Red Sneaker Effect

8/12/2019 Red Sneaker Effect

1/21

Journal of Consumer Research Inc.

The Red Sneakers Effect: Inferring Status and Competence from Signals of NonconformityAuthor(s): Silvia Bellezza, Francesca Gino, and Anat KeinanSource: Journal of Consumer Research, (-Not available-), p. 000Published by: The University of Chicago PressStable URL: http://www.jstor.org/stable/10.1086/674870.

Accessed: 21/02/2014 20:39

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at.http://www.jstor.org/page/info/about/policies/terms.jsp

.JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of

content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms

of scholarship. For more information about JSTOR, please contact [email protected].

.

The University of Chicago PressandJournal of Consumer Research, Inc.are collaborating with JSTOR to

digitize, preserve and extend access toJournal of Consumer Research.

http://www.jstor.org
http://www.jstor.org/action/showPublisher?publisherCode=ucpresshttp://www.jstor.org/stable/10.1086/674870?origin=JSTOR-pdfhttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/stable/10.1086/674870?origin=JSTOR-pdfhttp://www.jstor.org/action/showPublisher?publisherCode=ucpress


2/21

000

2013 by JOURNAL OF CONSUMER RESEARCH, Inc. Vol. 41 June 2014

All rights reserved. 0093-5301/2014/4101-0003$10.00. DOI: 10.1086/674870

Please use DOI when citing. Page numbers are not final.

The Red Sneakers Effect: Inferring Statusand Competence from Signals of

Nonconformity

SILVIA BELLEZZAFRANCESCA GINOANAT KEINAN

This research examines how people react to nonconforming behaviors, such asentering a luxury boutique wearing gym clothes rather than an elegant outfit orwearing red sneakers in a professional setting. Nonconformingbehaviors, as costly

and visible signals, can act as a particular form of conspicuous consumption andlead to positive inferences of status and competence in the eyes of others. A seriesof studies demonstrates that people confer higher status and competence to non-conforming rather than conforming individuals. These positive inferences derivedfrom signals of nonconformity are mediated by perceived autonomy and moderatedby individual differences in need for uniqueness in the observers. An investigationof boundary conditions demonstrates that the positive inferences disappear whenthe observer is unfamiliar with the environment, when the nonconforming behavioris depicted as unintentional, and in the absence of expected norms and sharedstandards of formal conduct.

Your sweats, PJs and flip-flops are losing youmoney! . . . Do you crave more confidence,

respect and power? . . . Find out how imageconnects to success. (Eve Michaels, author ofDress Code)

I have a number of super-successful SiliconValley clients who dress in ripped denim, Vansshoes, and T-shirts. They are worth hundredsof millions, even more, but its a status symbolto dress like youre homeless to attend boardmeetings. (Tom Searcy, CBS Moneywatch)

Silvia Bellezza ([email protected]) is a doctoral candidate in mar-

keting, Francesca Gino([email protected])is associate professor of business

administration, and Anat Keinan([email protected])is associate professor

of marketing at the Harvard Business School, Boston, MA 02163. The

authors are grateful for helpful comments received from participants in

seminars and lab groups at Harvard and MIT. They thank John T. Gourville,

Max H. Bazerman, Zoe Chance, Michael I. Norton, and Natalie Zakarian

for their constructive feedback on earlier drafts of the article. The article

is based on part of the first authors dissertation.

Mary Frances Luce served as editor and Rebecca Ratner served as as-

sociate editor for this article.

Electronically published December 18, 2013

I n both professional and nonprofessional settings, indi-viduals often make a significant effort to learn and adhereto dress codes, etiquette, and other written and unwrittenstandards of behavior. Conformity to such rules and socialnorms is driven by a desire to gain social acceptance andstatus (see Cialdini and Goldstein 2004) and avoid negativesanctions such as social disapproval, ridicule, and exclusion(Kruglanski and Webster 1991; Levine 1989; Miller andAnderson 1979; Schachter 1951). In the present research,we propose that under certain conditions, nonconformingbehaviors can be more beneficial than efforts to conformand can signal higher status and competence to others. Weargue that while unintentional violations of normative codesand etiquette can indeed result in negative inferences andattributions, when the deviant behavior appears to be delib-erate, it can lead to higher rather than lower status and

competence inferences.Since nonconformity often has a social cost (e.g., Levine

1989; Schachter 1951), observers may infer that a noncon-forming individual is in a powerful position that allows herto risk the social costs of nonconformity without fear of losingher place in the social hierarchy. Signaling theory suggeststhat, for a signal to be effective, it must be costly and ob-servable by others (Feltovich, Harbaugh, and To 2002; Spence1973; Zahavi and Zahavi 1997). We propose that noncon-forming behaviors, as costly and observable signals, can act

This content downloaded from 132.235.61.22 on Fri, 21 Feb 2014 20:39:53 PMAll use subject to JSTOR Terms and Conditions
mailto:[email protected]:[email protected]:[email protected]://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jspmailto:[email protected]:[email protected]:[email protected]


3/21

000 JOURNAL OF CONSUMER RESEARCH


as a particular form of conspicuous consumption and lead toinferences of status and competence by observers. Such pos-itive inferences are consistent with Veblens classic theory ofconspicuous consumption (1899/1994), which suggests thatindividuals display status through the prominent, visible evi-dence of their ability to afford luxury goods. Similarly, weargue that nonconformity can lead to inferences of higherstatus and greater competence by providing visible evidencethat individuals can afford to follow their own volition. Basedon some of our experimental stimuli for nonconformity, welabel this potential positive outcome of nonconforming be-haviors the red sneakers effect.

As a preliminary test, we first explore the relationship be-tween nonconformity and status in the field by examining thedress style of conference participants and their professionalstatus. Next, five lab and field studies explore how noncon-forming behavior is perceived by others. In particular, whendo people interpret nonconformity as a signal of status andcompetence, and what are the processes underlying such in-

ferences? Our studies explore various consumption environ-ments and populations, including shop assistants at high-endboutiques, business executives, and college students.

Our investigation of psychological processes reveals thatinferences of status and competence derived from signalsof nonconformity are mediated by perceived autonomy. Wedemonstrate that nonconformity can fuel perceptions ofstatus and competence in the eyes of others because devi-ating from the norm signals that one has the autonomyneeded to act according to ones own inclinations and tobear the cost of nonconformity. Moreover, we show that therelationship between a persons nonconforming behaviorand observers perceptions of enhanced status and compe-

tence is moderated by observers need for uniqueness (Sny-der and Fromkin 1977), such that observers with high levelsof need for uniqueness tend to confer greater status andcompetence to nonconforming behaviors as compared toobservers with low needs for uniqueness. We further in-vestigate boundary conditions of the effect by manipulatingand measuring additional characteristics of the observers,the environment, and the nonconforming behavior.

Our research contributes to the conspicuous consumptionliterature and to research on nonconformity. First, we extendconsumer behavior research analyzing alternative and coun-terintuitive ways to display status, such as using less recog-nizable but more expensive luxury brands and products or

smaller logos (Berger and Ward 2010; Han, Nunes, and Dreze2010). Specifically, we investigate a different kind of con-sumer behavior and an alternative way of displaying status(e.g., violating a dress code rather than buying subtly brandedbut expensive luxury products). Second, in contrast to mostnonconformity research, which has focused on nonconform-ing individuals and the antecedents for their behavior, wefocus on the consequences of nonconformity and the percep-tions of external observers. Importantly, we concentrate onnonconformity-based inferences of status and competence.

THEORETICAL FOUNDATIONS

Society has powerful formal and informal mechanismsthat motivate individuals to conform to social norms andexpectations regarding appropriate conduct. While confor-mity is rewarded with group acceptance and social inclusion

(see Cialdini and Goldstein 2004), nonconformity can berisky and costly, often leading to social disapproval, rejec-tion, and punishment (Anderson et al. 2006, 2008; Levine1989; Lin et al. 2013; Marques et al. 2001; Miller and An-derson 1979; Schachter 1951; Wilson 1979). The power ofthese rewards and sanctions has been demonstrated in classicsocial psychology experiments. For example, in Aschs(1956) well-known studies examining the conformity of judg-ments and opinions in groups, participants often conformedbecause it was easier to follow the crowd than to face theconsequences of going against it (Crutchfield 1955). Morepowerful and disturbing evidence comes from Zimbardos(1973) prison experiment, in which volunteers who wererandomly assigned the roles of guards or prisoners be-

haved accordingly, and Milgrams (1963) obedience exper-iments, which demonstrated that people readily conform tothe social roles they are expected to play.

