the in-laws, well-attended Friday afternoon department meetings, whitewashed fences, and a variety of other behaviors that are socially, if not personally, desirable. Thus, contrary to what may be expected from prior research on persuasion and compliance, self-regulation may actually serve to facilitate rather than restrict the influence of injunctive norms. Self-Regulation and Descriptive Norms In contrast, the information communicated by descriptive norms appears less relevant to this conflict between immediate intrapersonal impulses and long-term interpersonal goals. By communicating what is typical in a given context, descriptive norms provide heuristic cues for behaviors that are likely to be desirable or immediately advantageous for the individual (i.e., aligned with intrapersonal motives). For example, out of countless flavors of ice cream, evidence that a certain variety is a shop’s “best seller” suggests that the choice of that flavor has a particularly strong likelihood of satisfying one’s hedonistic craving for a high-calorie, high-fat dairy product. Even though this information is derived from social behavior and injunctive inferences can be made about the possible approval or disapproval associated with those actions, this information is most clearly and most immediately relevant to intrapersonal (i.e., making accurate/efficient decisions) rather than interpersonal goals (Cialdini & Trost, 1998). Supporting this contention, and consistent with the ice cream example, individuals tend to use descriptive information to guide behavior especially in situations that are novel, ambiguous, uncertain, or threatening (Deutsch & Gerard, 1955; Griskevicius, Goldstein, Mortensen, Cialdini, & Kenrick, 2006; Sherif, 1936; Tesser, Campbell, & Mickler, 1983). Cialdini (2009) has described this use of descriptive social information as the decision-making heuristic of “social proof,” suggesting that it serves as a time- and effort-saving shortcut that can promote effective behavior across a wide variety of circumstances. Thus, the role of effortful selfregulation in responding to descriptive norms is likely to be similar as to that demonstrated for persuasive messages and compliance techniques—as a means to override the tendency to use this information as a heuristic cue for behavior. Although following descriptive norms may most often lead to behaviors that are more accurate or efficient than would have been possible without this information, this is not universally true in all cases. As is demonstrated by the phenomenon of pluralistic ignorance, using descriptive norms as behavioral cues can at times lead to very negative outcomes, like an increased tendency toward binge drinking (e.g., Prentice & Miller, 1993) or a decreased tendency to escape emergency situations involving personal peril (e.g., Latane´ & Darley, 1968). Additionally, Berger and Rand (2008) have provided evidence that social influence attempts using descriptive norms can actually decrease (rather than increase) compliance when the descriptive norm is associated with an undesirable group. This implies that individuals tend to exert at least a limited degree of self-regulatory effort toward considering the implications of the descriptive norm and possible behavioral alternatives— helping to counteract the possibility that following the norm may not be consistent with self-interest. Given this tendency, impairing the ability to engage in such effortful cognitive activity (via self-regulatory depletion) should be expected to increase the degree of conformity to descriptive social norms. Implications for Cognition and Affect Consideration of these different roles for injunctive and descriptive norms in social life also leads to the expectation of differences in the typical cognitive and affective responses to the norms. The salience of an injunctive norm should lead individuals to consider their behavior as it relates to both the intrapersonal goal of accuracy/efficiency and the interpersonal goal of social approval because it suggests the possibility of a situation in which social standards may not be aligned with immediate self-interests. In contrast, thoughts associated with social approval should not necessarily be stimulated by the salience of a descriptive norm. Rather, as a cue for accurate or efficient behavior, the descriptive norm should lead individuals to consider this information about the behavior’s prevalence primarily as it relates to the intrapersonal goal of choosing correctly. The norms also might be expected to have different consequences for attention directed toward intrapersonal versus interpersonal aspects of self. To the extent that injunctive norms encourage cooperative and socially beneficial behaviors, they should stimulate a comparatively greater interpersonal focus of attention (i.e., focus on group memberships and social relationships). In contrast, descriptive norms should stimulate a focus on more intrapersonal aspects of self because they are most immediately relevant to intrapersonal rather than interpersonal goals. Finally, the salience of dual and potentially dueling goals for an injunctive norm should lead individuals to experience conflict over decisions 1 Bicchieri (2006) uses the term social norms, and Lahti and Weinstein (2005) use the term moral norms to refer to the category of norms we label as “injunctive norms.” 436 JACOBSON, MORTENSEN, AND CIALDINI This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly
RESPONSES TO INJUNCTIVE AND DESCRIPTIVE NORMS 437 to conform or not conform.Conversely,the descriptive norm drew on as ily relevan ry orga antic representati cluding goal The Present research these norms inm Evide point for the present s regardless of the actual beh sed to the norm A body o the norm (a)that goals and motiv were involve soclgnal-diected Aarts Trot Aarts 2007 for a review)It has even heen shown that this Sheeran et al task that sublimina (e.g. ro&Bargh 2009 s between the two types of the t oals on the cial envi ing relat 0gmemseg,lo red specch volume:Aarts apersonal goal of behaving accuratel that he ory respons nost co In co ocial life.Thus,interestingly,very subtle differences in no ase the framing may have sfor individu Thus.to paire goal words (a val)and e esponse (ie and.ult onsequences Method e and injuncti on cognit ted to ment 2.we xperience of dec making contl tcdhtowardaclasr ttentic control task to ulate sclf-res ulator ight partici who Word om on and goal ta tion of the third experiment,but we tive word primes we stream.rect ring rather than timulus for sel Experiment 1 mmon/prevalent (e.g typical and social approval.On the basis of the sus Thr reg the imary god d to the intrape f the ole of each no roval (e.g both ro-related matched in length with the target norm words were also included