In the context of consumer behavior, research demon-strates that assimilation and conformity motives can driveconsumption practices and choices in the marketplace. Con-sumers are motivated to behave like those around them andmake choices that are consistent with their in-group due toa need to increase affiliation and express desired identities(Bearden, Netemeyer, and Teel 1989; Escalas and Bettman2003, 2005; McFerran et al. 2010a, 2010b). In particular,individuals who feel socially excluded and lonely are morelikely to conform in an effort to avoid the negative evalu-ations of others by selecting products endorsed by most

consumers (Mead et al. 2011; Wang, Zhu, and Shiv 2012).Given these powerful social mechanisms, researchers

across disciplinesincluding sociology, social psychology,economics, and marketinghave devoted their attention tothe study of nonconformity and its antecedents. Noncon-formity is generally defined as a behavior or belief that isinconsistent with norms or standards (Nail, Macdonald, andLevy 2000). In the consumer psychology literature, the ten-dency to engage in nonconforming consumption has beenassociated with a desire to distance the self from dissimilar,disliked, or unattractive others or from out-group members(Berger and Heath 2007, 2008; White and Dahl 2006, 2007)or to establish ones uniqueness and distinctiveness (Arielyand Levav 2000; Griskevicius et al. 2006; Simonson and

Nowlis 2000; Snyder and Fromkin 1977).While this literature has focused primarily on the ante-

cedents of the nonconforming individuals behaviors, in ourwork we shift the focus of analysis to how external observersperceive and interpret nonconforming behaviors in terms ofstatus and competence. That is, rather than examining in-dividuals decisions to conform or not conform, we examinethe consequences of deviating from the norm in the eyes ofothers. A vast body of research on impression formationand status beliefs suggests that individuals rapidly make

http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp


4/21

BELLEZZA, GINO, AND KEINAN 000


inferences and judgments of others competence and statusbased on observable signals, such as appearance, verbal andnonverbal behaviors, attitudes, and consumption choices(Ambady and Rosenthal 1993; Dubois et al. 2012; Hall etal. 2005; Knapp et al. 2009; Magee 2009; Ridgeway andCorrell 2006; Ridgeway et al. 1998; Schmid Mast and Hall

2004; Todorov et al. 2005). We contribute to this literatureby investigating the conditions under which lay observersmakenonconformity-based inferences of targets economic,professional, and social status.

Nonconformity, Status, and Competence

As compared to low-status individuals, high-status indi-viduals have wider latitude for deviation and are relativelyfree from social constraints (Feshbach 1967; Hollander 1958;Peterson and Kern 1996; Phillips and Zuckerman 2001). Agroup member can be said to earn and maintain increasedstatus through idiosyncratic credits, or an accumulationof positive impressions in the minds of the rest of the group

(Hollander 1958). This accumulation is reflected in the de-gree to which the individual can deviate from group normswithout sanction. Thus, unlike low-status group members,high-status members and powerful individuals can afford todeviate from conventional behavior and common expecta-tions without social disapproval (Cartwright 1959; Galinskyet al. 2008; Haslam 2004; Sherif and Sherif 1964).

More specifically, in the domain of consumption, high-status individuals may voluntarily downgrade their lifestyleand adopt nonconforming consumption habits, such as ma-terial frugality, omnivoreness (consuming a broad rangeof products), and simplicity (Arnould and Thompson 2005;Brooks 1981; Holt 1998; Peterson and Kern 1996; Solomon1999). For example, high-status individuals may choose to

dress informally in business settings. Certain CEOs of majorcorporations, including Microsofts Bill Gates and Face-books Mark Zuckerberg, have been known to appear with-out ties or even wearing sweatshirts at interviews and formalgatherings such as the World Economic Forum (Etzioni2004); some successful entrepreneurs have a habit of at-tending their companies board meetings in casual dress,such as jeans and sneakers (Searcy 2011).

To provide empirical evidence of the relationship betweennonconformity to dress codes and status and competence,we conducted a pilot observational study examining thepotential relationship between the dress style of participantsin a professional academic conference and the number ofarticles they had published. In this pilot study, we focus on

the link between actual status and nonconforming behavior;in our remaining studies, we focus on the link between non-conforming behavior and observers perceptions of status andcompetence. We predicted that conference participants whohad gained greater status through research productivitywould dress more casually than other participants. Partici-pants were 76 randomly selected attendees of the 2011 As-sociation for Consumer Research conference. We recordedthe names of participants, as indicated on their name tags,and coded the formality of their dress. To code dress for-

mality objectively, we created a composite score rangingfrom a minimum of 0 points (less formal) to a maximumof 4 points (more formal) by discretely coding four clothingelements worn by each participant. More specifically, weadopted the following scoring system: 1 point for wearinga blazer, 0 points otherwise; 1 point for wearing a button-

down shirt or a dress, 0 points otherwise (e.g., for a T-shirt);1 point for wearing formal pants, 0 points otherwise (e.g.,for jeans); 1 point for wearing dress shoes, 0 points oth-erwise (e.g., for sneakers). We then collected data on thenumber of publications for each scholar in our sample (basedon information available online) as an objective measure ofstatus and competence in the academic community. To con-trol for gender and age, we also coded these variables (mea-suring age by years since PhD graduation) and includedthem in our regression analysis. We counted participantstotal number of peer-reviewed publications in academic

journals and their number of publications in top marketingjournals for consumer behavior scholars, namely, the Jour-nal of Consumer Research, the Journal of Marketing Re-

search, Marketing Science, the Journal of Marketing, Sci-ence,Psychological Science, and theJournal of Personalityand Social Psychology. We found that a less-conformingdress style was significantly correlated with research pro-ductivity (rp .35,p !.01). Interestingly, this correlationwas even stronger when focusing on publications in topmarketing journals (r p .53, p ! .001). These results,which hold even when controlling for gender and age,1 in-dicate that higher status and performance within a givencommunity is correlated with a stronger tendency to deviatefrom a conforming dress code (e.g., wearing jeans, sneakers,T-shirts rather than professional attire). Although these re-sults are correlational, they are consistent with Hollanders(1958) theoretical account that high-status and well-re-

spected individuals tend to engage in nonconforming be-haviors. But are these nonconforming behaviors actuallyinterpreted as a signal of status and competence by third-party observers? And if so, when and why does this happen?

We propose that nonconforming behavior can act as aparticular form of conspicuous consumption and lead toperceptions of enhanced status and competence in the eyesof others. Observers may infer that a nonconforming indi-vidual is in a more powerful position that allows her tofollow her volition in autonomy and bear the cost of de-viating from the norm. Research suggests that high-statusindividuals tend to avoid blatant and conspicuous displaysof wealth, status, or personal accomplishments, and insteadseek alternative ways to differentiate themselves from lower-status individuals (Berger and Ward 2010; Feltovich et al.2002; Han et al. 2010). For example, sophisticated luxuryconsumers elect to use less known and less conspicuous

1 In a linear regression, formality score was regressed on number of

top publications, gender (dummy coded 1 for male, 0 for female), and age.

The final model was significant (R2 p .29, F(3, 67) p 8.9, p ! .001).

There was a significant effect of number of top publications (bpublications p

.47, t(67) p 2.8, p ! .01), whereas the other independent variables

were not significant (bgender p .11, NS; bage p .02, NS).



5/21



luxury brands. We investigate the conditions under whichnonconforming behaviors, such as entering a luxury bou-tique wearing gym clothes rather than an elegant outfit orwearing red sneakers in a professional setting, can serve asan alternative, nonconventional form of conspicuous con-sumption. In line with research on status beliefs and im-

pression formation (Hollander 1958; Ridgeway and Correll2006; Ridgeway et al. 1998), we examine inferences of bothstatus and competence. Status is defined as a higher positioncompared to others on some dimension (wealth, hierarchy,etc.), and it relates to the respect one has in the eyes ofothers (Magee and Galinsky 2008). Competence refers tothe perceived ability to successfully pursue and performspecific tasks (Fiske et al. 2002).

We argue that inferences of greater status and competencefrom nonconforming behavior result from observers attri-butions of the nonconforming individuals autonomy. Au-tonomy refers to self-governance and self-regulation (Ryanand Lynch 1989). The central idea in the concept of auton-omy is indicated by the etymology of the term: autos(self)

andnomos(rule or law). Thus autonomous individuals tendto act independently and behave according to their ownrules. The significance of freedom and autonomy is builtinto the founding documents of the United States, and theidea that individuals are independent and autonomous ispervasive and generally greatly admired (Dworkin 1988;Markus and Schwartz 2010). Especially in Western culturesthat place high value on individualism and independence,resisting group pressure can be perceived as a brave andbold gesture (Baumeister 1982; Galinsky et al. 2008; Kimand Markus 1999). Here, we suggest that nonconformitycan be perceived as admirable behavior that reflects highlevels of autonomy and control. While being easily influ-enced by others is not an admired personal trait (see Jetten,

Hornsey, and Adarves-Yorno 2006), deviating from thenorm signals freedom and autonomy from the pressure toconform (Phillips and Zuckerman 2001; Thompson et al.2006) and thus can fuel positive inferences in the eyes ofothers (Van Kleef et al. 2011; Simonson and Nowlis 2000).Hence, we predict that observers will infer enhanced statusand competence from signals of nonconformity because theybelieve that nonconforming individuals have greater auton-omy to act according to their own volition and bear the costof nonconforming:

H1: Nonconforming behavior can lead to greater in-ferences of status and competence as compared toconforming behavior.