to conform or not conform. Conversely, the descriptive norm should not stimulate such conflict because it is primarily relevant to a single intrapersonal goal (i.e., choosing correctly). The Present Research A critical point for the present research is our contention that these somewhat unique forms of responding may be associated with the salience of the two norms regardless of the actual behavior advocated by the norm. In other words, we predicted that these different types of response tendencies would be engaged for the norms even in cases for which the same behaviors were involved. A spate of evidence has demonstrated that goal-directed behavior, including goal-directed responses to social norms (e.g., Aarts & Dijksterhuis, 2003), can be initiated automatically in response to salient cues in the environment (see Dijksterhuis, Chartrand, & Aarts, 2007, for a review). It has even been shown that this nonconscious initiation of goal-directed responding can lead to effortful self-regulatory responses (e.g., Ackerman, Goldstein, Shapiro, & Bargh, 2009). Just as cues for particular social environments (e.g., library) can engage tendencies toward behaviors that are habitually associated with those environments (e.g., lowered speech volume; Aarts & Dijksterhuis, 2003), we suggest that descriptive or injunctive frames for a behavior can lead to response tendencies that reflect the cognitive, affective, and self-regulatory responses most commonly associated with these forms of normative information in social life. Thus, interestingly, very subtle differences in normative framing may have relatively major consequences for individual decision making—altering the evaluation of a behavior’s relevance for the self and stimulating somewhat different emotional and self-regulatory responses. We report four experiments in which we tested hypothesized differences in the cognitive, affective, self-regulatory, and, ultimately, behavioral consequences of descriptive and injunctive norm activation. In Experiment 1, we examined the effects of descriptive and injunctive norm salience on cognitions related to the goals of accuracy/efficiency and social approval. In Experiment 2, we explored differences in self-awareness and in the experience of decision-making conflict that result from descriptive versus injunctive framing of a behavioral norm. Using a lab-based attention control task to manipulate self-regulatory capacity, we examined in Experiment 3 the effects of depletion on conformity to descriptive versus injunctive norms. Finally, Experiment 4 provided a conceptual replication of the third experiment, but we used a naturally occurring rather than lab-based stimulus for selfregulatory depletion. Experiment 1 In Experiment 1, we examined the extent to which the two types of norms are cognitively associated with the fundamental goals of accuracy/efficiency and social approval. On the basis of the suggestions of focus theory regarding the primary goals driving conformity for each type of norm and on a functional analysis of the role of each norm in social life, we hypothesized that descriptive information would be associated with accuracy/efficiency-related concepts and that injunctive information would be associated with both accuracy/efficiency and social approval-related concepts. We drew on associative network and connectionist principles of memory organization (see E. R. Smith, 1998, for a review) to test this hypothesis, reasoning that the activation of norm-related concepts should spread to semantic representations (including goal- and motivation-related concepts) that are most closely associated with these norms in memory. Evidence for semantic associations between particular norms and particular goal concepts should therefore reflect the types of goal/motive considerations that are most likely to be present after being exposed to the norm. A body of research demonstrating (a) that goals and motive states can be automatically activated by environmental cues and (b) that this nonconscious activation can lead to behaviors aligned with those goals/motives provides strong evidence to support this rationale (e.g., Aarts et al., 2005; Aarts, Custers, & Holland, 2007; Bargh, Gollwitzer, Lee Chai, Barndollar, & Trötschel, 2001; Fitzsimons & Bargh, 2003; Moskowitz, Li, & Kirk, 2004; Sheeran et al., 2005). We designed a lexical decision task that involved subliminal priming with norm-related words to explore the cognitive associations between the two types of norms and the two goals. On the basis of our theory, priming with words related to descriptive norms (e.g., typical, widespread) should increase the accessibility of words related to the intrapersonal goal of behaving accurately/ efficiently (e.g., accurate, beneficial) but should not increase the accessibility of words related to the interpersonal goal of gaining/ maintaining social approval (e.g., approval, dependent). In contrast, priming with words related to injunctive norms (e.g., ought, responsibility) should increase the accessibility of both types of goal-related words. Thus, to test this hypothesis, we paired descriptive-related, injunctive-related, and neutral primes with target goal words (accuracy/efficiency and social approval) and examined the extent to which the norm primes decreased or increased response times for word/nonword decisions (i.e., facilitated or suppressed decision-making responses). Method Participants. Participants were 87 students (49 women, 38 men) from introductory psychology classes who received partial credit toward a class research requirement for their involvement in the experiment. Experiment sessions included groups of two to eight participants who were seated at private computer workstations and visually shielded from one another by partitions. Word primes and goal targets. Three categories of normative word primes were created. Neutral primes (e.g., stream, rectangle, fence, trees) reflected common nonsocial words selected on the basis of their lack of norm-related connotations. Consistent with the suggestions of focus theory regarding the essential information communicated by each type of norm, descriptive norm primes reflected the concept of common/prevalent (e.g., typical, widespread, usual, popular), and injunctive norm primes reflected the concept of social obligation (e.g., should, ought, responsibility, duty). Three categories of target words were also created. These included the focal categories of words related to the intrapersonal goal of accuracy/efficiency (e.g., accurate, beneficial, desire, efficient) and the interpersonal goal of social approval (e.g., approval, others, dependent, team). In keeping with the aims of a lexical decision task, nonwords (e.g., kourge, shrouthen, dwex, knurdge) matched in length with the target norm words were also included. RESPONSES TO INJUNCTIVE AND DESCRIPTIVE NORMS 437 This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly
JACOBSON.MORTENSEN.AND CIALDINI le whether a string of letters re ented a meaningful of word prime appro hed significance in this key on t 166 60,p1 m2= nd tare vords chosen at tandom without re lacen ded norms word prime that ms.Prev s research (e.g.. 997 uracy/efficiency goa d les the ately fo owing the masking stimulu n.ared unti tion effect for both ti ccuracv efed foll ving completion of the ta Results e in co nparison to nding on the he two classes of goal words. The mean for Discussion sequent analyses (76%of resp The results of Experiment 1 support our that indi es).Trials in onwords were not included in Nomainorinteraction of particp the a began by submitting the data to a 3(word prime:descriptive norm. predicted that the descriptive norm primes would increase re and Standard Deviations for Prime-Targe Trial Categories i Experiment2 Neutral prime Target word (goals) SD SD SD y 728594 6m3;1809%68947769
Procedure. Participants received instructions to complete a lexical decision task. Their objective was to indicate, as quickly as possible, whether a string of letters represented a meaningful English word (by pressing the 5 key on their keyboard) or represented a nonword (by pressing the A key). All participants completed 72 trials, made up of 36 words (18 words related to the goal of accuracy/efficiency and 18 words related to the goal of social approval) and 36 filler nonwords. Within each of the three target word categories, individual trials were paired in equal numbers with the three types of word primes. For each trial, word primes and target words/nonwords were chosen at random without replacement from within each of the categories. Thus, nine distinct types of prime–target pairs were presented to each participant. All task stimuli were presented and response times recorded using DirectRT v2006 software (Jarvis, 2006a). Each trial started with a fixation point (“”) presented on the center of the computer screen for 1,000 ms. This was immediately followed by a masking stimulus (“X”) that appeared for 70 ms, a word prime that appeared for 35 ms, and a masking stimulus that again appeared for 70 ms. Previous research (e.g., Wittenbrink, Judd, & Park, 1997) has indicated that prime words are detectable but not identifiable at similar stimulus intervals. Target letter sequences were presented immediately following the masking stimulus and appeared until participants made a word/nonword decision. A 1,500-ms pause separated each decision from the beginning of the next trial. All participants were fully debriefed following completion of the task. Results Initial considerations. We excluded three participants with accuracy below 60% in identifying words/nonwords (three standard deviations below the mean) due to random responding on the task. The mean percentage of correct responses for the 84 remaining participants was 95%. We excluded all incorrect lexical decisions from subsequent analyses (7.6% of responses). Additionally, based on considerations for response time data discussed by Bargh and Chartrand (2000), we excluded decisions with response times less than (i.e., faster than) 300 ms and those greater than (i.e., slower than) three standard deviations above the mean (2.3% of responses). Trials involving nonwords were not included in subsequent analyses. In this and all subsequent experiments, we examined the data for possible gender effects before testing our primary hypotheses. No main or interaction effects of participant gender were found in Experiment 1 or in subsequent experiments. Thus, this variable was not included in tests of our primary hypotheses. Hypothesis tests. Mean response times and standard deviations for each prime-target category are presented in Table 1. We began by submitting the data to a 3 (word prime: descriptive norm, injunctive norm, neutral) 2 (goal target: accuracy/efficiency, social approval) within-subjects analysis of variance (ANOVA). The main effect of word prime approached significance in this analysis, F(2, 166) 2.60, p .078, p 2 .030, and the main effect of goal target was significant, F(2, 166) 75.82, p .001, p 2 .477. However, as predicted, these effects were qualified by a significant interaction between word prime and goal target, F(2, 166) 17.20, p .001, p 2 .172. We then proceeded to a set of planned comparisons designed to compare response times for goal targets in the neutral prime condition against response times for those targets in the descriptive and injunctive norm conditions (i.e., the comparisons were focused on the facilitative/suppressive effects of the normative primes on response times). On the basis of theory, priming with words related to descriptive norms should facilitate response times for words related to an accuracy/efficiency goal but should not facilitate response times for words related to a social approval goal. Supporting these predictions, when preceded by descriptive norm primes, participants responded more quickly (in comparison to neutral primes) to words related to the accuracy/efficiency goal, F(1, 83) 12.77, p .001, p 2 .133. In contrast, participants actually responded less quickly to words related to the social approval goal when these were