H2: Positive inferences of status and competence fromnonconforming behavior will be mediated by ob-servers attributions of the nonconforming indi-viduals autonomy.

Characteristics of the Observers. We investigate the im-pact of individual differences in observers need for unique-ness (Nail et al. 2000; Snyder and Fromkin 1977) on howthey interpret signals of nonconformity. Consumers need for

uniqueness reflects individual differences in motivations fordistinguishing the self via consumer goods that manifest thewillful pursuit of differentness relative to others (Tian, Bear-den, and Hunter 2001). Individuals with a high level of needfor uniqueness are particularly sensitive to the degree to whichthey are seen as similar to others and are more likely than

others to exhibit behaviors that establish a sense of special-ness, such as acquiring unique or scarce products (Snyder1992).

Previous research on uniqueness motive demonstrates itsimpact on consumption choices and behavior (Ariely andLevav 2000; Chan et al. 2012; Cheema and Kaikati 2010;Irmak et al. 2010; Lynn and Harris 1997; Maimaran andWheeler 2008; Ratner and Kahn 2002; Simonson and Nowlis2000; White and Argo 2011). For example, consumers witha high level of need for uniqueness tend to prefer objects thatdeviate from norms over those that comply with norms (Lynnand Harris 1997; Snyder and Fromkin 1977; Tian et al. 2001);these consumers often demonstrate nonconforming prefer-ences in group contexts to distinguish themselves from others

(Ariely and Levav 2000). Our studies measure observersneed for uniqueness and examine how such uniqueness mo-tives impact the inferences observers make about a noncon-forming individual. That is, rather than examining how needfor uniqueness impacts consumers decision to conform ornot conform, we examine how it impacts the inferences theymake about other individuals who deviate.

We hypothesize that individual differences related to needfor uniqueness, as measured through the Tian et al. (2001)scale, will moderate the red sneakers effect. Research onuniqueness emphasizes that people who score relatively highon need for uniqueness often deviate from the norm in orderto assert their differentness, affirm strong character and, thus,enhance their social-image (Gross 1977; Tian et al. 2001).

Moreover, high need-for-uniqueness individuals are rela-tively free from social constraints imparted by others andexhibit high levels of autonomy (Snyder and Fromkin 1977).Accordingly, we expect those individuals who score highon need for uniqueness to associate (as external observers)higher autonomy, status, and competence with nonconform-ing behaviors relative to individuals with low need foruniqueness. Therefore, we predict that:

H3: Positive inferences of status and competence andattributions of autonomy from nonconforming be-havior will be moderated by observers need foruniqueness; observers with high levels of need foruniqueness will infer more status, competence,

and autonomy from nonconforming rather thanconforming behavior as compared to observerswith low levels of need for uniqueness.

We also examine observers familiarity with the environ-ment and consumption context as an important boundarycondition for the red sneakers effect. Familiarity with thecontext is critical in verifying and understanding the signalof nonconformity (Hollander 1958). In the consumer be-havior domain, individuals who are familiar with the context



6/21



are capable of finer, more articulated discrimination in thatspecific environment compared to those who are not andthey rely less on the ownership of prototypical product sym-bols (Solomon 1999). Accordingly, we predict that onlyobservers who are familiar with the consumption contextand have experience observing and interpreting individu-

als behavior in this specific context will infer greater statusand competence from signals of nonconformity rather thanfrom signals of conformity. For example, in the setting ofour observational study, conference participants may inferthat their colleagues are violating the dress code with theircasual attire, but observers who are not part of this specificcommunity or who have no past experience with it willnot make such attributions. Instead, they might infer thatnonconforming individuals are unaware of the dress codeor cannot afford nicer clothes. Thus, inferences of statusand competence require the observer to be somewhat fa-miliar with the environment. We therefore predict that:

H4: Nonconforming behavior will be interpreted as a

positive signal of status and competence as longas the observer is familiar with the environment.

Characteristics of the Environment. Nonconformity sig-nals can only occur in environments that have strong normsand shared standards, with a social expectation of confor-mity to these norms (Walker and Heyns 1962). For example,formal and prestigious contexts (e.g., a black-tie event, abusiness meeting) are typically characterized by an officialdress code or behavioral etiquette. Accordingly, in all ourexperiments we examine prestige settings with expected normsof appropriate conduct (e.g., luxury boutiques, golf clubs,professional symposiums at business schools), and we definenonconforming behaviors as those that deviate from suchbehavioral standards. Consistent with the notion that indi-viduals integrate specific environmental information intotheir overall evaluations, inferences, and choices (Belk 1975;Herr 1989; Swait and Adamowicz 2001), we investigatethe role of context on perceptions of status and competencederived from signals of nonconformity. We suggest thatnonconformity is likely to fuel perceptions of greater statusand competence in prestigious contexts, with expected be-havioral norms and relatively high standards of conduct.For example, in the setting of our first study, luxury bou-tiques, we find that shop assistants attribute higher poten-tial to a prospect wearing casual gym clothing than to onewearing an elegant dress. We would not expect to detecta similar red sneakers effect in the context of an ordinarystore that lacks the expected norm of being nicely dressed.Thus, we propose that a nonconforming behavior will sig-nal status and competence in the eyes of others in prestigecontexts with shared standards of formal conduct. We pre-dict that:

H5: Nonconforming behavior will lead to higher in-ferences of status and competence in prestigiouscontexts with expected norms than in less pres-tigious contexts.

Characteristics of the Nonconforming Behavior. We in-vestigate the extent to which a specific nonconforming be-havior is perceived as deliberate and intentional as anotherimportant boundary condition for the red sneakers effect. Abehavior is defined as deliberate if the actor set out to pro-duce the action (Malle and Knobe 1997), and an intention

is generally understood as a determination to engage in aparticular behavior (Atkinson 1964). We suggest that ob-servers attribute heightened status and competence to non-conformity when they believe that the nonconforming in-dividual is purposely deviating from an accepted, establishednorm. That is, the observer assumes that the nonconformingindividual is both aware of the norm and potentially ableto conform, but deliberately decided to adopt a noncon-forming conduct. In contrast, we expect that when a non-conforming behavior is perceived as unintentional, it willno longer be associated with enhanced perceptions of statusand competence. For example, when a nonconforming be-havior appears dictated by lack of a better alternative (as inthe case of observing a poorly dressed person who is home-

less) rather than by a deliberate dress choice, it will not leadto positive inferences in the eyes of others.

Similarly, we expect enhanced attributions of autonomyderived from signals of nonconformity to dissipate whenthe deviant behavior is perceived as unintentional. This pre-diction is in line with research suggesting that intentionalityand autonomy are strongly and positively associated (Deciand Ryan 1987). Thus, we hypothesize that:

H6: When a specific nonconforming behavior is per-ceived as unintentional, it will no longer be as-sociated with status, competence, and autonomy.

Overview of the Present Research

We test our hypotheses and theoretical framework, de-picted in figure 1, in five laboratory and field studies thatemploy different types of nonconformity and different pop-ulations of participants. Resistance to conformity pressurescan take distinct forms across individuals. Of particular rel-evance to our work is Tian et al.s (2001) conceptualization,which suggests that consumers exhibit three main behavioralmanifestations of nonconformity. First, creative choicecounterconformity refers to the tendency of some consum-ers to seek social differentness by selecting original, novel,or unique consumer goods (e.g., wearing a colorful, unusualtie to a formal event). Second, unpopular choice counter-

conformity reflects the selection or use of products andbrands that strongly violate and disrupt existing norms ofproper conduct (e.g., wearing a tie around ones head in aformal context). Finally, avoidance of similarity entails adowngrading of ones consumption style and refers to a lossof interest in, or discontinued use of, possessions to moveaway from the norm and reestablish ones differentness (e.g.,not wearing a tie in a formal context). In our research, wefocus on behavioral dimensions of nonconformity that entailsome deviance from the norm but are not perceived as a



7/21



FIGURE 1

THEORETICAL FRAMEWORK AND HYPOTHESES

strong disruption and violation of the norm. Accordingly,the manipulations in our studies center on creative noncon-formity and avoidance of similaritythat is, manifestationsof nonconformity within the realm of commonly acceptedbehaviors.