preceded by descriptive norm primes, F(1, 83) 6.78, p .011, p 2 .076. For injunctive norms, we hypothesized a facilitation effect for both the accuracy/ efficiency and social approval classes of goal words. Supporting this prediction, participants responded more quickly to goal targets when these had been preceded by an injunctive in comparison to a neutral prime, F(1, 83) 8.28, p .005, p 2 .091. Figure 1 depicts the facilitative/suppressive effects of the norm primes for the two classes of goal words. Discussion The results of Experiment 1 support our contention that individuals associate descriptive information primarily with the goal of accuracy/efficiency but associate injunctive information with the dual goals of accuracy/efficiency and social approval. Following subliminal presentation of norm-related words, word/nonword responses for the goal of accuracy/efficiency were faster than control in both the descriptive and injunctive norm conditions. For the goal of social approval, responses were slower than control in the descriptive norm condition and faster than control in the injunctive norm condition. Thus, thoughts about making accurate and efficient decisions appear particularly likely following the salience of a descriptive norm, whereas thoughts about both accuracy/ efficiency and social approval are likely to follow from the salience of an injunctive norm. Although we had not explicitly predicted that the descriptive norm primes would increase reTable 1 Means and Standard Deviations for Prime-Target Trial Categories in Experiment 2 Target word (goals) Descriptive norm prime Injunctive norm prime Neutral prime M SD M SD M SD Accuracy/efficiency 607.28 159.14 663.53 141.02 696.47 216.56 Social approval 721.97 224.84 618.35 100.76 659.94 170.12 438 JACOBSON, MORTENSEN, AND CIALDINI This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
RESPONSES TO INJUNCTIVE AND DESCRIPTIVE NORMS 439 Target Word (Goal) Metho Participants and design. Participants were 80 students (49 xneriment Experiment sessions included roups of two to eight participants o were seated at privat 32.9 .41.50 (nom type:injunctive or descriptive) Intrapersonal versus interpersonal self-awareness. This 983D Injunctive ences and asked to c al prime (negative value ngular pronc (eg. n plura e.g po pronou consistentith the idea that descriptivenorms are per Four self-re sider tems ofa potentialy competin but not partic ularly relevant,goal. d三 about the Experiment 2 g the tu-iem he po Experiment 2 had two primary objectives.First.we sought Clark. evalid.then ught tha as well activating som fest itself as n the cTmtgal-relaticdconcepts.Pa PANAS-e( 10040 additionl six-item subscale measuring the emotic of self decreased confidence or boldness)than those expe nt predicted differ conflict.For all PAN Amiresponded-point of norms Because info ion stimulates MediaLab s ing ks. ot the d ld lend that th degree of decision-mak ichen the beha ywould have the optior advocate the same behavior (in this case.volunteering to complete the form of independentt-mnsurveys that were part of an
sponse times for social approval-related concepts (we had only predicted that they would not facilitate these responses), this remains consistent with the idea that descriptive norms are perceived as relevant to intrapersonal goals and not necessarily to interpersonal goals. Potentially, a dissociation between social approval concepts and descriptive norms could make these norms more effective by decreasing the likelihood that individuals consider the norm in terms of a potentially competing, but not particularly relevant, goal. Experiment 2 Experiment 2 had two primary objectives. First, we sought additional evidence to support our contention that the two norms lead individuals to focus on somewhat different goals. If our hypotheses regarding the norms’ cognitive effects are valid, then one might also expect exposure to the norms to stimulate somewhat different forms of self-awareness as well as activating somewhat different goal-related concepts. Particularly, to the extent that injunctive but not descriptive norms engage a concern with social approval as well as with accuracy/efficiency, they should lead to a greater focus on interpersonal aspects of self (e.g., group memberships, social roles, interpersonal relationships) than descriptive norms. Thus, evidence to support this proposition would provide converging support for the theorized effects of the norms on goal-related cognition. Second, we also tested in this experiment predicted differences in the affective experiences that follow from exposure to the two types of norms. Because injunctive information stimulates thoughts about dual/dueling goals but descriptive information does not, the decision to follow or disregard the norm should lead to a greater degree of decision-making conflict when the behavior is framed as an injunctive norm than as a descriptive norm. As we have argued, this should be the case even when the two norms advocate the same behavior (in this case, volunteering to complete extra surveys) due to differences in the response tendencies associated with these two normative frames. Method Participants and design. Participants were 80 students (49 women, 31 men) from introductory psychology classes who received partial credit toward a class research requirement for their involvement in the experiment. Experiment sessions included groups of two to eight participants who were seated at private computer workstations and visually shielded from one another by partitions. The experiment was a single factor, two-cell, betweensubjects design (norm type: injunctive or descriptive). Measures. Intrapersonal versus interpersonal self-awareness. This construct was measured using Wegner and Giuliano’s (1980, 1983) pronoun selection task. For this widely used measure (e.g., Silvia, Phillips, Baumgaertner, & Maschauer, 2006; Stephenson & Wicklund, 1984), participants are given 20 fill-in-the-blank sentences and asked to complete each by choosing one of three pronouns. These include first-person singular pronouns (e.g., I, me, my, or mine) that reflect a more intrapersonal focus and third-person singular/plural (e.g., hers, them) or first-person plural (e.g., we, our) pronouns that reflect a more interpersonal focus. An example item reads, “Someone stopped (them, me, us) to get directions to the stadium.” The interpersonally oriented pronoun selections (20 possible) were summed to provide an overall score. Conflict over the conformity decision. Four self-report indices were used to measure the level of conflict participants experienced when deciding to conform/nonconform following the normative message. For the primary measure, participants were asked directly about their experience of conflict over the decision. Participants responded to this item, “Right now, I feel conflicted about the decision to either complete or not complete extra surveys” on a 5-point scale ranging from 1 (strongly disagree) to 5 (strongly agree). In addition, the two 10-item subscales of the Positive and Negative Affect Schedule (PANAS; Watson, Clark, & Tellegen, 1988) were used to create indices of general positive and negative emotionality. Although these measures offered somewhat less direct assessments of the phenomena of interest, it was thought that conflict over the conformity/nonconformity decision might manifest itself as lower positivity and/or higher negativity on these general emotional indices. Three items were also included from the larger PANAS-X scale (Watson & Clark, 1994) to provide an additional six-item subscale measuring the emotion of selfassurance (i.e., these were added to three existing self-assurance items from the PANAS). Particularly, it was thought that individuals experiencing a high degree of conflict over the conformity decision would be likely to report feeling less self-assured (i.e., decreased confidence or boldness) than those experiencing less conflict. For all PANAS items, participants responded on a 7-point scale ranging from 1 (disagree) to 7 (agree). Procedure. MediaLab software (Jarvis, 2006b) was used to administer all instructions, tasks, and measures. Participants first received information that the primary experimental session tended to end about 20 –25 min early and that they would have the option of either leaving the lab at that time or staying to complete extra questionnaires. It was explained that the questionnaires would take the form of independent 1- to 2-min surveys that were part of an -89.18 -32.94 62.85 -41.59 -100 -75 -50 -25 0 25 50 75 100 Descriptive Injunctive Accuracy/Efficiency Social Approval Target Word (Goal) Response Time Facilitation / Suppresion (inms) Figure 1. Facilitation/suppression of target word response times as a function of norm type in Experiment 2. Values reflect deviations (in milliseconds) from the neutral prime condition (negative values indicate response time facilitation). RESPONSES TO INJUNCTIVE AND DESCRIPTIVE NORMS 439 This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly
JACOBSON.MORTENSEN.AND CIALDINI of the primary study.Immdiately following by the survey questions be nted to limit the extent to area might e and a full deb might be strongly ciated withex ut the pursuit or avoidance of certain eha Results and Discussion artic We began by submitting the dependent measures ton st the likely to choos oriented pro shaveended,most havec orm co ndition.F(.78)832. 005. T2=096.A1s with predictions,participants eportedrelinelesself& this manipulation. participant emo n in the (1980.1983) then read a 、oned fecing marganallyiton c ptive 60, for the n by informing parti ants that the uppo that point study exposed to a descript nine out rsonal goals (e.g soci he veys while n out o ent read "p nine out of ten former pa n the orms' er the decision to c han when it had been 1994 cales that o assess the degree of 5000 we hav is supported by s likely that this is the result of differe in the r In this m that they hac hesefor behavior Descriptive (n 41) Injunctive (n 39) SD SD 1,78) 039 296 12.00 189 8.32 005 elf-A 87
ongoing and unrelated study that had begun in the previous semester. No information was given regarding the topics addressed by the survey questions because we wanted to limit the extent to which a particular topic area might evoke preexisting norms that could interfere with or attenuate our norm manipulation (e.g., health-related questions might be strongly associated with existing injunctive norms about the pursuit or avoidance of certain behaviors). Participants were told that they would be asked to indicate their choice regarding the surveys after all primary experimental tasks were complete and that no penalties or extra rewards would be associated with their decision. Participants in the descriptive norm condition then read the following: “In past instances in which study sessions have ended early, most students have chosen to stay for the full hour and complete extra surveys.” In the injunctive norm condition, participants instead read, “In a survey conducted last semester, most students indicated that, in instances in which study sessions end early, they felt that participants should be willing to stay for the full hour and complete extra surveys.” Immediately following this manipulation, participants completed the measure of intrapersonal versus interpersonal selfawareness (Wegner & Giuliano, 1980, 1983). They then read a short message designed to serve as a “booster shot” for the initial norm manipulation, used due to the presence of multiple dependent measures. This message began by informing participants that they had completed roughly one third of the study to that point in the session and reminded them that they would be given the option to complete extra surveys at the end of the primary study. In the descriptive norm condition, the message concluded with the following statement, “Roughly nine out of ten former participants have decided to complete the optional surveys while one out of ten has decided to leave early.” In the injunctive norm condition, this concluding statement read, “Roughly nine out of ten former participants indicated that they thought others should stay to complete the optional surveys while one out of ten has thought this isn’t necessary.” After this norm booster shot, participants completed items from the PANAS (Watson et al., 1988) and PANAS-X (Watson & Clark, 1994) scales that were included to assess the degree of conflict participants experienced over the conformity/nonconformity decision. Participants then received a second reminder message. In this message, participants