Study 1 examines the responses of shop assistants in lux-ury boutiques in Milan, Italy, and illustrates that noncon-formity, as compared to conformity, leads to inferences ofhigher status among individuals who are familiar with theenvironment. Study 2 investigates the effect of nonconform-ity and the role of the prestige context in a professionalsetting by testing students responses to the dress style oftheir professors. Study 3 delves into the underlying mecha-nisms of the red sneakers effect and demonstrates that infer-

ences of status and competence are mediated by the autonomythat participants perceive in the individuals nonconformingbehavior. Moreover, this study shows that the positive in-ferences dissipate when the nonconforming behavior is per-ceived as unintentional. Study 4 examines the moderatingrole of observers need for uniqueness and shows that par-ticipants with high levels of need for uniqueness tend toattribute more status and competence to nonconforming be-haviors as compared to participants with lower needs foruniqueness. Finally, in a follow-up study we increase thevalidity of our findings by examining nonconformity andneed for uniqueness outside the laboratory.

RESEARCH DESIGN AND FINDINGS

Study 1: Status Inferences Based on aNonconforming Dress Style

We begin our investigation by examining whether indi-viduals confer greater status to people who do not conformto shared norms of behavior in a given context. We createvignettes that describe a potential client who enters an el-egant boutique and engages in either conforming or non-conforming behavior. We employ two different operation-

alizations of nonconforming behavior to luxury stores. Instudy 1A, we manipulate nonconformity through style ofdress (e.g., entering the store wearing a gym outfit vs. adress), and in study 1B we examine consumers use of well-known brand names (e.g., wearing a Swatch watch vs. aRolex). We selected luxury boutiques as an experimental

setting due to their established and commonly perceivedbehavioral norm of elegant dress. Thus, we expect that inthis specific context, rather than in ordinary stores, a casualdress style will stand out and will be perceived as a deviationfrom the established standard. Moreover, we consider thecase of people who are either familiar or less familiar withthe context being examined by comparing the responses oftwo samples: shop assistants in luxury shops in downtownMilan, Italy, and women recruited in Milans central trainstation. We predict that participants who are familiar withthe context (i.e., the shop assistants) will confer greaterstatus to the nonconforming client rather than to the con-forming one. However, we expect the effect to be attenuatedor even reversed for participants who are less familiar with

the context (i.e., pedestrians recruited at the central station).Prior to running studies 1A and 1B, we conducted an ex-ploratory interview in Rome with the store manager of aprestigious luxury brand. According to him, shoppers athigh-end boutiques generally wear elegant and expensiveclothing, in keeping with the stores luxury atmosphere. How-ever, he admitted that his stores very top clientele alsoincludes customers who dress quite casually. Therefore, weexpect shop assistants in luxury boutiques, who are familiarwith the environment and are motivated to determine thestatus of potential customers, to be able to detect and in-terpret the unconventional behavior of potential customersas a signal of status.

Method. We recruited 109 female adults in downtownMilan. All participants responded to a short survey in Italianand received a pen for participating in the study. Fifty-twoparticipants were shop assistants working in boutiques sell-ing luxury brands such as Armani, Burberry, Christian Dior,La Perla, Les Copains, and Valentino (i.e., individuals fa-miliar with the environment of a high-end boutique). Onaverage, shop assistants had 12 years of experience in thefashion sector and an estimated net income ranging fromi14,400 to i16,800 per year. The other 57 female partici-pants were recruited at Milans central station (i.e., individ-uals less familiar with the environment of a high-end bou-tique). The estimated average net income for women livingin Milan is i15,800 (Registry Office Milan 2011). Thus,

both groups share comparable demographic profiles (gender,age, income, and nationality), but the first group is betteracquainted with the environment of luxury boutiques thanthe second group.

Participants completed the study in their own environment(boutique or train station) and were randomly assigned toone of two conditions: conforming versus nonconformingpotential client. Participants in each condition read a vignetteabout a potential client entering a luxury boutique. In study1A, the text read: Imagine that a woman is entering a luxury



8/21



FIGURE 2

STUDY 1A RESULTS: THE EFFECT OF NONCONFORMITY

AND FAMILIARITY WITH THE ENVIRONMENT

boutique in downtown Milan during winter. She looks ap-proximately 35 years old. Participants in the nonconform-ing condition next read: She is wearing gym clothes anda jacket. Participants in the conforming condition next read:She is wearing a dress and a fur coat. After reading thevignette, participants answered questions, using 17 scales,

to assess whether they believed the individual describedcould be a potential client at the luxury store and whethershe might be a VIP or a celebrity. Specifically, participantsanswered three questions assessing the womans status as aluxury client: 1. How likely is the woman described topurchase something in the store? (1 p very unlikely, 7 pvery likely). 2. Imagine that the woman described were tobuy something. Would she spend more or less than the av-erage store client? (1 p less than average, 7 p more thanaverage). 3. Can she afford the most expensive items inthe store? (1 p definitely yes, 7 p definitely no). Weaveraged the three items and used the resulting measure ofstatus as a luxury client as the first dependent variable inour analyses. Next, participants answered one question as-

sessing the clients perceived status as a celebrity, the seconddependent variable: Is she likely to be a VIP or a celebrity?(1 p very unlikely, 7 p very likely). In study 1B, partic-ipants in both conditions read: Imagine that a woman isentering a luxury boutique in downtown Milan during sum-mer. She looks approximately 35 years old. Next, partic-ipants in the nonconforming condition read: She is wearingplastic flip-flops and she has a Swatch on her wrist. Par-ticipants in the conforming condition read: She is wearingsandals with heels and she has a Rolex on her wrist. Afterreading the description of the client, participants answeredthe same questions as in study 1A.

Results(Study 1A). A 2 (nonconforming vs. conforming

client) # 2 (familiar vs. unfamiliar observer) between-sub-jects ANOVA using ratings of the potential as a luxury client(a p .88) as the dependent variable revealed a significantmain effect for nonconformity (F(1, 104) p 4.9, p ! .05),no main effect for familiarity (F(1, 104) p .82, NS), anda significant interaction (F(1, 104) p 37.7, p ! .001) de-picted in figure 2A. Shop assistants familiar with the en-vironment granted greater status to the nonconforming clientrather than to the conforming client (Mnonconforming p 4.9 vs.Mconforming p 3.8, t(49) p 2.8, p ! .01). In contrast, partic-ipants unfamiliar with the context granted less status tothe nonconforming client than to the conforming one(Mnonconforming p 3.5 vs. Mconforming p 5.7, t(55) p 5.8, p !.001). We conducted a similar ANOVA using celebrity status

as the dependent variable and found a similar pattern ofresults. The main effect for nonconformity was significant(F(1, 104) p 8.1, p ! .01), while the main effect for fa-miliarity was not (F(1, 104) p 1.7, NS). Consistent withour predictions, we found a significant interaction betweenconformity and familiarity (F(1, 104) p 15.5,p !.001; seefig. 2B). Shop assistants believed that the nonconformingclient was more likely than the conforming client to be acelebrity or a VIP (Mnonconforming p 4.9 vs. Mconforming p 2.5,t(49) p 5.4, p !.001); there was no significant difference

between conditions for participants unfamiliar with the lux-ury boutiques (Mnonconforming p 4.0 vs. Mconforming p 4.3, NS).

Results(Study 1B). A 2 (nonconforming vs. conformingclient) # 2 (familiar vs. unfamiliar observer) between-sub-

jects ANOVA using ratings of the status as a luxury client(a p .74) as the dependent variable revealed a significantmain effect for nonconformity (F(1, 104) p 10.9,p ! .001),no main effect for familiarity (F(1, 104) p .02, NS), and

a significant interaction (F(1, 104) p 35.0,p !.001). Shopassistants granted greater status to the nonconforming clientthan to the conforming one (Mnonconforming p 4.8 vs.Mconformingp 4.2,t(50) p 2.1,p ! .05); participants with no familiaritywith the environment did just the opposite (Mnonconforming p3.4 vs. Mconforming p 5.6, t(54) p 5.9, p ! .001). A similarANOVA using perceived celebrity status as the dependentvariable revealed no main effect for nonconformity (F(1,103) p 2.5, NS) or for familiarity (F(1, 103) p .96, NS)and a significant interaction between these two factors (F(1,



9/21



103) p 7.9, p ! .01). Mimicking the results of study 1A,shop assistants granted greater status to the nonconformingclient than to the conforming one (Mnonconforming p 4.6 vs.Mconforming p 3.0, t(49) p 3.2, p !.01), and there was nosignificant difference between conditions for pedestrians(Mnonconforming p 3.9 vs. Mconforming p 4.4, NS).

Discussion. Consistent with our hypotheses, the resultsof study 1 show that observers grant higher status to anonconforming individual than to a conforming one, as longas the observers are familiar with the environment. Specif-ically, shop assistants at luxury boutiques perceived a clientto be more likely to make a purchase and to be a celebritywhen she was wearing gym clothes or a Swatch than whenshe was wearing an elegant dress or a Rolex. The effect forpedestrians recruited at Milans central station was attenu-ated or even reversed. These participants, of similar back-ground but less familiar with the luxury boutique environ-ment, tended to perceive the shopper with the elegant outfitas being of higher or similar status relative to the poorly

dressed shopper. The shop assistants status inferences areconsistent with research demonstrating that conspicuousconsumption of brands and explicit use of other status sym-bols is often associated with low-status groups (Feltovichet al. 2002; Han et al. 2010; Mazzocco et al. 2012). Weanalyzed shop assistants open-ended comments and con-ducted follow-up questions to clarify their status inferences.Interestingly, the shop assistants in our study seemed tobelieve that the nonconforming client was purposely devi-ating from the accepted norm of appropriate behavior in anattempt to distinguish herself from the average shopper.Some participants in the nonconforming condition believed,in the words of one of them, that poorly dressed shoppersare often playing a role and doing it on purpose. One shop

assistant stated that wealthy people sometimes dress verybadly to demonstrate superiority and that if you dare enterthese boutiques so underdressed, you are definitely goingto buy something. In contrast, it did not occur to pedestriansthat a shopper might purposely enter a luxury store poorlydressed. These remarks suggest that status inferences maybe driven by perceived deliberateness of the individualsnonconforming behavior. The scenarios tested in this studybring to mind the famous scene from the film Pretty Womanof Julia Robertss character shopping on Rodeo Drive inBeverly Hills. Consistent with our conceptualization, thecharacters nonconforming dress style did not lead to in-ferences of high status by store clerks because it did notappear to be intentional. We will directly test this proposition

in study 3 by manipulating the extent to which noncon-forming behavior is depicted as deliberate or unintentional.

In study 2, we seek to complement the findings of study1 by testing our hypotheses in a more controlled laboratoryenvironment. We will examine the consequences of non-conformity through down-dressing in a professional settingand the role of the prestige of the context with relativelyhigh standards of conduct. Since in study 1 we demonstratedthe boundary condition of familiarity with the environment,in the studies that follow, we focus on participants who are

familiar with the particular environment, and we assess thedegree of their familiarity as a precondition.

Study 2: Nonconformity in Professional Settings

In this study, we examine the effect of nonconformingbehavior in a more professional context by testing studentsresponses to the dress style of their professors. In profes-sional settings, nonconformity, and casual dress style in par-ticular, are typically viewed as costly behaviors that couldpotentially damage ones employment or promotion pros-pects (De Souza et al. 2003; Michaels 2012). Given thecontext, we measure status in terms of respect by others andcompetence in terms of workplace performance. In addition,we investigate the role of prestige of the context by manip-ulating between-subjects the reputation of the setting de-scribed in the experiment. We predict that students will per-ceive a male professor who wears a T-shirt and is unshaven(i.e., nonconforming) as having higher professional status

and competence than a professor who wears a tie and shaved(i.e., conforming), but only when the professor teaches at atop school where established norms exist regarding formalattire at work.

Method. We recruited 159 respondents (55% female,Magep 23) at Harvard University in Boston who participated ina series of unrelated lab studies. The vast majority of re-spondents were current students at local universities (83%),thus ensuring that our sample was familiar with the experi-mental stimuli (i.e., descriptions of professors). We randomlyassigned participants to one of four conditions, in a 2 (con-forming vs. nonconforming individual) # 2 (prestige contextvs. nonprestige context) between-subjects design. We manip-ulated the conformity of the dress style of the professor bytelling participants that the professor typically wears a tie andis clean-shaven (conforming) or that he typically wears a T-shirt and has a beard (nonconforming). To manipulate theprestige of the context, we varied whether or not the universitythe professor teaches at was described as a top-tier university.Participants read the following description: Mike is 45 yearsold and teaches at a university [a top-tier university]. Hetypically wears a tie [a T-shirt] to work and is clean-shaven[has a beard]. Participants then assessed the professors pro-fessional status and competence by answering three questions.1. How well respected is Mike by his students? (1 p notrespected at all, 7 p extremely well respected). 2. How doyou expect him to perform in class as a teacher? (1 p poor,7 p excellent). 3. How do you expect him to perform as aresearcher? (1 p poor, 7 p excellent). We averaged thethree items (a p .80) and used the resulting measure as thedependent variable in our analysis. Participants also answereda manipulation check for our conformity manipulation: Howconforming to his work environment is Mikes style? (1 pnot conforming at all, 7 p extremely conforming).

Results. The manipulation check confirmed that theshaved professor wearing a tie was perceived to be signifi-cantly more conforming to his work environment than the



10/21



FIGURE 3

STUDY 2 RESULTS: THE EFFECT OF NONCONFORMITY IN

PRESTIGE CONTEXTS

unshaved professor wearing a T-shirt (Mconforming p 5.7 vs.Mnonconforming p 2.9, t(156) p13.8, p ! .001). As a furthermanipulation check, we conducted a 2 (conforming vs. non-conforming individual) # 2 (prestige context vs. nonprestigecontext) between-subjects ANOVA using ratings of confor-mity as the dependent variable. The analysis revealed a sig-

nificant main effect for nonconformity (F(1, 154) p 165.2,p !.001), a significant main effect for prestige context (F(1,154) p 5.0, p !.05), and a significant interaction (F(1, 154)p 6.0, p ! .05). Given the statistical significance of bothtreatment variables and their interaction, we proceeded withan analysis of the effect sizes to compare the relative impactof each factor (Perdue and Summers 1986). The effect sizeof the nonconformity manipulation ( p .52) was2hnonconformityrespectively 16 and 14 times larger than the effect size ofthe prestige context manipulation ( p .03) and of the2hprestigeinteraction ( p .04), suggesting that our noncon-2hinteractionformity manipulation was successful.

Next, we conducted a similar 2 # 2 between-subjectsANOVA using ratings of the professors professional status

and competence as the dependent variable. The analysisrevealed a significant main effect for prestige context (F(1,155) p 6.1, p ! .05), a nonsignificant main effect for non-conformity (F(1, 155) p .70, NS), and a significant inter-action (F(1, 155) p 13.0, p ! .001), depicted in figure 3.As predicted, when the school was described as prestigious,students attributed significantly more status and competenceto the nonconforming professor than to the conforming one(Mnonconforming p 5.7 vs. Mconforming p 5.0, t(83) p 3.3, p !.01). In contrast, when no information about the school wasprovided, there was a marginally significant difference be-tween conditions in favor of the conforming individual(Mnonconforming p 4.8 vs. Mconforming p 5.2, t(72) p 1.8, p p.07). These results suggest that people attribute higher status

and competence to individuals who are nonconformingrather than conforming in prestigious contexts with expectednorms of formal conduct. We also checked whether partic-ipants gender would impact status attributions. We con-ducted the same ANOVA analysis including gender as acontrol variable and found no significant effect for thisdemographic variable.

Discussion. Study 2 extends the findings of study 1 re-garding the relationship between signals of nonconformityand perceptions of status and competence by examining thislink in a different domain. We find that students perceivean unshaven professor who wears a T-shirt to have higherprofessional status and competence than a shaven professor

who wears a tie, but only in a prestigious context, withrelatively high standards of conduct.

Individuals can deviate from the norm and nonconformin several ways (Tian et al. 2001). While studies 1 and 2operationalized nonconformity as dressing down, in the nextstudy we examine nonadherence to dress codes through anoriginal product choice (i.e., wearing a red bow tie at aformal black-tie party at a country club). We also examinethe role of perceived autonomy as the key mediating mech-anism underlying status inferences. Since in this study we

tested and demonstrated the boundary condition relative toprestige contexts, in the experiments that follow we examinethe red sneakers effect in similar contexts with shared normsof formal conduct (e.g., country clubs, business schools).

Study 3: Nonconformity Depicted asUnintentional and Perceived Autonomy

The goal of study 3 is threefold. First, it examines non-conforming behaviors in the domain of nonadherence todress codes from a different angle than studies 1 and 2. Tian

et al.s (2001) theoretical account suggests that consumersresistance to conformity pressures can have distinct behav-ioral manifestations, such as downgrading of the consump-tion style (so-called avoidance of similarity) or the selec-tion of original and novel consumer goods (creative choicecounterconformity). While studies 1 and 2 manipulatednonconformity through casual dress styles, study 3 inves-tigates nonconformity through original dress styles. Specif-ically, we test how participants grant status within a mem-bership club to an individual attending a formal black-tieparty. We describe the individual as conforming or non-conforming by manipulating the color of the bow tie he iswearing at the party (black vs. red). Second, in this studywe examine another necessary condition of the red sneakers

effect. We expect that when a specific nonconforming be-havior is unintentional (i.e., it was not his intention todress in a way that potentially deviates from the norm),the nonconforming conduct no longer will be associatedwith enhanced status, competence, and autonomy, as hy-pothesized. Finally, in study 3 we delve into the mechanismsunderlying status and competence inferences resulting fromnonconformity. Specifically, we seek to demonstrate that ob-servers attribute higher status and competence in responseto signals of nonconformity because they believe that the



11/21



nonconforming individual is autonomous and in control, andcan afford to act according to his volition, as predicted byhypothesis 2.

Method. We recruited 141 participants who responded toa paid online survey on Amazon Mechanical Turk (MTurk;

45% female, Mage p 35). Participants were randomly as-signed to one of four experimental conditions in a 2 (con-forming vs. nonconforming dress style) # 2 (no intentionto deviate vs. control) between-subjects design. We manip-ulated conformity of dress style by telling participants aboutCharles, an individual attending a formal black-tie party whowas either wearing a red bow tie (nonconforming) or a blackbow tie (conforming). Participants read the following de-scription: Imagine Charles, a 40-year-old man who likesto play golf. This year, Charless golf club is hosting a black-tie holiday party. Charles decides to wear a red [black] bowtie to the party. Most of the other male invitees are wearinga black bow tie. In addition, we manipulated the deliber-ateness of the behavior by either depicting the choice as

unintentional through an extra statement at the end of thedescription (no-intention-to-deviate condition) or by omit-ting this information (control condition). Specifically, par-ticipants in the no-intention-to-deviate condition read: Itwas not Charless intention to dress in a way that poten-tially deviates from the expected dress code. After readingCharless description, participants assessed his perceivedautonomy by rating the following two items (a p .66):1. The extent to which Charles can afford to do what pleaseshim (1 p He can never afford to do what he wants, 7 pHe can always afford to do what he wants). 2. The extentto which Charles is in control over the decision of what towear (1 p not in control at all, 7 p completely in control).Participants then answered two questions on membership

status in the golf club and performance as a golf player:1. How likely is Charles to be one of the top membersof the country club? (1 p not likely at all, 7 p extremelylikely). 2. Do you think Charles has won golf competi-tions/prizes in the past? (1 p not likely at all, 7 p ex-tremely likely). We averaged the two items (a p .62) andused the resulting measure as the dependent variable inour analyses. Subsequently, participants answered threemanipulation-check questions about, respectively, the per-ceived (1) nonconformity, (2) creativity, and (3) deliber-ateness of the behavior described: 1. To what extent doesCharless bow tie conform to the dress code? (1 p notconforming at all, 7 p extremely conforming). 2. Howcreative is Charless bow tie choice? (1 p not creative at

all, 7 p extremely creative). 3. How deliberate is Charlessbow tie choice? (1 p not deliberate at all, 7 p extremelydeliberate). Finally, we asked participants if they ever at-tended parties or events with formal or semiformal dresscodes to assess the general level of familiarity with formalgatherings in the sample.

Results. The vast majority of participants (89%) have at-tended formal gatherings in the past, thus guaranteeing a sat-isfactory level of familiarity with the scenario being tested.

Manipulation Checks: Nonconformity and Creativity. Asexpected, the manipulation checks confirmed that partici-pants perceived wearing a black bow tie to the party as amore conforming and noncreative behavior as compared towearing a red bow tie. In particular, wearing a black bowtie was perceived as significantly more conforming than

wearing a red bow tie (Mconformity p 6.5 vs. Mnonconformity p2.3, t(139) p 21.2, p ! .001), and as significantly less cre-ative (Mconformity p 2.0 vs.Mnonconformity p 4.8,t(139) p 11.8,p ! .001). As a further check to assess the success and thevalidity of the manipulation (Perdue and Summers 1986),we also conducted a 2 (nonconforming vs. conforming dressstyle) # 2 (no intention to deviate vs. control) between-subjects ANOVA using ratings of conformity as the depen-dent variable. As expected, the analysis revealed a signifi-cant main effect only for nonconformity (F(1, 137) p 453.3,p ! .001) but not for deliberateness (F(1, 137) p 1.4, NS),nor for the interaction between the two factors (F(1, 137)p .81, NS). The same analysis performed on ratings ofcreativity as the dependent variable revealed an analogous

pattern of results with a significant main effect for noncon-formity (F(1, 137) p 138.3, p ! .001), and nonsignificanteffects for both deliberateness (F(1, 137) p .23, NS) andthe interaction (F(1, 137) p .01, NS).

Manipulation Check: Deliberateness. Participants per-ceived the described behavior as less deliberate in the no-intention-to-deviate condition than in the control condition(Munintentional p 4.4 vs.Mcontrol p 6.2,t(139) p 6.8,p ! .001).As a further check, we conducted the same 2 # 2 ANOVAusing ratings of deliberateness as the dependent variable.The analysis revealed a significant main effect for noncon-formity (F(1, 137) p 4.4,p !.05), a significant main effectfor deliberateness (F(1, 137) p 51.9, p ! .001), and a sig-

nificant interaction (F(1, 137) p 13.3,p !.001). Given thestatistical significance of both treatment variables and theirinteraction, we proceeded with an analysis of the effectsizes to compare the relative impact of each factor (Perdueand Summers 1986). The effect size of the deliberatenessmanipulation ( p .28) was nine times larger2hdeliberatednessthan the effect size of the nonconformity manipulation( p .03) and three times larger than the effect2hnonconformitysize of the interaction ( p .09), suggesting that our2hinteractiondeliberateness manipulation was successful.

Inferences of Status and Competence. Next, we con-ducted a 2 (nonconforming vs. conforming dress style) #2 (no intention to deviate vs. control) between-subjects

ANOVA using ratings of status within the country club andcompetence as a golf player as the dependent variable. Theanalysis revealed a nonsignificant main effect for noncon-formity (F(1, 137) p 1.5, NS), a significant main effect fordeliberateness (F(1, 137) p 4.1, p ! .05), and a significantinteraction (F(1, 137) p 3.7, p p .05), depicted in figure4A. In line with the findings of our previous studies, whenthe behavior was deliberate, participants granted signifi-cantly more status and competence to the individual wearingthe red bow tie than to the one wearing the black bow tie



12/21



FIGURE 4

STUDY 3 RESULTS: UNINTENTIONALITY OF THE

NONCONFORMING BEHAVIOR AS BOUNDARY CONDITION

(Mcontrol nonconformity p 4.9 vs. Mcontrol conformity p 4.2, t(68) p2.4,p ! .05). In contrast, there was no significant differencebetween conditions when the behavior was depicted as un-intentional (Munintentional nonconformity p 4.0 vs.Munintentional conformityp 4.2, NS). Importantly, when comparing the two noncon-forming conditions, we found that the positive status andcompetence inferences associated with wearing a red bow tiesignificantly decreased when the nonconforming behaviorwas clearly depicted as unintentional (Mcontrol nonconformity p 4.9vs. Munintentional nonconformity p 4.0, t(68) p 2.6, p ! .001). Wealso checked whether participants gender would impactstatus attributions. We conducted the same ANOVA analysisincluding gender as a control variable and found no significanteffect for this demographic variable.

Perceived Autonomy. We then performed a similar anal-ysis using ratings of autonomy as the dependent variable.The analysis revealed a significant main effect for noncon-

formity (F(1, 137) p 35.3, p ! .001), a significant maineffect for deliberateness (F(1, 137) p 4.1, p ! .05), and asignificant interaction (F(1, 137) p 4.8, p ! .05), depictedin figure 4B. As predicted, when we provided no informationabout the intentions of the described individual and the be-havior was interpreted as deliberate, participants perceived

the nonconforming individual as having significantly higherautonomy than the conforming one (Mcontrol nonconformity p 6.0vs. Mcontrol conformity p 4.5, t(68) p 6.0, p ! .001). The non-conformity manipulation elicited a significant difference be-tween conditions also when the behavior was depicted as un-intentional (Munintentional nonconformity p 5.2 vs. Munintentional conformity p4.5, t(69) p 2.6,p ! .05). Importantly, the comparison betweenthe two nonconforming conditions revealed that the perceivedautonomy participants associated with wearing a red bow tiewas significantly weakened when this nonconforming behaviorwas depicted as unintentional (Mcontrol nonconformity p 6.0 vs.Munintentional nonconformity p 5.2, t(68) p 3.4, p ! .001), as wepredicted. In sum, we find that enhanced perceptions of thenonconforming individuals status, competence, and auton-

omy dissipate when observers perceive the nonconformingconduct as unintentional.

Mediated Moderation Analysis. To test moderation bydeliberateness and mediation by perceived autonomy, weconducted a mediated moderation analysis (Edwards andLambert 2007) examining whether perceived autonomy me-diated the detected interaction between nonconformity anddeliberateness. As reported above, deliberateness signifi-cantly moderated both the dependent variable (status andcompetence) and the mediator (autonomy). Moreover, whenstatus and competence were regressed on nonconformity,deliberateness, their two-way interaction, and autonomy, themediator was significant (B p .38,t(137) p 4.1,p !.001),

and the effect of the interaction between nonconformity anddeliberateness on status and competence became nonsignif-icant (from B p .82, t(137) p 1.9, p p .05, to B p .51,t(136) p 1.2, NS). In a bootstrap analysis, we found thatthe 95% bias-corrected confidence interval for the size ofthe indirect effect excluded zero (95% CI p .033 to .717),suggesting a significant indirect effect.

Taken together, the results of study 3 deepen our under-standing of the interactions among the underlying processesof the red sneakers effect. We find that nonconformity leadsto inferences of heightened status and competence, as longas the deviant conduct is perceived as deliberate. Moreover,we show that autonomy mediates the interaction betweenthe nonconformity manipulation and deliberateness on status

and competence inferences.

Discussion. Study 3 extends our previous findings byexamining deviance from the norm through a dress choicethat denotes originality. We demonstrate that participantsperceive an individual wearing a red bow tie at a black-tieparty in a country club as a higher-status member in theclub and as a better golf player relative to a conformingindividual wearing a black bow tie. Importantly, this studyexplores the role of perceived deliberateness as a boundary



13/21



condition of the red sneakers effect. As predicted, we findthat when the behavior is perceived to be unintentionallynonconforming, the positive inferences associated with non-conformity dissipate. Finally, we provide evidence in sup-port of our proposed mediating mechanism and show thatparticipants infer enhanced status and competence because

they believe that the nonconforming individual has the au-tonomy to follow his preferences and deviate from the norm.

Study 4: Inferences from a NonconformingPresentation Style and Observers Need forUniqueness

Thus far, our experiments have manipulated nonconform-ity as nonadherence to expected dress codes. In this study,we extend our findings by examining a different way ofdeviating from the norm: the styles people use in theirPowerPoint presentations in a prestigious competition. Wetest how participants confer status and competence to a con-testant in the MITs well-known $100K business competi-tion. We manipulate between subjects whether the contestantadopts his own PowerPoint presentation layout (noncon-forming condition) or MITs official layout (conformingcondition). Importantly, we clearly establish the behavioralnorm by stating in both conditions that other participants inthe contest are using MITs official background. In thisstudy, we test whether the relationship between noncon-formity and perceptions of greater status and competence ismoderated by respondents level of need for uniqueness.Consistent with hypothesis 3, we expect participants withhigh levels of need for uniqueness, as compared to partic-ipants with low levels of need for uniqueness, to confergreater status and competence to the nonconforming indi-vidual. Moreover, we seek further support for hypothesis 2and for the findings of study 3 on the mediating role ofperceived autonomy. Finally, in this study we measure theactors perceived awareness of the typical PowerPoint slidestyle to confirm that in all conditions the contestant is viewedas knowledgeable of the norm.

Method. We recruited 149 participants who respondedto a paid online survey on Amazon MTurk (50% female;Mage p 37). Participants were introduced to the study andread a description of the actual MIT $100K competition.They were told, The MIT $100K Competition is one ofthe nations premier business plan competitions. The capitalraising contest is aimed at helping students and researchersin the MIT community start up their firms. The MIT $100K

brings together a network of resources (venture capitalists,entrepreneurs, mentors, and more than $350K in cash andprizes) to help participants through the funding process ofnew ventures. Then participants read about John, a can-didate in the competition. The study manipulated betweensubjects whether John adopted his own PowerPoint presen-tation layout (nonconforming condition) or, like most othercontestants, the MIT official layout (conforming condition).Participants read the following description: Imagine John,a 22-year-old student at MIT, who is participating in the

MIT $100K competition. John has already passed the firstround of the contest and is about to participate in the secondround. As he is preparing the slides for the presentation ofhis business plan, he could pick the official MIT backgroundor use a background of his choice for the slides. His slideswould have a more unusual and less conventional back-

ground. The majority of the other participants are using theofficial MIT background for the slides. Eventually John de-cides to use his own [the MIT official] layout for the slidepresentation.

Subsequently, participants answered a series of questions.In order to avoid potential order effects, we counterbalancedthe appearance of the measures. Specifically, the order ofappearance of the dependent variable (perceived status andcompetence) and the mediator (autonomy) was interchanged,and the moderator (need for uniqueness) appeared either atthe beginning or at the end of the survey. We assessed statusand competence by asking participants to answer four ques-tions: 1. How likely is John to win the MIT $100K com-petition? (1 p not likely at all, 7 p extremely likely). 2.

How likely is John to become a millionaire entrepreneurone day? (1 p not likely at all, 7 p extremely likely). 3.How do you think Johns business idea compares to otherbusiness proposals in the contest? (1 p below average, 7p above average). 4. How well respected is John by hisfriends? (1 p not respected at all, 7 p extremely wellrespected). We averaged the four items to create a measureof perceived status and competence (a p .82) and used itas the dependent variable in our analyses. Similarly to study3, participants assessed Johns perceived autonomy by ratingthe following two items (a p .81): 1. The extent to whichJohn can afford to do what he wants (1 p He can neverafford to do what he wants, 7 p He can always afford todo what he wants). 2. The extent to which John is in control

(1 p not in control at all, 7 p completely in control). Asa manipulation check for our nonconformity manipulation,participants answered the following question: How con-forming to competition standards is Johns presentationstyle? (1 p not conforming at all, 7 p extremely con-forming). Moreover, participants were asked to judge Johnsawareness of the norm: Is John knowledgeable about theappropriate slides style for the competition? (1 p notknowledgeable at all, 7 p extremely knowledgeable). Wethen assessed the level of acquaintance with the behaviordescribed by asking respondents to rate their familiarity levelwith PowerPoint or similar presentation programs (1 p notfamiliar at all, 4 p somewhat familiar, 7 p very familiar).Participants completed the 31-item scale developed by Tianet al. (2001) to measure the degree to which individualspursue differentness and uniqueness (e.g., I actively seekto develop my personal uniqueness by buying special prod-ucts or brands).

Results. Preliminary analysis revealed no significant dif-ferences in the patterns of results between respondents gen-der groups; thus, we analyzed the results jointly. The ma-nipulation check confirmed that participants perceived thestudent as nonconforming when he was using his own back-



14/21



FIGURE 5

STUDY 4 RESULTS: OBSERVERS NEED FOR UNIQUENESS

AND RESPONSES TO NONCONFORMITY

(SPOTLIGHT ANALYSIS)

ground for the slides (Mnonconforming p 5.7 vs. Mconforming p2.7,t(147) p 14.1,p !.001). Moreover, participants meanlevel of familiarity with PowerPoint was fairly high (Mp5.1) and significantly above the scale midpoint (4) in bothconditions (Mconforming p 4.8, t(66) p 3.8, p ! .001, andMnonconforming p 5.3, t(80) p 6.9, p ! .001).

As expected, participants granted more status and com-petence to the nonconforming individual than to the con-forming one (Mnonconforming p 5.0 vs.Mconforming p 4.2,t(147)p 5.3, p ! .001). Moreover, they perceived the noncon-forming individual as more autonomous and more able toafford his preferences (Mnonconforming p 5.6 vs. Mconforming p4.3, t(147) p 7.4, p !.001). Participants thought that Johnwas more knowledgeable about the appropriate PowerPointslide style for the competition when he was nonconformingrather than conforming (Mnonconforming p 5.8 vs.Mconforming p5.2, t(144) p 4.1, p ! .001), indicating that in this casenonconforming behavior is not associated with ignorance ofexpectations in the given context.

Perceived Autonomy as Mediator. We examined whetherperceived autonomy mediated the relationship between non-conformity and greater status and competence inferences,as we hypothesized. First, the nonconformity manipulationaffected status and competence inferences (B p .38,t(147)p 5.3, p !.001). Second, the nonconformity manipulationsignificantly affected autonomy (B p .66, t(147) p 7.4, p! .001). Finally, the influence of the independent variableon status and competence became nonsignificant when au-tonomy was included in the model (fromB p .38,p ! .001,to B p .11, NS). In a bootstrap analysis, we found that the95% bias-corrected confidence interval for the size of theindirect effect excluded zero (95% CIp .164 to .389). Theseresults suggest that participants attributed higher status and

competence to the nonconforming individual, relative to the

conforming individual, because he was perceived as moreautonomous.

Need for Uniqueness as Moderator. Next, we examinedthe moderating role of observers need for uniqueness (ap.97). We analyzed responses using a linear regression with

perceptions of status and competence as the dependent var-iable and with the following independent variables: a var-iable for the behavior coded as 1 for using a nonconformingpresentation layout and 1 for using a conforming presen-tation layout, need for uniqueness scale (standardized forease of interpretation), and their interaction. This analysisrevealed a main effect of condition (B p .38,t(145) p 5.5,p ! .001), a nonsignificant main effect for observers needfor uniqueness (B p .08, NS), and a significant interactionbetween these two variables (B p .28, t(145) p 4.1, p !.001).2 To further explore this interaction, we performed aspotlight analysis (Fitzsimons 2008) that considered the ef-fect of nonconformity among those participants with higherand lower need for uniqueness. As illustrated by figure 5,

a spotlight analysis at one standard deviation above the meanof need for uniqueness revealed a significant difference (Bp .66,t(145) p 6.8,p !.001): participants with high needfor uniqueness conferred significantly more status and com-petence to John when he engaged in a nonconforming be-havior as compared to when he engaged in a conformingbehavior. In contrast, a similar spotlight analysis performedat one standard deviation below the mean of need for unique-ness showed a nonsignificant difference between conditions(B p .09, NS). Thus, nonconformity predicted higher in-ferences of status and competence when respondents scoredhigh in need for uniqueness, but such relationship did notexist for respondents who scored low in need for uniqueness.

Additionally, we examined the slopes of need for unique-

ness in each condition. The slope was positive and signif-icant when John was depicted as adopting a nonconformingpresentation format (B p .36, t(145) p 3.9, p ! .001),indicating that participants scoring high on the need foruniqueness scale attributed more potential to John when hisbehavior was perceived as deviant (as compared to the re-actions of participants with lower levels of need for unique-ness). However, when John was described as adopting amainstream presentation style, the slope of need for unique-ness was negative (B p .20, t(145) p 2.0, p p .05),suggesting that participants high in need for uniquenessgranted less status and competence to John when he fol-lowed the same behavior of other contestants.

Next, we also examined the moderating role of observers

need for uniqueness on the mediator, perceived autonomy.We analyzed responses using a linear regression with per-ceived autonomy as the dependent variable and with the

2 The 31 items of the Need for Uniqueness scale, as conceptualized

by Tian et al. (2001), can be further divided into three subsets: (1) Creative

counterconformity (12 items); (2) Unpopular counterconformity (11 items);

(3) Avoidance of similarity (8 items). We performed the same moderation

analysis with each of the three subscales and found a significant interaction

with each of them.



15/21



following independent variables: a variable for the behaviorcoded as 1 for using a nonconforming presentation layoutand 1 for using a conforming presentation layout, needfor uniqueness scale (standardized for ease of interpretation),and their interaction. This analysis revealed a main effectof condition (B p .66, t(145) p 7.5, p ! .001), a nonsig-

nificant main effect for observers need for uniqueness (Bp .02, NS), and a significant interaction between these twovariables (B p .19, t(145) p 2.1, p !.05). To explore thisinteraction, we performed a spotlight analysis (Fitzsimons2008) that considered the effect of nonconformity on per-ceived autonomy among those participants with higher andlower need for uniqueness. A spotlight analysis at one stan-dard deviation above the mean of need for uniqueness re-vealed a significant difference (B p .85, t(145) p 6.7, p! .001): participants with high need for uniqueness per-ceived John as significantly more autonomous when he en-gaged in a nonconforming behavior as compared to whenhe engaged in a conforming behavior. A similar spotlightanalysis performed at one standard deviation below themean

of need for uniqueness also showed a significant differencebetween conditions (B p .47,t(145) p 3.8,p ! .01). Whilethe nonconformity manipulation elicited a significant reac-tion for participants with both high and low need for unique-ness, the magnitude of the effect for respondents with highneed for uniqueness was almost double compared to the sizeof the effect for respondents with low need for uniqueness(Bhigh uniqueness p .85 vs. Blow uniqueness p .47).

Mediated Moderation Analysis. To test moderation byneed for uniqueness and mediation by perceived autonomy,we conducted a mediated moderation analysis (Edwards andLambert 2007) examining whether perceived autonomy me-diated the detected interaction between the nonconformity

condition and need for uniqueness. As reported above, needfor uniqueness significantly moderated both the dependentvariable (status and competence) and the mediator (auton-omy). Moreover, when status and competence were re-gressed on nonconforming behavior, need for uniqueness,their two-way interaction, and autonomy, the mediator wassignificant (B p .37,t(144) p 6.6,p !.001), and the effectof the interaction between nonconforming behavior and needfor uniqueness on status and competence decreased (fromB p .28, t(145) p 4.1, p ! .001, to B p .21, t(144) p3.5, p ! .001). In a bootstrap analysis, we found that the95% bias-corrected confidence interval for the size of theindirect effect excluded zero (95% CI p .004 to .162),suggesting a significant indirect effect.

In sum, the results of study 4 shed light on the interactionsbetween the underlying processes of the red sneakers effect.Need for uniqueness moderated both the direct effect ofnonconformity on status inferences (dependent variable) andthe first stage of the indirect effect of nonconformity onperceived sense of autonomy (mediator), suggesting thatparticipants high in need for uniqueness attributed higherstatus and competence and heightened sense of autonomyin response to nonconformity signals, relative to participantswith low levels of need for uniqueness. Moreover, we find

that autonomy mediated the interaction between the non-conforming manipulation and need for uniqueness on statusand competence inferences.

Discussion. Study 4 extends our findings about noncon-formity-based inferences beyond the domain of nonadher-

ence to dress codes. We find that participants perceive acontestant in a prestigious competition as having higherstatus and competence when he adopts his own layout forthe presentation rather than the standard background. In ad-dition, we show the moderating role of observers need foruniqueness on inferences of heightened status and compe-tence and perceived autonomy. Relative to participants withlow levels of need for uniqueness, participants with highlevels of need for uniqueness attributed greater status, com-petence, and autonomy to the nonconforming individualrather than to the conforming one. In line with hypothesis2 and the findings of study 3, we provide further evidencein support of our proposed mediating mechanism and showthat participants infer higher status and competence because

they believe that the nonconforming individual has the au-tonomy to follow his volition.

In the next, follow-up study, we examine responses tononconformity outside the laboratory and provide furthersupport to the moderating role of need for uniquenessthrough a behavioral proxy.

Follow-Up Study: Stepping Outside the Lab withRed Sneakers

This follow-up study aims to increase the ecological va-lidity of our findings by employing a real-world manipu-lation of nonconformity and by examining a product-related

behavioral proxy for need for uniqueness. Specifically, weexamine the reactions of executives attending a formal sym-posium in a prestigious business school at which a professorwears red sneakers while teaching in the classroom. In ad-dition to measuring need for uniqueness through conven-tional scale items, we collect information on whether par-ticipants own shoes that have an unusual color and thus doengage in less conventional consumption in their daily lives.Relative to individuals with low need for uniqueness, weexpect individuals with high need for uniqueness to ownmore unusual pair of shoes and to attribute more status andcompetence to signals of nonconformity, in line with hy-pothesis 3 and with the results of study 4.

Method. Participants were 59 male executives (Mage p46) attending the Inner City 100 Urban Small Business Sym-posium.3 At this 1-day event, executives gather for net-working opportunities and a full day of management edu-cation. We decided to focus our analysis on male participantsbecause almost all female respondents (28 out of 30) saidthey owned a pair of distinctive-looking shoes. In the case

3 This number excludes 10 participants who did not notice that the

professor was wearing red sneakers, possibly because the shoes were not

visible from where they were sitting.



16/21



of male participants, 40 individuals out of 59 indicated thatthey owned this type of shoes. Because this behavior seemsto be the prevailing norm for female individuals, it is not adiscriminating behavioral proxy for nonconforming anduniqueness motives. Nevertheless, we also analyzed thesample in its entirety and, as reported below, the nature and

significance of the results did not change. In this study, thefemale negotiations professor taught her 90-minute sessionwearing a pair of (nonconforming) red Converse sneakers.At the end of the class, participants were asked to completea short survey. Participants assessed the professors profes-sional status and competence by answering four questionssimilar to those used in our previous studies: 1. How highis [professors name] status within [school name] (comparedto colleagues in her cohort)? (1 p definitely low, 7 pdefinitely high). 2. How likely is she to be the head of thenegotiations unit at [school name] 10 years from now? (1p very unlikely, 7 p very likely). 3. How likely is [pro-fessors name]s research to be featured in the Harvard Busi-ness Review? (1 p very unlikely, 7 p very likely). 4. In

your opinion, how likely is she to be selected to present herresearch at the prestigious [school name] research sympo-sium? (1 p very unlikely, 7 p very likely). We averagedthe four items (a p .75) and used the resulting measure inour analysis. Next, we asked participants whether they hadever owned a pair of distinctive-looking shoes: Did youever own a pair of shoes that had a distinctive color? (yes,no). To make sure that owning a pair of distinctive shoeswas a valid behavioral proxy for uniqueness motives, weasked participants to answer three items (two questions re-lated to distinctiveness and one specific to nonconformity)selected from Snyder and Fromkins (1977) need for unique-ness scale: 1. Do you typically prefer to conform to dresscodes? (1 p strongly avoid, 7 p strongly prefer, reverse

coded). 2. Do you like to dress in a way that is distinctive?(1 p dislike extremely, 7 p like extremely). 3. Please rateyour agreement with the statement: Whenever I take part ingroup activities, I am somethin

Red Sneaker Effect

Documents

Transcript of Red Sneaker Effect