were informed that they had completed two thirds of the study to that point in the session and were reminded of the option to complete extra surveys at the end of the primary study. Immediately following this second reminder message, participants responded to the single-item measure of decision conflict. The session concluded with a funneled suspicion probe and a full debriefing. No extra surveys were actually administered to participants. Results and Discussion We began by submitting the dependent measures to one-way ANOVAs, with norm type as the independent variable. Table 2 summarizes these analyses. Consistent with predictions, participants were more likely to choose interpersonally oriented pronouns in the injunctive norm condition than in the descriptive norm condition, F(1, 78) 8.32, p .005, p 2 .096. Also consistent with predictions, participants in the injunctive norm condition reported experiencing a greater degree of conflict over the conformity decision, F(1, 78) 4.39, p .039, p 2 .053, and reported feeling less self-assured, F(1, 78) 4.50, p .037, p 2 .055, than in the descriptive norm condition. Participants also reported feeling marginally less positive emotion in the injunctive than in the descriptive condition, F(1, 78) 2.60, p .111, p 2 .032, but did not report feeling greater negative emotion (F 1). The results of this experiment provide additional evidence to support our claims regarding the norms’ effects on goal-related cognition. Participants exposed to an injunctive norm were more likely to direct attention toward interpersonal aspects of self than those exposed to a descriptive norm. This supports the idea that the two norms lead individuals to focus on somewhat different sets of goals—with interpersonal goals (e.g., social approval) being more salient after exposure to injunctive norms than descriptive norms. Critically, this experiment also provides evidence for differences in the norms’ effects on the experience of conflict over the conformity/nonconformity decision. Supporting our theoretical argument, participants reported experiencing a greater degree of conflict over the decision to complete extra surveys when it had been framed using an injunctive norm (most think you should do these) than when it had been framed using a descriptive norm (most have chosen to do these). As we have argued, and as is supported by evidence from both the present experiment and Experiment 1, it seems likely that this is the result of differences in the primary goals that underlie conformity to each type of norm and in the differential presence of secondary goals that can compete with these for control over behavior. Table 2 Means, Standard Deviations, and Analysis of Variance Results for Dependent Variables in Experiment 2 Descriptive (n 41) Injunctive (n 39) M SD M SD F(1, 78) p p 2 Interpersonal self-awareness 10.39 2.96 12.00 1.89 8.32 .005 .096 Conflict over conformity decision Single item 2.24 1.14 2.80 1.22 4.39 .039 .053 Self-Assurance (PANAS) 4.30 1.04 3.76 1.22 4.50 .037 .055 General Positivity (PANAS) 4.42 0.99 4.04 1.13 2.60 .111 .032 General Negativity (PANAS) 2.18 0.91 2.21 0.86 0.03 .871 .000 Note. PANAS Positive and Negative Affect Schedule. 440 JACOBSON, MORTENSEN, AND CIALDINI This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
RESPONSES TO INJUNCTIVE AND DESCRIPTIVE NORMS 44 Experiment 3 With evidence from the first two experiments supporting our ide ationalcfordi crences in the Participants then received information about completing extra ment 3 to examining the role of effortful self-regulation in re mmediately rather than after finishing a series of other tasks italizing on tendencies e (Burkle vill you complete?"Afte veys they would agree to complete,par e the full debriefing No xtra surveys were actually administered to ected self-regulator y depletion to have the oppositee ause njunctiv ehaneoeteonh Results and Discussion ng goals of accuracy/eff and socia As expected,a significant interaction was obtained between val)actua g the 4sondniol 5.39, pein吗goa igni t (Fs 24).Also s expected simple effects test Method 115 ur the mM=422 hen sel redit toward a cl h reau nt fo beir in en ns ed grou ed at dhigh-fregu her by partition 691a8 vith the inants woule whe depleted o experin ntal instruct e or negativ participants by the by the intera irst part of the rved as a m and Krull 1988)0 trated that willful control of atten ofa ion alse 2007 In the task nartici ants viewed a siler t yideo clin of a w g intervie vords (e. disk th might h tely woma app nstructions regarding these words.As a check on possible mood on.we opted against its inclusion in the experimer
Experiment 3 With evidence from the first two experiments supporting our rationale for differences in the cognitive and affective reactions to injunctive and descriptive norms, we shifted our focus in Experiment 3 to examining the role of effortful self-regulation in responding to the norms. Specifically, we manipulated the capacity for effortful self-regulation through use of an attention control task and then measured conformity to a descriptive or an injunctive norm for completing optional surveys. If one were to assume that both norm types influence behavior by capitalizing on tendencies toward heuristic responding, then, as with persuasion (Burkley, 2008; Wheeler et al., 2007) and several compliance techniques (Fennis et al., 2009), depletion should be expected to increase the effectiveness of both norms. However, we have made this assumption only for descriptive norms. For injunctive norms, we expected self-regulatory depletion to have the opposite effect— decreasing conformity. On the basis of our theory, this is because injunctive norms create a conflict between the dual/ dueling goals of accuracy/efficiency and social approval. Thus, following an injunctive norm (and pursuing the goal of social approval) actually requires self-control in the form of resisting the competing goal of accuracy/efficiency. Depleting individuals of the capacity for such resistance should therefore decrease the effectiveness of the injunctive norm. Method Participants. Participants were 116 students (61 women, 55 men) from introductory psychology classes who received partial credit toward a class research requirement for their involvement in the experiment. Experiment sessions included groups of two to six participants who were seated at private computer workstations and visually shielded from one another by partitions. Design and procedure. The design was a 2 (self-regulatory capacity: low or high) 2 (norm: injunctive or descriptive) between-subjects factorial.2 All experimental instructions, tasks, and measures were administered via computer using MediaLab v2006.1.10 experiment delivery software (Jarvis, 2006b). In the first part of the experiment, participants were asked to engage in what was described as an “impression formation” task, which actually served as a manipulation of self-regulatory capacity. Gilbert, Pelham, and Krull (1988) originally developed this task as a manipulation of attention control. However, subsequent research has demonstrated that willful control of attention also requires self-regulatory capacity, and, thus, this manipulation can be used as an effective stimulus for self-regulatory depletion (e.g., DeWall, Baumeister, Stillman, & Gailliot, 2007; Gailliot et al., 2007; Schmeichel, Vohs, & Baumeister, 2003; Vohs et al., 2005). In the task, participants viewed a silent video clip of a woman being interviewed and were instructed to form an impression of her on the basis of her nonverbal behaviors. Throughout the approximately 4-min interview clip, neutral words (e.g., disk, tire, pole) appeared in the bottom right-hand corner of participants’ screens. In the low-self-regulatory capacity condition, participants were asked to avoid reading these words and, if distracted, to immediately redirect their attention to the woman. In the high-selfregulatory capacity condition, participants received no explicit instructions regarding these words. As a check on possible moodrelated effects of the self-regulation task, participants completed the PANAS scale (Watson et al., 1988) immediately following the video clip. Participants then received information about completing extra surveys that was consistent in all respects to the scenario presented in Experiment 2 except that they expected to make their decision immediately rather than after finishing a series of other tasks. Directly following the delivery of the descriptive or injunctive norm, all participants received the dependent measure, which took the form of the following question: “You can choose to complete up to 10 extra surveys. Each short survey takes about 1–2 minutes to complete. How many extra surveys will you complete?” After indicating how many surveys they would agree to complete, participants responded to a funneled suspicion probe and received a full debriefing. No extra surveys were actually administered to participants. Results and Discussion As expected, a significant interaction was obtained between self-regulation condition and norm type, F(1, 112) 5.39, p .022, p 2 .046 (see Figure 2). Neither of the main effects was significant (Fs 2.4). Also as expected, simple effects tests revealed that those exposed to the injunctive norm agreed to complete significantly fewer of the optional surveys when selfregulatory capacity was low (M 2.21) than high (M 4.84), F(1, 112) 7.73, p .006, p 2 .065. Furthermore, and also consistent with our theory, the injunctive norm (M 2.21) was less effective than the descriptive norm (M 4.22) when selfregulatory capacity was low, F(1, 112) 4.23, p .042, p 2 .036. Although we found no significant difference between lowand high-self-regulation conditions for participants exposed to the descriptive norm, the pattern of means (Mlow 4.22; Mhigh 3.69) was consistent with the expectation that participants would be more likely to follow the descriptive norm when depleted of their capacity to self-regulate. The number of surveys participants volunteered to complete was not predicted by positive or negative affect as measured by the PANAS scale, nor by the interaction between affect and norm type, thus reducing the possibility that confounding mood effects were responsible for differences in conformity levels. 2 We decided against the inclusion of a no-norm control condition based in part on Aarts and Dijksterhuis’ (2003) evidence that certain social situations (e.g., a library context) can automatically evoke cognitive (e.g., heightened accessibility of the concept “silence”) and behavioral (e.g., decreased voice intensity) responses that are aligned with norms associated with those situations. We thought it likely that participants would have existing conceptions of norms associated either with the laboratory context or with the scenario of volunteering for optional tasks and that these conceptions could influence behavior in the absence of an explicit norm manipulation. Particularly problematic for a “control” condition, we thought it unlikely that the normative associations with an optional survey scenario would be as uniformly similar across participants as one might expect for a library scenario (i.e., completing extra surveys might be perceived by some as typical and by others as atypical; by some as moderately approved and by others as strongly approved). Thus, without the ability to actually control normative associations with the scenario in a no-norm condition, we opted against its inclusion in the experiment. RESPONSES TO INJUNCTIVE AND DESCRIPTIVE NORMS 441 This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
JACOBSON.MORTENSEN.AND CIALDINI 86 Norm Type related conformity in a more naturalistic en 4 方3 o for.our laboratory-base High ow Self-Regulatory Resource Capacity incre of s The results of expe ment 3 demonstrated that an rkday).Thus,we sought to explore th This esponse a the same mec as demonstrated Io onin uristie respo d haveincreased rather than ereased ontexts tha Finally,as with the cognitive and affective responses examined namely,that clas activities have been used successfully as was obtained even when the injunctive and descriptive noms for the experiment also provided tw Thus the nommative framing of the behavior as somethin by che side the lab cte e norm und sibe that this and computer-driven instructions to re racting words)B possibility that we tested in Exper ment Second,using a naturally occu Experiment 4 uscdcoatrocg of the norms on goal-related cognition.(b) ppear at the bottom right-hand comer of their screen as the opcuo o s of rms.In ate reactance against the norm and
The results of Experiment 3 demonstrated that an experimentally induced depletion of self-regulation capacity decreased the influence of an injunctive but not a descriptive norm on behavior. This supports our theorized role for self-regulation in individuals’ responses to injunctive norms—as a means to resist drives toward intrapersonal goals that can interfere with norm-consistent behavior. Additionally, these results suggest that injunctive norms do not influence behavior via the same mechanism as demonstrated for persuasion and multiple forms of compliance (i.e., capitalizing on heuristic response tendencies). If this were the case, then depletion should have increased rather than decreased conformity to the injunctive norm. Finally, as with the cognitive and affective responses examined in Experiment 2, it is important to note that the selective decrease in the influence of the injunctive norm in the depletion condition was obtained even when the injunctive and descriptive norms advocated the same behavior (i.e., completing extra surveys). Thus, the normative framing of the behavior (i.e., as something one should do versus something many people do) rather than a particular quality of the behavior itself was responsible for differences in the effectiveness of the norms. Although the pattern of means was consistent with predictions, the expected increase in conformity to the descriptive norm in the depletion condition was not significant. Although this may suggest that self-regulation does not serve as a means to restrict or override heuristic responding to a descriptive norm, it is also possible that this failure may be a consequence of a relatively weak manipulation of self-regulation depletion (i.e., a brief video clip and computer-driven instructions to resist distracting words)—a possibility that we tested in Experiment 4. Experiment 4 To this point in our research program, we had used controlled laboratory designs to provide evidence for (a) differential effects of the norms on goal-related cognition, (b) differential effects of the norms on the experience of conflict over the conformity decision, and (c) differential effects of self-regulatory depletion on the likelihood of conforming to the two types of norms. In our view, the critical next step, consistent with a full-cycle approach to social psychology (Cialdini, 1980; Mortensen & Cialdini, 2010), was to examine the effects of self-regulatory depletion on normrelated conformity in a more naturalistic environment. In addition to providing increased confidence in the ecological validity of our research, we viewed this as an opportunity to emphasize the real-world implications of our recent findings. We thought this focus on behavioral consequences would be especially valuable given the burgeoning field of social norms marketing (e.g., Moreira, Smith, & Foxcroft, 2009; Perkins & Berkowitz, 1986) and the important implications of this phenomenon for applied practitioners. Particularly, our laboratory-based results suggest that certain conditions might lead to “boomerang” effects of such social norms approaches—reductions rather than increases in conformity. For example, conformity to an injunctive norm could be attenuated if influence targets possess particularly low acute levels of self-regulatory resources (e.g., college students currently under the influence of alcohol or other drugs) and conformity to a descriptive norm could be attenuated if targets possess particularly high acute levels of such resources (e.g., employees just beginning a new workday). Thus, we sought to explore the generalizability of the phenomenon by using a common setting outside the laboratory that would involve a more complex (and realistic) array of variables. Fortunately for us, the university classroom provided an easily accessible social environment for which depletion is a predictable consequence of normal classroom behaviors involving effortful self-regulation (e.g., focusing and controlling attention, engaging in logical reasoning, regulating emotions, suppressing irrelevant thoughts, presenting the self strategically). In this way, the classroom is similar to many real-world occupational contexts that involve highly effortful responding. The choice of a classroom as the setting for our experiment also had another advantage— namely, that classroom activities have been used successfully in past research as naturally occurring manipulations of selfregulatory resources (e.g., Gailliot et al., 2007). Our choice of setting for the experiment also provided two additional benefits. First, by choosing to apply our methodology outside the laboratory, we were able to capitalize on a naturally occurring stimulus for self-regulatory depletion that we felt would be much stronger than the manipulation used in Experiment 3. Recall that in that experiment, we did not obtain the expected increase in conformity to the descriptive norm under conditions of depletion. We suggested that this might have been the consequence of a relatively weak depletion manipulation (i.e., a brief video clip and computer-driven instructions to resist distracting words). Because we expected the activities involved in the classroom setting to produce a much higher level of self-regulatory depletion than our laboratory manipulation, the use of this setting allowed us to evaluate the viability of this as a possible explanation. Second, using a naturally occurring stimulus for depletion also provided an opportunity to rule out a potential alternative explanation for the results of Experiment 3. In Experiment 3, participants in the low-self-regulatory capacity condition were instructed to avoid attending to a series of phrases that would periodically appear at the bottom right-hand corner of their screen as they watched a video clip. One might argue that these instructions could be seen as an initial attempt to influence behavior. If, subsequently, a norm were to be perceived as a second attempt to influence behavior, then this could stimulate reactance against the norm and Figure 2. Number of extra surveys participants agreed to complete as a function of norm type and self-regulation capacity in Experiment 3. 442 JACOBSON, MORTENSEN, AND CIALDINI This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly
