This item is only available as the following downloads:
GROUNDED IDEOLOGIES: THE FEELING OF POWER AND SYSTEM JUSTIFICATION BY TIMOTHY NEST A Thesis Submitted to the Division of Social Sciences, the Division of Humanities, and the Division of Natural Sciences New College o f Florida in partial fulfillment of the requirements for the degree Bachelor of Arts Under the sponsorship of Charlene Callahan Sarasota, Florida May, 2011
The Feeling of Power and System Justification | ii GROUNDED IDEOLOGIES : Tim othy Nest New College of Florida, 2011 ABSTRACT Cont rary to traditional approaches in the psychological study of group behavior, System justification theory has argued that social actors are not only motivated to sustain their local social affiliations, relevant group identifications, and ingroup' relevan t belief systems but also possess a basic motivation to preserve and maintain dominant social structures through system affirming ideologies. Since the rise in popularity of the Implicit association test (IAT), System justification theory has shown that s ocially marginalized group members or those in positions of limited social power, tend to exhibit high levels of in group derogation and out group favoritism compared to members of dominant, or high status social groups Research conducted on social power has d emonstrated dynamic effects of power relevant affect on various aspects of social cognition, categorization and stereotyping, as well as major positive physiological effe cts, such as stress reduction. In light of the social psychological literature o n feelings of power and their effect on social cognition, categorization, and stress reduction, it is hypothesized that subjective feelings of power could function to ameliorate or eliminate system justification tendencies among socially marginalized indi viduals. The current study aims to explore the connection between subjective power and system justifying processes related to gender role stereotyping in men and women.
The Feeling of Power and System Justification | iii Implicit and explicit measures of attitudes related to gender and authority were used to assess whether participants across three "power" conditions differed significantly in their system justifying tendencies. While results supported a significant effect of system justification in men and women, subjective power had only a marginally sign ificant effect on implicit system justification in women, and held for only one of the two pools of data collected. Despite statistical weakness, explicit power priming did reverse system justification effects for women on average ; supporting the hypothesi s that transient power related affect can under certain conditions mitigate ideological sway Under the sponsorship of Charlene Callahan Division of Social Sciences
The Feeling of Power and System Justification | iv TABLE OF CONTENTS Abstract........................................ .......................................................... i i Table of Contents................................................................................... iv Foreword.. ........................... ...................................... 1 Introduction. ................................................................. 2 Review of Literature ................................................... .............. 6 Group Behavior in Context...................................................... Social Identity...................................................... 10 Individuals vs. Groups.............. ............................ 12 Anxiety, Subjective Uncertainty, and Group Identification......... 15 Ideology and System Justification ... 18 Complementary Stereotyping and Gender ............. ... ................ 26 Categories .... .................................................... .... .. 32 Categorical Cognition ......................... ........................... ....... 34 Categoric al Social Cognition 3 6 Cognition as Emotive .. 39 Embodied Cognition. .. 41 Social Cognitive Perspectives on Power .. 43 The Present Study... ............................................................................. ... 46 Method ................................................................................... ............... 47 Participants....... ......................................................................... Materials.................................................................................. 48 Procedure............................ ...................................................... 50
The Feeling of Power and System Justification | v Results .................................................................................................. 51 Discussion................................................. ............................................. 54 References.............................................................................................. 57 Tables .............................................................. .................................... 66 Appendix A ................................................................................. 70 B ...................................................................... ........... 73 C ................................................................................ 81 D. Notes on Programming and Experimental Design ...... 83
Foreword Th e present thesis involves three major components. First, a critical investigation of prominent social psychological perspectives on social reasoning, motivation, and group behavior is undertaken, focusing on the phenomenon of what has been called by socia l psychologists and political theorists alike, "false consciousness." Here a novel theory of social behavior and ideological mechanisms is offered, suggesting against the prominent essentialist reading of ideology or "system justification," as it is call ed that self damaging ideological phenomena have as their basis an affective propensity for self relevant anxiety, and can be mitigated at the affective level through affective states characteristic of "power." The second component was an empirical psyc hological study, testing this hypothesis through implicit and explicit measures of culturally salient attitudes about gender, power, and authority. Finally, a substantial computer science component was included, involving the development of an original ve rsion of the popular methodological framework known as the Implicit Association Test, within a Java programming framework. The online distribution of the experiment, which included various priming measures in addition to the IAT and survey, also involved s ome programming; specifically in MySQL, and PHP environments Documentation of this work can be found in Appendix D.
The Feeling of Power and System Justification | 2 Introduction "Here we see the ego, in its essential resistance to the elusive process of Becoming, to the variations of Desire. This il lusion of Unity, in which a human being is always looking forward to self mastery, entails a constant danger of sliding back again into the chaos from which he started; it hangs over the abyss off a dizzy Assent in which one can perhaps see the very essenc e of Anxiety Lacan In November of 2006, Mike Jo nes, a former escort, leveled an accusation against the evangelical preacher, Ted Haggard. The claim, which would later prove true, was that Haggard, a vehemently anti gay, family oriented figurehe ad for the Christian right, had for some time been paying Jones to have sex with him. Without revealing his name, level of influence, or the source of his wealth, Haggard had also, on occasion, smoked crystal methamphetamine with Jones. Beyond a level of d ivine justice detectable in Haggard's, widespread media exposure, the ordeal illustrates a very traditional conception of ideology whose influence is notable even today: the understanding of ideology as articulating a fundamental rift between public and pr ivate life Public or private; ideas or phenomena; this has always been the question at hand. The history of Western philosophy, theology, and metaphysics, is the history of a long painful tension between these terms. Ideology is invoked as the modern, s ecular form of the moral dictate s of religious life, the Idea, or God whose a priori status renders all private existence, contingent or accidental phenomena contrary and threatening to the high flung Platonic or Medieval Christian realms of divine purity. But as Marx who was among the first, at least in the Hegelian tradition, to frame the question differently realized, there has been little grounds for distinction to begin with. "T he human essence is no abstraction inherent in each single individual. In its reality it is the ensemble of the social relations (Theses on Feurbach. Marx, 112)
The Feeling of Power and System Justification | 3 Accordingly, our investigation and the research that accompanies it, rests on the distinction between two images of ideology. The first is the one described above, which locates between the public and the private an un bridgable divide. This perspective has its highest realization in the process Marx refers to as "false consciousness," whereby social actors unknowingly advocate for, and serve social forces which con tradict their own self interests. !i"ek sums up this position as follows: a doctrine, a composite of ideas, beliefs, concepts...destined to convince us of its 'truth', yet actually serving some unavowed particula r power interest ( !i"ek, 10 )." The maj or alternative to this is the embodied or anti ideological perspective suggested by Marx, and advocated by Deleuze and others; this approach regards ideology as immanent to social reality, and immediately related to material, affective, and bodily factor s. Deleuze in particular offers a perspective in which phenomena denoted by the term, ideology' can be adequately accounted for through the interactions between organizations of power, and desire. Indeed, according to Marx, perspectives which would essent ialize any dichotomy between public and private are the epitome of ideology itself. Lacan invokes a kind of primordial anxiety as the negative springboard for the genesis of the ego The ego, in its association with the quintessentially ideological symb olic order, is assimilated to various groups, socio cultural institutions or systems, and defined accordingly. Its beliefs, knowledge, and desires, are dictated first from above, vis a vis the Nom du pere the No, or Name, of the father ; and button tie bri dging meaning and language and then from below, as desire is oriented and organized in perfect accord with the material interests of the system in question, as reflected in the family.
The Feeling of Power and System Justification | 4 Anxiety, from the perspective of psychoanalysis, is imputed as the pri meval source of all determinate, or manifest culturally restrained desire; it arises from confrontation with death, "essential resistance to becoming." For empirical psychology, however, its role in forming desire, motivation, and behavior, and the role o f affect in general, therein, has been considerably underplayed since the rise of cognitivism as a dominating framework As empirical approaches and methods have evolved, however, a wave of social psychological research has arisen in recent years granting new credence unbeknownst to it to an emotionally grounded discourse for the study of social life, which does not exclude psychoanalytic, biological, and affective factors from the s ocial etiology. In spite of this, empirical social psychology and its th eories have offered somewhat weak theoretical accounts of social reality, ideology, and discriminatory behavior, relying ineluctably on unclear constructs whose function an effectuality shifts with every concrete experimental contingency. Thus it has been necessary, through a critical reading of the social psychological literature on ideology and false consciousness, affect, and power, to expose the theoretical presumptiveness of this work, the psychoanalytic accountability of empirical social psychology, a nd its utility for refining and securing a substantial and comprehensively psychological science Here, we tackle what is surely the strongest extant account of false consciousness in social psychology today. Termed System justification theory (SJT), this perspective extends an d supersedes those "group justification," or culturally oriented accounts favored in past years giving priority instead to the role of cross cultural, group transcendent, and system inscribed values in determining social attitudes Hierarchies
The Feeling of Power and System Justification | 5 concerning age, sex, race, and wealth are exemplary among the evaluative spectra involved in System justification processes While SJT posits system maintenance and other ideological processes as essential motivating factors in all social beha vior, a critical look at the social psychological literature reveals the plausible priority of affect. System justification, in other words, is one means of coping with the kinds of primordial anxiety, or originary, undifferentiated, and fragmentary desire cited by Lacan. It is a means, as Nietzsche says, of giving a "single sense" to the manifold of desire; but it is not the only means. For this purpose, modes of aff ect related to power, and quite distinct of System justifying processes are also adequate. In the literature review to follow, four basic assumptions are supported: (a) Forms of social reasoning are socially conferred. (b) Social identification, and ideology are both grounded in the body, and maintained by subjective uncertainty. (c) All reaso ning, and particularly social reasoning, is grounded in the body, in affect. (d) Affect related to power, autonomy, and self control has a significant impact on cognition, and is one means by which subjective uncertainties may be attenuated. The research and discussion that follow support the additional premise that power related affect does in some significant way, function under certain conditions, to mitigate false consciousness behaviors among socially marginal individuals. Results thus support the Lac anian assertion that the proper location of ideology is in the body.
The Feeling of Power and System Justification | 6 Review of Literature The aim of the present work is to develop an original synthesis of at least two distinct lines of research in social psychology one concerning the physiologic al and behavioral effects of what has in the relevant literature been termed "the feeling of power," and the other approaching an empirical, social scientific solution to the deeply elusive question of ideology, beginning with work on belief formation and intergroup conflict. Before presenting at length any new theoretical formulations however, it is necessary to set these in context, and to consider in detail the disparate genealogies and assertions of the aforementioned research. Group Behavior in Contex t In the social psychological literature on group behavior, attention has been traditionally given to patterns of stereotyping in individuals and the motivational consequences of group membership. Though it is almost unanimously conceded that cultural fac tors, group identifications, and individual differences all play a major role in determining the nature of inter group discriminati on, much controversy has arisen generating divergent views on how, to what extent, and at which level these factors conditio n the outgroup derogatory or alternatively, ingroup affirmative behaviors that such modes of discrimination frequently manifest. Early accounts of these phenomena point quite reli ably to two distinct approaches: one which imputes prejudice, or group bias to mainly low level intra personal sources, and another which tends to emphasize the etiological role of large scale social conflict and group interests embodied by social actors, even at the expense of individual characteristics distinct of the group in question. Following the conventions of Tajfel & Turner (1970), we shall refer to these below as
The Feeling of Power and System Justification | 7 "person oriented" or "low level" theories; and "group oriented," "realistic conflict," or "high level" theories, respectively. Of the former position, one note worthy example can be seen in the theory of the Authoritarian Personality (TAP) and eponymously titled book. Based on survey data measuring anti Semitism, racism, and the tendency to glorify socially dominant or otherwise authoritarian figures, Adorno, Fre nkel Brunswick Levinson, and Sanford (1950) argued that discriminatory practices, and racial or ethnic biases derive mainly from paranoid, chauvinistic, and authoritarian personalities which are the motivational consequence of particular modes of upbring ing, early childhood development, and other interpersonal dynamics. Outgroup derogation, in other words, is the primary effect not of high level social conflicts, or group based competition over resources, but of a "fascistic" personality trait cultivated only indirectly by society at large. Despite certain empirical concerns, TAP has proven quite influential within the social psychological theory of personality and social behavior, yielding potent theoretical measures like Right Wing Authoritarianism (RWA ) and paving the way for the now popular theory of Social Dominance Orientation (Sidanius, Pratto, & Bobo, 1994; Kahn, Ho, Sidanius, & Pratto, 2009) to be discussed here in greater detail later. A major alternative to such accounts had its first incarnat ion in what has been called Realistic Conflict Theory (RCT). RCT posits that social prejudice exists to greater and lesser degrees as a consequence of real group interests and intergroup competition over resources. The "robbers cave experiment" from which this theory arose consisted of the random division of 24, 12 year old boys into two distinct groups at a common summer camp. In spite of pre existing inter group friendships, signs of hostility
The Feeling of Power and System Justification | 8 spontaneously emerged between groups within days. The two grou ps developed names ("Rattlers" and "Eagles"), symbols of identification and distinct territories over which they were eager to compete. When the Rattlers heard the other group playing on "their" ball field, they made remarks expressing the feeling that they considered others playing there as intrusion. Even without coming into physical contact with "those boys at the other end of the camp," the Rattlers had built up a highly competitive mood in relation to them." Conflicts between these groups, which w ere artificially aggravated by controlled "friction producing activities," ultimately escalated to such a point that researchers deemed them unsafe, and were forced to truncate this stage of research. In a final portion of the study, researchers showed tha t they could resolve inter group conflict through the introduction of "superordinate goals," requiring the cooperation of both groups. In the wake of this cooperative exercise, campers bonded to such an extent that they insisted on riding the same bus home For researchers, this was taken as definitive proof that inter group conflict is basically economic in nature, and grounded in the struggle for m aterial resources (Sherif, Harvey White, Hood, & Sherif, 1954). While the proponents of RCT demonstrated the clear effectuality of introducing common inter group goals, and common inter group resources in resolving serious group conflict, the basic premise from which these researchers departed: that inter group conflict is necessarily mediated by material group interests, and that these same interests
The Feeling of Power and System Justification | 9 are what motivate human beings to assimilate to groups is dubious. That individuals strongly identify with groups solely for reasons of power and security is a primitive position indeed. If individuals are drive n towards group identification by predominately material or economic concerns, then what sust ains the fervent discrimination identified by Adorno and others, against marginal ethnic, or religious groups posing no clear economic threat? What, to employ a ti mely example, might underlie a vehement and widespread popular opposition, even among those in lower socio economic margins, to the state funded provision of free nation wide health care? The question remains; what motivates group identification, and the v arious modes of discrimination accompanying it? In response to Sher if & Sherif (1954), Tajfel, Billig, Bundy, and Flament (1971) developed a procedure to evaluate the minimum necessary conditions under which discriminatory inter group behavior can be expe cted to evolve. A series of experiments demonstrated how even arbitrary group assignment with little to no value connotation can yield discernible derogation effects targeted at outgroup members, and powerful identification with, or favoring of ingroup me mbers. Results supported the hypothesis that even in absence of group competition, clear differences between group members or substantive material grounds for inter group conflict, group identification occurs rapidly and spontaneously. Participants in one experiment proved so loyal to their ingroups (established on the basis of artistic preference for either Klee or Kandinsky) that even in circumstances where the benefits of inter group cooperation clearly outweighed those of ingroup exclusivity, group memb ers still unambiguously favored members of their own ingroup. Furthermore, Billig & Tajfel (1973) demonstrated effects of ingroup favoritism and outgroup discrimination between groups established on an explicitly random basis.
The Feeling of Power and System Justification | 10 Tajfel's Minimal Group Para digm (MGP), as it was called, poses obvious problems for RCT, not only for its thorough repudiation of the notion that inter group dynamics are essentially reducible to utilitarian concerns of self or eve n common interest, but also because of the heavy e mphasis it places on non competive and seemingly arbitrary motivations for group formation. While opposing claims to valuable social resources such as wealth, power, territory, or prestige may be sufficient for group discrimination of the form identified a bove, they are not, as RCT understood it, necessary MGP shows, on the contrary, that the minimal necessary conditions for group formation may be near impossible to establish, as ingroup favoritism can be observed even among those with little to no basis f or group distinction. These same criticisms might also be applied to the more personality oriented or interpersonal theories described above. Though individual persons may show to be more or less inclined towards explicit derogation of outgroup members, st ereotyping, and ingroup favoritism, these activities are modalities of all social agents, and arise to varying degrees in all social contexts. Social Identity Theory In light of MGP, and the apparently arbitrary nature of group identification and group mot ivated discrimination it posits, Tajfel & Turner developed a more comprehensive account of mechanisms underlying group formation, known as Social Identity Theory (SIT). Social Identity Theory makes three primary claims: first, that individuals strive to es tablish and preserve a positive social identity; second, that positive social identity is derived from group membership as well as perceived value of ones ingroup in relation to a comparable outgroup; and finally, that when social identity is unsatisfactor y, due perhaps to the subordinate position of a relevant ingroup, group
The Feeling of Power and System Justification | 11 members will attempt to either leave their ingroup, or cultivate a positive socia l value for their group (Tajfel & Turner, 1979). Implicit in this last claim is the notion that alongsi de the particular groups in question stands a superior social dimension, with reference to which the value or status of various social groups can be determined. SIT, however, claims that not all social groups enter into a state of mutual discrimination, a nd that distinct groups must be salient for one another as groups worthy of comparison, in order to spawn the kinds of discriminatory behaviors relevant to this discussion. In other words, group comparisons which might determine one group's social value re lative to another's, are made not on the basis of predominating social structures and their corresponding standards of comparison, or even necessarily through reference to objective measures of group status such as wealth, but rather on the basis of value systems internal to the groups themselves. Beyond the obvious method of simply assuming a new social identity through association with other, more attractive groups, or engaging in direct competition, Tajfel and Turner (1979) cite three means by which in dividual members of a socially subordinate group might plausibly manipulate the terms of the comparative situation in order to increase perceived group status These strategies include (1) changing the outgroup in terms of which group identity is defined, (2) comparing the ingroup to the outgroup on some new di mension, and (3) reclaiming a previously negative quality as a positive one (i.e. "black is beautiful"). Since its inception, So cial Identity Theory has undeniably dominated the theoretical discussio n of inter group behavior within social psychology. Certain of its suppositions have nonetheless fallen out of favor. Chief among SIT's more dubious
The Feeling of Power and System Justification | 12 claims is the hypothesis that self esteem acts as a primitive motivation for group identification. While th e notion of positive group identification related to t he finding, as Fiske and Taylor (p 261 ,2008 ) put it, that "people favor us' more than they explicitly disfavor them'" was supported to some extent in Tajfel's research with MGP, discrimination on the basis of group affiliation fails in empirical tests to significantly influence "trait" self esteem." Despite a transient positive self assessment following outgroup derogatory or ingroup affirmative behaviors, the effect of group discrimination on the self assessment of one's longer term trait' characteristics is null. Additionally, research has failed to support SIT's assumption that low self esteem would encourage group discriminatory behaviors (Rubin & Hewstone, 1998). For this reason, it is unlike ly that group members would necessarily act to find new group affiliations in order to increase self esteem, as SIT had previously suggested. Individual vs. Groups Over the years, a number of, some quite imaginative, have been developed, in an attempt to resolve SIT's apparent shortcomings and explain the motivational push in individuals toward group formation. In 1985, Turner revised SIT to develop Self Categorization Theory (SCT), preserving many of the assumptions of SIT, while placing a greater emphasi s on the contextual quality of inter group relations, and the socially dynamic nature of the individual actor. Depending on the salience of group distinctions in a given social context, he found that a singular actor may conceive herself as belonging to an y number of social categories. When an individual is attending class, that is, she may identify more strongly as a student, than when visiting a nursing home (in which case she might identify as 'young'), managing a cash register, or attending a political rally.
The Feeling of Power and System Justification | 13 Furthermore, people tend to behave in ways consistent with stereotypes appropriate to salient social categories. Shih, Pittinsky, and Ambady (1999) demonstrated this effect in research with Asian American women. In two experiments, researchers showe d that when participants' ethnic identity was made salient, they performed better on a simple math test, than when gender identity was made salient. Salience of group stereotypes, in other words, can promote stereotype confirming behavior. This quality of "social mobility" the tendency of individuals to identify with different social groups in different social contexts had for SIT denoted an ideological position characteristic of those individuals most inclined to abandon or dissociate from their releva nt social groupings; for SCT it is taken as a theoretical precept. Social grouping, that is to say, is first and foremost psychological, and accordingly situational in nature. Each one of us carries a vast array of group identifications available in situat ions where group based identities become relevant. Consistent with this understanding is the well known O ptimal distinctiveness theory (ODT; Brewer, 1991) which suggests that social actors aim to achieve an optimal balance between personality traits disti nct of group affiliation and those related to particular group affiliation. Perhaps surprising in light of cross cultural research, particularly the finding that non westerners tend to view themselves as more inter dependent than westerners (see Markus & K itayama, 1991), this theory has found support among western and non western part icipants alike (Eckes, Trautner & Behrendt, 2005). Though social actors may seek to balance group based characteristics and individual distinctiveness from other group members, certain individuals and certain groups may do so more than others. Research in right wing authoritarianism (RWA)
The Feeling of Power and System Justification | 14 following in the tradition of the theo ry of Authoritarian Personality has shown with considerable reliability that personalities high in RWA ( who might notably be affiliated with any political party) tend to obey, or glorify powerful leaders, and exhibit derogatory attitudes toward an outgroup more than those measuring low in RWA. When a relevant outgroup is perceived as threatening to the convi ctions of individuals high in RWA, those individuals tend more fervently to derogate outgroups and affirm ingroup beliefs (Altemeyer, 2004). In a similar vein, Social dominance theory (SDT) aims to restore to such intra personal accounts of discriminator y behavior as RWA, a high level social basis. For social dominance theorists, social hierarchies are a necessary development of human life, which those in positions of privilege unconsciously affirm through system justifying prejudices, group discriminatio ns, and practices. Accordingly, SDT, not unlike the Realistic conflict theories noted earlier, holds that political ideologies and discriminatory practices mirror group affiliation, group interest,' and fundamentally self interest within a given social order. While the Social dominance orientation (SDO) measure is similar to RWA, high SDO is seen to be representative of a primary motivation towards ingroup advantage rather than outgroup aversion. Those testing high in SDO tend to regard the world as mer ciless, violent, and characterized by brutal competition over social resources. Examples of such people include men with strong masculine identification, while examples of those testing low in SDO include ethnic minorities who do not view the social order they belong to as being legitimate (Kahn, Ho, Sidanius & Pratto, 2009). The basic distinction, with which this review of relevant psychological literature
The Feeling of Power and System Justification | 15 began, was one between high and low level approaches to group phenomena in social psychology. The relevant question, roughly, was whether individuals are basically determined in their relations to others by the groups they identify with, or their individual characteristics. SCT, ODT, and related theories have emphasized the social dynamism of the indiv idual, supplanting the status of the social group', with that of the social predicate' and lending support to the notion that behavior, discriminatory or otherwise, is determined purely by neither intra personal, nor inter group factors, but both. Socia l behavior, that is, is inter personal in origin. The perspective of SDT is notable for the reason that unlike any of the approaches described above, it bridges in a coherent way, the concept of the individual and that of the group. The individual, as Hege l expresses, takes the entirety of her meaning from the predicates attached to her, or the groups she represents. Thus it would be expected that that members of socially marginal groups should favor their own group, resist those in positions of social aut hority, and in general rate lower in SDO than high status group members. Research in SDO, however, presents us with a confound. What can be said of those belonging to an ethnic minority who do regard existing social hierarchies as legitimate? Sidanius, Pra tto & Bobo, (1994) have shown that such people tend to score high on SDO measures, further corroborating the finding that group identification is not motivated by self esteem. Anxiety, subjective uncertainty and group identification If social actors truly seek positive group affiliation habitually and unconsciously for reasons other than increasing perceived self worth, it is of course worthwhile to ask what, if any, factors motivate group identification, group formation and group discriminatory behaviors. Solutions proposed range from the tautological to the bizarre.
The Feeling of Power and System Justification | 16 One provocative answer and potent theoretical complement to SCT comes in the form of Subjective uncertainty reduction theory (SRT; Hogg, 2000). SRT claims that people are driven not positively towards group identification by a basic need for self esteem, but rather negatively, through an essential aversion to self relevant uncertainty. Hogg writes, The processes of self categorization and prototypical depersonalization responsible for socia l identification and group behaviors are well suited to subjective uncertainty reduction; they contextually assimilate self to a prescriptive prototype that guides and consensually validates perception, cognition, affect and behavior. This account of gr oup cohesion is wholly consistent with the aversive and discriminatory behaviors observed in persons rating high in either RWA or SDO. The fact that people commonly maintain group consistent beliefs in spite of disconfirming evidence, especially in cases o f perceived group threat is well supported in all realms of social psychology (Pomerantz, Chaiken & Tordesillas, 1995). Salient outgroups, alien belief systems, or other situations threatening to group identity arouse anxieties that frequently provoke stre ngthened group held convictions. A number of theories in social psychological research have in recent years constellated around this finding that anxieties grounded in subjective uncertainty may play a major role in motivating self categorization, confor mity to ingroup stereotypes, as well as group relevant discrimination. Most notable (and provocative) among these is Terror Management Theory (TMT), which imputes all forms of human behavior to a basic existential fear of
The Feeling of Power and System Justification | 17 mortality. Research in terror ma nagement has shown that individuals primed with mortality salient stimuli tend to show a marked increase in prejudice, group discrimination behaviors, and a certainty in the convictions of relevant ingroups, termed "worldview defense". Effects of mortality salience include, but are certainly not limited to, strengthened conviction in religious belief (Vail, Rothschild, Weise, Solomon, Sheldon, Pyszczynski & Greenberg, 2010), nationality, profession, political affiliations, and even increases in culturally c onsistent cons umption practices (Arndt, Solomon, Kasser, Sheldon 2004; Chatard, Arndt & Pyszczynski, 2010; Pyszczynski Greenberg, Solomon & Sheldon, 1999). R esearchers have also revealed a considerable terror management effect of "creatureliness" salien ce, supporting the assumption central to TMT that people affiliate strongly with institutions and social groupings in order to achieve a kind of "symbolic transcendence" of d eath, or animality (Goldenberg Pyszczynski, Greenberg, Solomon, Sheldon, Kluck & Cornwell 2001). This notion fits nicely with that of "timescale separation," which speaks to a basic variability in the rates of change for embedded and mutually constitutive hierarchical systems. Take a simple Hegelian example: the singular individual g ains its meaning at the most immediate level in the family that outlives it; the family, that is, operates at a higher timescale The family in turn supports the social structure in which it is embedded in our case a global capitalist economy by assimi lating its offspring through education to the various values and moral structures that maintain it; the social structure surely outlives the family and individual both, thus it operates on an even greater timescale. For those with religious convictions Go d transcends all mortality, and corresponds to a
The Feeling of Power and System Justification | 18 humanity untainted by what Goldenberg et al. (2001) call "creatureliness." In this way terror management theory in conjunction with the concept of timescale separation accounts for major aspects of religious conviction, the exceptional social power of which lies much in the promise of eternity, as well as the fervent and sometimes irrational defense of nationalist and consumer behaviors, personal vocation, and other self group or system relevant practice s which grant the possibility of "symbolically" transcending the otherwise inevitable truth of individual mortality. From the perspectives summarized above four major points can be gleaned: (1) Group identifications are formed rapidly, spontaneously, and beneath conscious awareness, varying along with social situations and context. (2) The functional dimension on which groups arise, develop, and exist, is that of social value which, though it may be vestigially related to the kinds of instrumental and eco nomic value invoked by Realistic conflict theory, is conceptually a nd operationally distinct. (3) The extent to which people do identify with and express relevant ingroup beliefs, derogatory prejudices, etc. differs between individuals and situations. And most importantly, (4) despite ephemeral increases in "state" self esteem, group, or social identification is very likely motivated not by a need for positive self relevant affect, but by an unconscious aversion to anxiety provoked by self relevant uncertai nty including those related to mortality. Thus self categorization may be motivated tautologically, by a fundamental need for socially mediated self certainty self control, or identity. Ideology and System Justification In line with the concept of time scale separation presented above, System justification theory (SJT) offers the basic distinction between three accounts of motivated
The Feeling of Power and System Justification | 19 social and inter group behavior these representing three distinct timescales. In ascending order they are ego group a nd system justifying processes. Jost, and Banaji (1994) define system justification as "the process by which existing social arrangements are legitimized, even at the expense of personal and group interest." Traditional theories of group behavior, they arg ue, have often favored ego or individual self interest as the fundamental motivating factor in social activity, and inter group discrimination. Traces of this kind of psychological rationalism can be seen in the realistic conflict theory described above, w hich grants economic self interest a primary position in group formation, and group related conflict resolution. Groups, according to RCT, form largely for the sake of competition over social resources and it is ultimately individual self interest which mu st be appealed to in order to terminate heated inter group conflicts. At the next level, these researchers address "group justification" theories as those which give motivational, attitudinal, and ideological priority to group affiliation. Group justifica tion theories, they claim, "hold that people are driven by ethnocentric motives to build ingroup solidarity and to defend and justify the interests and identities of fellow ingroup members against those of outgroup members." Among such theories, the author s cite Social identification theory/Self categorization theory, and Social dominance theory as exemplary. Though both perspectives acknowledge the plausible existence of system justifying tendencies, they unequivocally treat such cases as exceptional, or, as is the case for Social dominance theory, assign them to a restricted and theoretically specific range of influence determinable by social status. System justification theory is founded on the assumption that system justification is by no means exception al, but touches all corners of the social fabric. Whereas group related motivational influences are variable, and
The Feeling of Power and System Justification | 20 whereas social actors are mobile, system maintenance motives are pervasive. One major methodological factor differentiating research in syste m justification from that conducted in the considerably more dated group justification theories is the relatively recent innovation of implicit association testing. Where research in SDT and SIT has typically relied on explicit self report or survey data, the Implicit association test (IAT) has allowed researchers to test largely unconscious or automatic biases which most participants would refuse to acknowledge, due to a well established "social desirability" bias. Not insignificantly, Green wald, Nosek & B anaji, (2003), who developed system justification theory, have also generated much research contributing to the development of the IAT, contributing, among other things, a revised and more reliable scoring algorithm. The fact that explicit accounts of gro up discrimination diverge considerably from implicit data is well established (Rudman & K ilianski, 2000; Uleman, Saribay & Go nzalez, 2008; Rudman, Greenwald & McGhee, 2001). Using the conventional scoring metho d, Greenwald, et al. (2003) showed implicit ex plicit correlations of .11, .20, .29, and .69, respectively, for the Age, Gender Science, Race, and 2000 Election IATs. People are dramatically more explicit about, or aware of their political biases, than they are about their biases in gender, age, and ra ce. Because these latter categories carry such heavy social stigma, and because of the IAT's ability to vitiate the kinds of social desirability effects seen on self report measures, this is not surprising. System justification theory aims to give theor etical expression to ideological mechanisms which by and large extend beyond the purview of traditional "group" or "ego" justification approaches in group psychology. While ingroup favoritism and
The Feeling of Power and System Justification | 21 outgroup derogation are surely not uncommon modes of social interaction, it is equally sure that these are not the only modes of social interaction which must be accounted for. As an example, one nationally representative survey showed that African American respondents frequently accepted derogatory stereotypes of their "ingroup," and even endorsed these prejudices to a greater extent than "European American" respondents (Sniderman & Piazza, 1993). Much research in system justification processes has indeed supported the finding that members of disadvantaged, devalue d, and marginalized social groups tend to hold derogatory or conflicted views about their own group membership and comparatively favorable views of socially dominant, or advantaged "outgroups." A now well developed body of research on system ju stification has revealed system justifying tendencies within in a number of disadvantaged groups. In the early 2000's, a series of implicit and explicit tests measuring system justification across a broad range of social categories were administered publicly on the w ebsite: (http://tolerance.org). Data collect ed from these tests demonstrate a stronger effect of explicit ingroup favoritism among African American participants than for European American participants, implicit measures, however, revealed the opposite. Whi le there was no effect of implicit ingroup favoritism for African American participants, European American participants exhibited strong ingroup favoritism. The system justifying trend holds to even greater degrees (in terms of disparity between implicit a nd explicit data) for measures of age and sexual orientation as w ell (see Figure 1; Jost, Banaji & Nosek, 2004).
The Feeling of Power and System Justification | 22 From Jost, Banaji, and Nosek, 2004 Contrary to the group justification theories invoked by Jost and others, and exemplified by the social domi nance hypothesis, research has shown with a great deal of consistency that material, group, and economic interests fail considerably in predicting political affiliation (Fong, 2001; Gilens, 1999). In one study, Rudman, Feinberg, and Fairchild (2002) orga nized high and low status groups according to perceived status gap. Data collected on implicit measures alone revealed a strong trend such that the lar gest perceived status gaps (i.e. rich vs. poor) correlated with strong ingroup favoritism amongst high st atus group members, and strong outgroup favoritism amongst low status members, while smaller status gaps (i.e. Whites vs. Asians, or Christians vs. Jews) corresponded with strong ingroup favoritism for high status participants, and weak ingroup favoritism for low status participants. In addition, Feather (1979) demonstrated a clear negative correlation between income levels, as well as education levels, and conservatism, showing that conservatism is better predicted by age and sex. That the group and ego justification premises of SIT and SDO have been largely supported by self report or explicit measures exclusively is thus as predictable as the
The Feeling of Power and System Justification | 23 theories themselves. What such theories do demonstrate is an ideologically grounded, culturally conditioned and world historically embedded tendency among individuals to turn towards rational and "self interest" explanations for their behaviors and beliefs reflected strongly in explicit measure c s; what these theories and their methods cannot demonstrate is some un iversal quality of group interaction and social belief. Major examples of research done on inter cultu r al self construal, indeed, show a strong tendency toward self interest, and self oriented thinking, in "western" more than "non western" participants (Ma rkus & Kitayama, 1991). The findings of IAT research in system justification theory, of course, are not immune to criticisms similar to those directed at the SIT/SCT and SDO perspectives discussed above. Though implicit association testing clearly mitiga tes effects of social desirability, the artificial conditions under which it is necessarily administered are no more a reflection of authentic inter group and interpersonal social interactions than are those which have been utilized in support of the afore mentioned group justification paradigms. Performance on implicit association tests, indeed, appears in some cases to be highly sensitive to test design (Rudman, Greenwald & McGhee, 2001). Nonetheless, a number of studies have demonstrated system justificat ion behaviors in various real world contexts. I n research conducted by Correll, Park, Judd & Wittenbrink (2002), police officers took part in a video game simulation featuring armed and unarmed African American and white targets. Participants were given a limited time frame in which to decide whether or not the target was armed, in which case they were to shoot. Data showed that both African American and European American participants made the decision to shoot armed black targets more quickly than armed wh ite targets, and that
The Feeling of Power and System Justification | 24 both were faster in deciding not to shoot white unarmed targets than black unarmed targets. African American participants, furthermore, were no more likely to hesitate in shooting a black target, than were European American participan ts. From a simple political perspective, the findings of System justification theory are both intuitive and concerning. If social life is constituted mainly by the interactions between dominant and subservient "groups," as SDO, and other group justificat ion theories have determined, the question necessarily arises: why do low status groups so frequently act, or affiliate politically against their own interests? Furthermore, what prevents or has in the past prevented political revolutions and other forms o f serious, large scale social conflict across the world from occurring with greater frequency? Marx, in the Critique of the Philosophy of Right, begs a similar question. Why did Feudalism, a notoriously inegalitarian insti tution, subsist throughout the Mid dle A ges? The simple answer: "The Middle Ages were the democracy of unfreedom (Marx, 22)."System justification, or false consciousness, as Marx called it, is prevalent in all societies, and has served historically to conceal actual social relations from th ose who suffer them most. Countless historical examples speak to the fact that although slavery, flagrant inequality, and subjugation of ethnic, religious, or other social groups is frequent, actual antagonism between dominant and subordinate groups is rel atively rare. The reality that system justification theory supports is the same reality that Marx points to in the above quoted remark: in order for any large scale, long term social structure to endure, the social groups and individuals which comprise it must be psychologically invested if unconsciously in its preservation and maintenance. That is, people internalize the beliefs and stereotypes proper to their respective social structures, even, as we have seen,
The Feeling of Power and System Justification | 25 against their own interests, and better j udgment. Marx's discussion of ideology in the above quoted text points to a historical situation in which marginalized social actors simply lack the language to address the problem of their servitude. Where the Middle Ages were a period of actual dualism ," the modern age is one of abstract dualism." The liberal and enlightenment era discourses which arose in parallel with the institution of democracy have served, in other words, to illuminate in the abstract, the actually existing power dynamic character istic of the monarchic or feudal state a power dynamic imperceptible to those who lived it immediately. Once this dynamic is raised to the level of discourse and common understanding, a monarchy becomes difficult to maintain. Of this, the French Revoluti on is a superb example. While the political landscape of today differs considerably from that of the Middle and even Modern ages, those who criticize it frequently employ the same discourses which arose within a burgeoning democratic capitalist system ag ainst the monarchic and feudal states preceding them. The ideals of political equality, personal freedom, and private enterprise are invoked often to various political ends, fleshing out a strong distinction between private and political life, yet any curs ory analysis shows that under all extant social structures lie real status differences between people of varying economic, ethnic, religious, sexual, and gender identifications. That political pundits like Glenn Beck should invoke the ideals of "freedom" a nd "equality" to endorse an anti immigration law which institutionalizes racial profiling is evidence of the exceptional value given to these ideals (Fox News, 2010, 3:25). With knowledge of the institutionalization, conventionality, and negative impact of stereotyping and other modes
The Feeling of Power and System Justification | 26 of group discrimination so widespread, one might be inclined to ask how discriminatory modes of thinking manage to propagate. How do ideology and system justifying stereotypes operate today? One definite mode of operation for system justify ing processes is what Zanna, Jost, Manisodza, Sherman, Petrocelli & Johnson (2007) have identified as Panglossian Ideology, commonly manifest in the form of complementary stereotyping. Complementary stereotypes are those which purport to ex press the socially redeeming qualities of low status group members, and the corresponding drawbacks of high status group membership. Perceptions of poor people as "happier" and "more honest" than rich people, or African Americans as more "athletic," are pr evalent examples of complementary stereotypes. While these prejudices are sometimes endorsed by members of the groups they describe, they also serve to reinforce perceptions of marginalized groups as fundamentally distinct in nature from dominant or more p owerful social groups, thereby reifying group distinctions, rationalizing extant social hierarchies and maintaining institutions of social inequality. Research conducted by Zanna et al. (2007), indeed, showed that participants perceive a social order as mo re legitimate after exposure to complementary (poor and happy vs. rich and poor) stereotypes, than non complementary ones (poor and unhappy vs. rich and happy). Surely no better example of complementary, system justifying stereotypes can be found than in t he case of gender. Complementary Stereotyping and Gender Man is defined as a human being and a woman as a female whenever she behaves as a human being she is said to imitate the male. Simone de Beauvoir In The Second Sex, Simone de Beauvoir ad dresses the problem of womanhood.
The Feeling of Power and System Justification | 27 The question she offers is not why women are treated poorly, disliked, or generally evaluated negatively by men, but what constitutes a woman. How are women defined in relation to man? And how is a woman seen to fit into a social world generally controlled, maintained, and engineered by men? "One is not born a woman, but becomes one," de Beauvoir writes, in what is perhaps her most famous statement. Women conform to roles appropriate to women. When they do not, they are abj ected, demonized. Fitting with the assumptions of traditional approaches towards prejudice and group interaction, empirical work on gender stereotypes has in the past dealt with predominately hostile forms of sexism. In recent years, however, studies have consistently supported the finding that for both women and men, attitudes towards women are largely favorable. Eagly and Mladinic (1993) found that women, indeed, are favored to men by men and women alike, and warrant more benevolent stereotypes than men. Recent literature has given increasing focus to complementary, or "benevolent" forms of sexism chivalrous sexism, that is, which idealizes those women who are willing to accept conventional gender roles. Examples of benevolent stereotypes frequently att ributed to women include, warmth, trustworthiness, and cooperation (Rudman, Greenwald & McGhee, 2001; Rudman & Kilianski, 2000). True to the system justifying tendencies concerned here, women frequently embrace such complementary stereotypes while consist ently rejecting more outwardly negative or hostile gender stereotypes (Rudman et al. 2001; Rudman & Kilianski, 2000; Glick & Fiske, 2001; Kay, Gaucher, Peach, Friesen, Laurin & Zanna, 2009; Zanna, et al. 2007). Despite potentially widespread acceptance of such seemingly benign stereotypes by women, the effect of these stereotypes effect is unambiguously damaging. By means
The Feeling of Power and System Justification | 28 of these, men and women alike rationalize, affirm and secure a "complementary, but not equal" position for women within the social sphere Bem & Bem (1970) write, In 1954 the United States Supreme Court declared that a fraud and hoax lay behind the slogan "separate but equal." It is unlikely that any court will ever do the same for the more subtle motto that successfully keeps the woman in her place: complementary but equal" The ideological rationalization that men and women hold complementary but equal positions in society appears to be a fairly recent invention. In earlier times and in more conservative company than today it was not felt necessary to provide the ideology with an egalitarian veneer (p. 96)." The negative impact of complementary stereotyping has even seen empirical support. Using a 22 item survey distributed to 15,000 participants across 19 different countries, Glick and Fiske (2001) measured attitudes about gender, with half of the survey items presented assessing benevolent and the other half hostile stereotypes. Results showed these two modes of sexism to be overwhelmingly complementary, with a high correlation between acceptance of benevolent stereotypes by women, and overall sexism in a given nation. Responses on measures of both hostile and benevolent sexism, furthermore, independently predicted levels of national gender inequality, confirming the assertion th at benevolent stereotypes are never benign. This study also demonstrated the cross cultural prevalence of the stereotype that women are, in addition to being more communal, warm, nurturing, and refined than men, more "deserving of protection."
The Feeling of Power and System Justification | 29 The exist ence and widespread acceptance of such beliefs has clear and far reaching consequences for women and their relationships with the social world. While the kinds of stereotypes listed above surely favor women across some evaluative dimensions, they do so th rough the assignment of women to a determinate and limited position in society. That women are "deserving of protection" is indeed tantamount to saying that women are socially, physically, and financially impotent: that they are in need of protection. The gender role hypothesis put forth by Eagly (1987) suggests that people hold strong implicit beliefs about men and women, and the roles appropriate to each, which are based on traditional labor divisions of gender. If women are held to be more communal, and generally domestic than men, exposure to women in positions of authority would be expected to elicit anxiety and discomfort among men and women alike. In research measuring implicit and explicit attitudes about female authority, Rudman and Kilianski (2000) revealed a clear implicit bias against women in positions of authority among both male and female participants. Though participants' explicit attitudes co varied with explicit gender authority beliefs, men and women, regardless of their explicit beliefs, favored men in positions of power to authoritative women on all implicit measures. In spite of findings so thoroughly representative of system justification research in non complementary stereotyping, evidence exists that system justification phenomena, especially in cases of complementary gender stereotyping may be somewhat more complex. Rudman, Greenwald, and McGhee (2001) tested male and female participants in a series of implicit association tests on the gender stereotypes 'warmth' and 'potency'. Re sults showed that when a value connotation was available for either stereotype (the
The Feeling of Power and System Justification | 30 implication that men are inferior to women in warmth, or that women are inferior in potency), men and women, surprisingly, tended to score high on implicit sexism only for stereotypes that were complementary, and reflected positively on their gender. When the terms used reflected positively on both women and men, differences between genders were eliminated. Most interesting is that while implicit self esteem showed no associ ation with implicit gender stereotypes for male and female participants, implicit self concept was strongly associated with implicit gender stereotype. In other words, participants regardless of gender expressed relatively strong gender stereotypes, only w here these stereotypes were complementary for them; yet when categories typically associated with stereotypes about men and women were neutral, participants revealed no implicit stereotyping, and associated gender qualities with implicit self concepts. Tho se men and women who identified as warm, also identified their respective genders as being warm; those who identified as potent, extended potency to their gender. This strong associative link between implicit self and gender concept, has been shown in other research as well (Greenwald, Banaji, Rudman, Farnham, Nosek & Mellott, 2002). Such a finding supports the position advocated earlier that group formation is not motivated by a basic need for self esteem, but rather by anxiety related to subjective un certainty, and the parallel need for positive self concept through group identification or other means of symbolic mortality transcendence. Nonetheless, it entails a major influence of gender stereotypes on self concept for stereotypes which are reliably e xpressed by women on implicit measures. Though it is clear that group identification bears a powerful affinity to self concept, it is unclear where causality lies. In the case of gender, for instance, we might
The Feeling of Power and System Justification | 31 ask whether an individual's positive self co ncept precedes, and is subsequently generalized to her/his gender as a whole, or whether stereotypes and concepts linked to one's particular group identifications play a role in determining self concept. In fact, findings support a form of mutual causality While in the research just described, there is a causal priority of self concept when negative gender stereotypes are not salient, certain situations, particularly those in which relevant ingroup stereotypes are salient, show a clear motivational priorit y of group identification. Interestingly, research suggests that the valence and context of gender stereotypes may play a significant role in mediating the relation between group identification and self concept. Earlier, a study by Shih, et al. (1999) was described ; it showed that salience of complementary stereotypes of Asians improved the math scores of Asian women. Thus it was concluded that stereotypes alone can produce stereotype consistent behavior. Cheryan and Bodenhausen (2000) demonstrated that su ch effects do not always take place. Using an equivalent experimental paradigm, these researchers showed that when complementary stereotypes about the mathematical proclivity of Asian people were made exceedingly explicit, performance expectations of Asian American participants were raised and math scores actually decreased. This suggests that the context of ingroup salience may influence the extent to which group identification occurs. Specifically, when an individual is exceedingly conscious of group rela ted expectations, she or he may under certain circumstances dissociate from the group in question. Given research previously discussed, the largely unconscious nature of group identification, and the ubiquity of affective determination in social behavior i t is not unlikely that affect and emotion plays a major role in mediating the relation between groups and individuals.
The Feeling of Power and System Justification | 32 The fact that people are by and large inclined towards assimilating to group categories, even given relatively arbitrary modes of categ orization, points to a motivational basis in neither utility nor personality, but more likely implicates a quality universal to human cognition: a human need for cognitive parsimony, and motivation to categorize in general. How this need relates to terror management or subjective uncertainty reduction on the other hand, is unclear. In order to establish this connection in earnest, we must first consider the nature of social reasoning as well as the affective, bodily and motivational dimension of reasoning i n general. The Category Social psychology, like any academic discipline, is fraught with terms specific its object of study; no less arcane is its own vocabulary than are those particular to biology, chemistry, or psychoanalysis. Constructs such as 'socia l dominance orientation', 'terror management', and 'mood congruence', while perhaps more intuitively descriptive than the terms hysteresis', cathexis', or hydrolysis,' arise through a unique genealogy in thought, bearing with them a vast network of hist orically determined valences and associations drawn from various scientific milieus. Though science and its constructs should be treated appropriately as being the products fundamentally of a particular discursive history a figure occasionally arises wi thin academia in whom a radical shift in scientific theory and procedure are seen to be embodied. For the development of empirical social psychology, early American researcher and theorist Gordon Allport was such a figure. In addition to being one of the o riginal proponents of the empirical study of personality, Allport is responsible for having effectively founded the modern framework on which the empirical study of social reasoning, prejudice, intergroup conflict, and
The Feeling of Power and System Justification | 33 group behavior rests in his 1954 expo sition The Nature of Prejudice. Here the concepts of 'in group', 'out group', and 'stereotyping' in their full psychological significance are expounded for perhaps the first time. Despite their theoretical specificity, Allport understood each of the aforem entioned terms to be reducible to a primary process of 'categorization'. As he puts it: "Man has a propensity to prejudice. This propensity lies in his normal and natural tendency to form generalizations, concepts, categories, whose content represents an oversimplification of his world of experience. His rational categories keep close to first hand experience, but he is able to form irrational categories just as readily (27)." Allport cites five distinctive features of categorization, four of which are r elevant here: (1) "It forms large classes and clusters for guiding our daily adjustments." (2) "It assimilates as much as it can to the cluster." (3) "The category enables us quickly to identify a related object," and (4) "The category saturates all it con tains with the same ideational and emotion al flavor ." (Allport, 1954, p. 20 22). Though the immediate relevance this has to the ideological or system justifying processes concerned here is clear, the empirical literature on social perception and memory mor e generally, follows at least as closely these principles. Central to any thoroughgoing investigation of human perception indeed is the question of categorization and the re presentation or encoding of sensory data in accordance with preexisting conceptual structures or schemata. Thus the ideological question concerns above all else, cognition in general, as the basis of the
The Feeling of Power and System Justification | 34 kinds of ideological mechanisms concerned here. Categorical Cogntion If we didn't appear to men looking the way they pictu re us, they wouldn't see us at all. That's logic." Cocteau, The Infernal Machine Much of the research conducted in cognitive psychology departs from the Kantian assumption that all comprehension has as its p recondition concepts or associations which are essentially independent of, or prior to experience in short, general categories inferred from experience which simultaneously act as the conditions of possibility for experience itself. As a simple example, o ne might consider the observation of a hammer. It is only assuming that one already possesses a certain level of conceptual knowledge about hammers in general, i.e., a sense of what their qualities and attributes are, how they are held, and what they migh t be used for, that one can be said to comprehend, or know the object. Without the relevant category to give an object form, one has no substantive framework from which to approach it. This line of thinking is not at all new. Socrates offers an early exam ple of this in his Meno dialogue, when he asserts that all learning is remembrance; where phenomena are transient, categories belonging to the mind are eternal. Kant, whose critical project begins and (arguably) ends with the investigation of the grounds o f all knowledge, reduces all empirical knowledge to four categories of cognition: quantity, quality, modality, and relation. Categories, for Kant, as for the more longstanding tradition of idealism in western philosophy, define the content and the limits o f thought. Excluding the metaphysics, this principle holds some weight in empirical psychology as well.
The Feeling of Power and System Justification | 35 It is well established within cognitive psychology that all perception is to varying degrees category driven. Livingston, Andrews, & Harnad (1998) hav e defined categorical perception as the process whereby, "a continuous physical dimension such as intensity (mechanical or electromagnetic energy) or frequency (of electromagnetic radiation or of acoustic vibration) is partitioned into discontinuous or bou nded regions by some kind of perceptual mechanism." Simply stated, categorization is the assimilation of novel units to familiar classes, based on common features. Psychological research on categorical perception has taken primarily two forms: that concer ning color categories (Bornstein, 1987) and that concerning phoneme categories (e.g., Pastore, 1987). Both are categories functional at the lowest levels of human perception and communicability. In recent years, however, more direct research has been perf ormed on object categorization (Lin & Murphy, 1997), categorical learning, and the mechanisms underlying the cognitive phenomenon of unitization, or the process of construction of discrete conceptual units triggered by repeated exposure to similar percept ual or intensive configurations (Goldstone, 2000). Objects are categorized in accordance with a person's understanding of which features are integral and which are merely optional (accidental) for membership in the relevant category. In one of four exper iments conducted on object categorization, Lin & Murphy (1997) demonstrated that participants exposed to novel items belonging to two distinct categories were able to infer with relative ease which features were critical and which were non critical to memb ership in either category. Participants gave four times more positive responses to objects possessing the features integral to category membership, with differences in response patterns reliably accounted for by individual
The Feeling of Power and System Justification | 36 variations in feature knowledge. Researchers also concluded that knowledge of category membership quickens object perception, and that background knowledge of object categories facilitates the detection of object features at speeds which would otherwise render the objects indistinguishabl e. Goldstone (2000) in a similar vein, demonstrated that prolonged practice in object categorization yielded significantly faster responses in a categorization task as a function of component contiguity quantity and proximity Ease of categorization is particularly influenced by the presence or absence of category integral components within the user's "perceptual vocabulary." 1 When relevant components are present in a person's perceptual vocabulary, he or she must only direct attention to this component in order to make the appropriate categorical discrimination (see also Nosofsky, 1986; Kruschke, 1992). Categorical Social Cognition The use of categories in perception is certainly not restricted to the low level modes of cognition described above. Human beings use categories or stereotypes daily, to aid in all forms of social perception. Allport defines the stereotype as "an exaggerated belief associated with a categorywhether favorable or unfavorable. Its function is to justify (rationalize) our conduc t in relation to that category (Allport, 1954)." The term then encompasses in addition to a determined social category, our orientation toward that category and its members. Thus the very notion of a stereotype is directly suggestive of the in group/out gr oup distinction on which the social psychological edifice in the past 50 years has been in large part built. Social interactions of every variety are mediated by expectations, projections, and 1 Goldstein (2000) sees ones perceptual vocabulary' as comprising "the elementary building blocks used to create representations." Examples might include color (red, green), size, or basic shapes (circular, rectangular).
The Feeling of Power and System Justification | 37 assumptions on the part of social actors about others. Such as sumptions are formed automatically, and often within milliseconds of exposure to a given person. "By providing mental economy," write Macrae, Milne, and Bodenhausen (1994), "category activation (and application) enables perceivers to streamline cognition a nd increase the intelligibility of an otherwise dauntingly complex social world." Yet as Allport indicated, stereotypes distort available information. Personality traits can act as dominant conceptual prototypes, shaping our impression of others even at th e expense of useful trait knowledge present in the environment. Cantor and Mischel (1977) gave participants 200 trait adjectives to be rated for degree of relatedness (low, medium, high) to either one of two salient prototype concepts. During a subsequent memory recognition task, participants significantly recalled trait adjectives deemed highly related to dominant prototypes, but which were nonetheless absent from the original adjective list. This suggests that stereotypic or prototypic signifiers play a considerable role in our understanding of actual social phenomena, and quite likely predominate over, order, and determine in large part which of the available social data are salient to us. In order to conceive what the psychic determination of categoric al knowledge may look like, many psychologists have proposed topological models, introducing constructs such as "associative clustering," and "conceptual architecture." As early as 1936, Kurt Lewin, one of the originators of empirical psychology and method characterized "topological psychology" as the experimental counterpart to psychoanalysis which would offer psychoanalytic concepts the possibility of scientific legitimacy (Le win, 1936). Livingston, Andrews, and Harnad (1998) in similar fashion, speak of a
The Feeling of Power and System Justification | 38 multi dimensional "psychological similarity space," which is dynamically and morphologically compressed as a function of category learning. While objects varying along only one dimension follow with the principle that categories can be inferred by compo nent object structure, categories constructed for more complex, multi dimensional structures must be assimilated through experience. Given the deeply complex, and multidimensional nature of social information, it would seem that social categories too must be learned in order to take hold. In Macrae & Bodenhausen (2002a) participants were exposed to a series of common and uncommon male and female names. Responses for both male and female participants were substantially quicker when the names presented were f amiliar, suggesting that social categorization occurs in large part as a function of familiarity. Such methodological innovations as the implicit association test have further corroborated the finding that stereotyping is standard fare in person perception occurs automatically, and is associated with construct or category familiarity (Banaji, Greenwald, & Nosek, 2003; Greenwald, McGhee, Schwartz, 1998). Still, it is as yet unclear whence these associative structures gain their force. The answer suggested in the previous sections is by and large is simple perhaps exceedingly so: people think differently according to their particular social affinities. With culture or lesser forms of in group identification come particular modes of reasoning socially, cate gorizing others, and understanding the world. Culture is the dominant factor in determining those categories which dominate in social perception; it is the world historical mechanism that dictates which values we identify as our own, and which we disavow, and impute to an abject other, or out group.' So great is culture's
The Feeling of Power and System Justification | 39 influence on cognition, in fact, that research has supported significant differences in such basic perceptual processes as color perception according to culture (Uchikawa & Boynton, 1987 ). Categorization, then, appears to be a fairly adaptive phenomenon allowing for a markedly lighter cognitive load in even the most routine perceptual activities. Categorization also allows human beings a considerably richer representation of their immedia te environment, even if within these representations as Allport says, "even a kernel of truth may be lacking." The intimation that follows here is interesting. Categories, Allport claims, are frequently composed of "hearsay evidence, emotional projections and fantasy (Allport, p. 27)." This is especially compelling given the ostensibly disembodied 2 quality of categorical cognition. Is it really the case that human cognition is an essentially category driven, top down affair? Do human beings necessarily di sregard raw sensory data in favor of abstract, disembodied forms of knowledge? The verdict, unfortunately, is still out on this issue. While many die hard cognitivists insist that categorical knowledge is the real sine qua non of perception, others includi ng first and foremost, those following the mid century American psychologist James J. Gibson insist that the essence of perception is a preeminent sensorimotor relationship to the environment. Cognition as Emotive "Beneath all reason lies delirium; drift ." Deleuze No science is immune to the prejudices of hist ory, and in this the social sciences 1. Not to m ention the philosophical tradition which underlies it.
The Feeling of Power and System Justification | 40 are particularly suspect. Just as the work of anthropology is rooted in colonialist practice and the empirical determination of various categories of abjecti on, so has psychology developed from the soil of the clinical definition of the social Other. Foucault, in Madness and Civilization traced the discipline of psychology to the institutionalization of lepers and madmen; R.D. Laing revealed this pattern at wo rk well into the twentieth century, in his largely disregarded research on the clinical treatment of schizophrenics. The historical relativism of social psychology is emphasize d by Fiske and Taylor (2007, p. 11) albeit on a smaller scale, as they chart th e twentieth century shift seen in social psychology from the image of social actors as cognitive misers' to that of motivated tacticians', and activated actors'. Such examples make it overwhelmingly clear that any claims to objectivity made by social sc ientists should be taken with a grain of salt; however scientific' the human sciences purport to be, they shall always remain subservient in major respects to culturally dominant assumptions about what human beings ought to be. That the is' of social sci ence is always an ought', is consequent to the very structure of statistical analysis, which is in the end concerned with the (albeit provisional) categorization of people. The attribution of cognition in general to "emotional projection" was at one time the standard paradigm for psychology. With the decline of psychoanalysis, gestalt theory, and behaviorism, however, and the accompanying rise of cognitivist approaches, came a renewed interest in the structure and mechanisms underlying the storage of conce ptual information, and a greater concern for logical processes underlying human reasoning and empirical legitimacy. This trend in twentieth century psychology drew much inspiration in the sixties and later through a rising fascination with computers, the n ascent possibility
The Feeling of Power and System Justification | 41 of artificial intelligence, and the metaphor of the human organism as machine'. Yet with the increasing dominance of neurobiology in the late twentieth century has come a large scale return of empirical interest in, and restoration of l egitimacy to the notion that the human organism, in addition to being an immensely complex computational device, is above all else, an emotional machine determined not by the dictates of Universal Reason, but by various social and emotional investments, as well as a primeval, unconscious concern for efficiency. Embodied Cognition As much as group psychology rests on the theoretical distinction between groups and people, cognitive psychology rests on a parallel distinction between categories and "objects." The general strain in psychological research that favors the production and relation of categories to that of actual phenomena in the determination of subjective experiences is known as cognitivism. While no strong theory of cognition could be founded on the assumption that categories play a marginal role in the determination of subjective phenomena, it would be equally unreasonable to deny the importance of environmental stimuli. The concept of a "niche" accounts nicely for the interdependence and mutual constitution of environment and animal. Thus write Michaels and Carello in their 1948 work, Direct Perception, "if the animal is the knower, and the environment is the known, a full account of knowing (that is, perceiving) cannot be had by analyzing only one. Moreover, it is claimed that the two the animal and the niche cannot be disjoined logically since each owes its very identity to the other. An animal is what it is given that its
The Feeling of Power and System Justification | 42 niche is what it is; an animal's wings, gills, snout, or hands describe that animal's environment. Likewise, a complete description of a niche describes the animal that occupies it (pg. 14)." The environment affords the animal useful information, which is grounded in and contingent on the perceptual system of that animal. O ther notable alternatives to this perspective are found in the form of the theories of embodied and situated cognition, which posit that while categorical knowledge does play a major role in perception, categories are constantly and dynamically related to sensorimotor limitations and capacities, as well as situational information (Chaigneu, Barsalou, & Zamani, 2009). A related approach is that of "grounded cognition," which is supported to a large degree by the finding that emotion underlies all forms of co gnitive activity, including perception (Niedenthal, Setterlund & Jones, 1994), reason ing (Fiedler, Asbeck & Nickel, 1991), and memory (Forgas & Bower, 1987). The role of the body in cognition and affect is not limited to emotion and hormonal dynamics. In several studies bodily and gestural orientation were seen to play a role in decision making, and evaluation. Wells and Petty (1980) asked participants to shake (horizontally) or nod (vertically) their heads while being presented with an argument. Participa nts who nodded were found to agree with an argument more than those who shook their heads. Other research has found that people who hold a pencil in their teeth forcing them to smile found cartoons funnier than did a control group who instead held pen cils bet ween their lips (Strack, Martin & Stepper, 1998). The role of emotion in cognitive activity is especially pronounced in social
The Feeling of Power and System Justification | 43 perception, and c ategorization. In Macrae, Alnwick, Milne, Schloerscheidt (2002b), female students primed with either m ale and female faces, or random patterns, were asked to pair gender stereotypic target words with gender. Not only did gender primed participants more quickly complete the stereotyping task, but those determined to be at highest risk of c onception during t he experiment completed this task more quickly than those determined to be at low risk, suggesting that stereotyping tendencies in social perception are significantly affected at even the most immediate hormonal levels. The neurobiologist Antonio Damasio ( 1998) has also done a great deal of work implicating subcortical processes linked to emotion, which he calls "somatic markers," in all forms of reasoning. Damasio argues that psychology and neurobiology alike have neglected the predominating influence of e motion in all matters of mental activity, pushing for a thoroughgoing reconciliation of early psychological and psychoanalytic perspectives, with modern research and methods. Social Psychological perspectives on power Of all varieties of affect dealt wit h in social psychology, power is the most dynamic, mediating in a great variety of empirically observable social and non social phenomena. Due in part to the generality of this construct, and in part to the wide variety of power priming techniques utilized in an experimental context, the psychological effects imputed to power, or the feeling thereof are shown to be quite diverse, highly dynamic, and in some cases contradictory. Before discussing specific research on power, though, it is important to first make clear what precisely the term "power" does and does not denote. Early theories of power in social psychology defined this term as a potential to
The Feeling of Power and System Justification | 44 influence or control others' behavior in a social context (French & Raven, 1959). However, power is not to be conflated with status or social dominance (Lewis 2002). While social authority and high social status may relate to overall social power, these factors do not predict power in the highly situational circumstances to which power can be applied. Many scenarios can be noted in which social status and power possession do not merge. In the workplace, for example, a long time employee may be known to possess valuable insight into various problems, and thus have a great deal of influence on the behavior an d decisions of other employees; it is nonetheless quite possible that someone in a high status managerial position provides less insight to subordinate employees, and is yet granted less direct influence over the work in question. Accordingly, power is und erstood to be highly situational in nature, and may be as frequently exhibited by those who identify with low status groups as it is by high status group members. For these reasons, power is to be treated here as a basic and distinctive variety of affect, aroused by feelings of personal agency and control, which has lasting effects on behavior and social judgments (Weick & Guinote, 2008). Thus, for the purpose of clarity, the term "power related affect" has been employed here in lieu of the more general and elusive term utilized in much of the literature. Numerous social and perceptual effects of power have been noted as well, including increases in goal directed activity (Guinote, 2007a), attentional flexibility (Guinote, 2007b), a tendency to focus on "t he big picture" rather than minute details (Smith & Trope, 2006) and stereotyping (Vesc io, Gervais, Heidenreich & Snyder, 2006; Guinote, Willis & Martellotta, 2010). While the tendency of power to increase implicit prejudices and stereotyping behavior is w ell supported, some studies have demonstrated
The Feeling of Power and System Justification | 45 the opposite effect, such that powerful individuals demonstrate greater cognitive variability, and a pronounced tendency to disregard irrelevant environmental information and attend to relevant environmental in formation in forming social judgments. Thus, powerful individuals are more inclined to rely on situational information about individuals than prominent stereotypes when such information is available, although these same individuals display heavier reliance on, ease of retrieval for, and confidence in automatic judgments. In order to account for such contradictory findings, Guinote (2007a) developed a Situated focus theory (SFT) of power. Specifically, SFT proposes that power related affect is entirely si tuational, and that it concerns personal agency, independence, and a decreased reliance on others, more so than direct control over others. The concrete influence that power relevant affect has, that is, is contingent on social context. Additionally SFT suggests that the reason for contradictory or paradoxical effects of power related affect on stereotyping is the fact that power related affect promotes what Weick, and Guinote (2008) call "temporal stability" in the judgments of individuals. In other wor ds, powerful individuals tend to adhere more strictly to their previous decisions than others. Among the many plausible asocial consequences of the "experience of power" are benefits for reproduction, cardiovascular health (Sapolsky, 2004), i mmune function (Sheridan, Stark, Avitsur & Padgett, 2000), neurological health and espe cially mental health (Buwalda Kole, Veenema, Huininga, de Boer & Korte, 2005). While many confounding factors likely contribute to these physiological outcomes, one vital influence, given the hormonal dynamics involved, is power's role in mitigating, and
The Feeling of Power and System Justification | 46 eliminating social stress. Indeed, it is not implausible, in light of the earlier assertion made that group and system justification tendencies relate to subjective uncertainty and s urrounding anxieties, that mitigation of social stress is the primary affective influence of power related affect. Given these observations, and social power's plausible role in ameliorating along with stress, depression and neuroticism both seen to be rel ated to system justification in marginalized group members is it possible that prompting the experience of power among low status group members might alleviate, or even eliminate the kinds of System Justification effects otherwise seen? The Present Study Theoretically, the purpose of this study is to integrate existing research on the cognitive impact of subjective power, or power related affect, and System justification theory's claims surrounding the ubiquity of system maintenance motives in human bein gs. This integration is realized through subjective uncertainty reduction and terror management theory's mutual claim that system justification and group identification alike are motivated in some fashion by subjective anxiety and the negative affect that constitutes it. The basic theoretical ass umption of this research is : power related affect mitigates system justification tendencies by combating subjective uncertainty related affect at the level on which it originates. In other words, ideology, or the system justification motive, is theorized not to be an essential quality of human social cognition, but a phenomenon whose source is simultaneously physiological, cognitive, and social: an anxiety by which human beings are unconsciously compelled to seek o rder. Though the anxiety in question
The Feeling of Power and System Justification | 47 is frequently resolved through social identification and system maintenance strategies, it is possible that subjective power, or power related affect might also serve to mitigate subjective anxieties, and thereby reduce the need for system justification. To test these assumptions an experiment was conducted that assessed the relationship between power related affect and system justification. This experiment aimed first to demonstrate a system justification effect of gen der, through explicit and implicit measures of attitudes about women and men in positions of authority, and second, to show an effect of power related affect on system justification. Power related affect was stimulated through either a priming condition re quiring participants to flex their muscles, or a priming condition in which participants were asked to recall and write about a situation in which they felt "independent, strong, and in control." Method Participants One hundred and ninety seven undergra duate and graduate students volunteered to participate in this study. All participants were over the age of 18 and enrolled at various universities within the United States. Data from 4 volunteers were discarded due to abnormally large or small response la tencies on the IAT, leaving 193 participants (107 female, 86 male). Participants were all contacted online in two discrete sampling phases, roughly three weeks apart. Using a snowball method, an initial sample of 112 participants (73 female, 39 male) was a cquired through the public social networking site: http://facebook.com. Later, a post was made to the largely anonymous psychology forum: http://reddit.com/r/psychology, from which second sample of 81 participants (34 female, 47 male) was established.
The Feeling of Power and System Justification | 48 Each student was randomly assigned to one of three pr iming conditions including, (a) an embodi ed power priming condition, (b) a verbal p ower priming condition, and (c) a control condition that included no priming task. Though the implicit association test (IAT) and accompanying survey are accessible to all on: http://aufbees.com/experiment, data were filtered according to responses on a number of demographic survey items. Only data belonging to American students, over the age of 18 were recorded. Other than these demographic responses, no information was stored about participants. Therefore participants remained anonymous in completing both the IAT and the survey portion of this experiment. For further information on the respective frequencies of gender, an d priming condition across sample, see Tables 1, 2, and 3. Materials Priming Materials. Implicit attitudes were assessed using a priming measure. Three written prompts served as primes. Participants in the embodied priming condition received a message asking the test recipient to "flex [their] muscles" for thirty seconds. Tho se in the written power priming condition were asked to recall and describe in one paragraph or less a situation in which they felt themselves to be powerful, while those in a contr ol condition received no prime. Following their respective primes, all participants were presented with an excerpt from a news article, which they were asked to read carefully. Primes for each condition can be found in Appendix A. Implicit Association Tes t. Stimuli for the IAT portion of the experiment consisted of 30 target words: 8 familial roles typically assigned to women, 8 familial roles typically assigned to men, 7 words signifying power, to be paired with the label: strong', and 7 words signifying subservience, or powerlessness to be paired with the label: weak'. The
The Feeling of Power and System Justification | 49 strength and weakness salient words were generated by the author. The selected male and female salient words were borrowed from the gender science IAT provided at: https://implicit .harvard.edu/implicit/demo. All target words were color coded according to the category to which they belonged. Words descriptive of gender roles and those related to strength, appeared in white and green respectively. All target words were judged to be un ambiguously classifiable as belonging to their respective categories by the author. The 30 stimuli words utilized in this experiment, along with examples of all stages of the IAT are included in Appendix B. The test was administered using custom software written in the java programming language, and was composed of 7 distinct trial blocks, each with a novel arrangement, as well as a unique, randomly generated sequence of words. Each block began with a screen displaying the novel arrangement, as well as b asic instructions specific to that block. A variable number of trials (either 20 or 40, depending on block) followed instruction, each exhibiting a new target word. Importantly, the presentation order alternated between participants such that approximately half were given a 'female'/'strong' pairing before receiving a pairing of 'male'/'strong', while half received these in the opposite order. Statistical analysis, nonetheless, reflected no difference attributable to order on either explicit ( T ( 1 ) =0.54, p= 0. 59 ) or implicit ( T= 0.13, p= 0.9) measures. The basic arrangement of this IAT can be found in Table 4. Self Report Measure. Explicit attitudes were assessed by Rudman and Kilianski's (2000) Gender and Authority Measure (GAM). The GAM contains 15 items o n which respondents indicate preference for male versus female authorities. Each item was rated according to level of agreement on a six point scale, with 5 expressing strong agreement
The Feeling of Power and System Justification | 50 and lower scores expressing greater levels of disagreement. Examples of items include: "I would rather work for a man than a woman" and "In general, I would rather take orders from a man than a woman". The GAM is scored by averaging respondents' agreement with the 15 items, after reverse scoring appropriate items. High score s indicate preference for male versus female authorities. In the original analysis, the GAM showed adequate internal consistency. The contents of this survey are listed in Appendix C. Apparatus. The IAT was presented using Java Applet, PHP, and MySQL data base technology. For each individual session of the implicit association test, stimuli was presented in the form of an applet, using the respondent's computer, and user response latencies were analyzed according to a scoring algorithm internal to that appl et. Test results for each trial were determined as showing either strong, medium, or weak association contrasts either in favor of male or female "strength". For example, results reflected the relative strength of participants' automatic preferences for m ale' versus female' gender, or the differential association of male versus female, with strong'. Individual latencies were measured relative to the clock rate particular to the operating system supporting the respondent's Web browser (e.g., 18.2 Hz for W indows systems). This limitation, however, was not significantly detrimental to the integrity of data due to the noise generated in averaging data over approximately 200 trials. Procedure Participants were presented first with an instruction page and co nsent form informing them that they would be completing a psychology experiment that assessed attitudes about gender. They were asked to read this and press continue' as soon as they were ready. Next, participants were presented with a series of question s determining
The Feeling of Power and System Justification | 51 whether their data were to be stored for analysis, before receiving the prompt appropriate to their priming condition. Those in the No prime control condition received no initial prime. Following the priming task, students were referred to a default task (reading and memorizing a newspaper article) before receiving the IAT proper and the subsequent GAM, a 15 item survey. This experiment was a 3(priming condition) x 2(gender) mixed factorial design. The implicit dependent variable was the d ifferential score calculated in accordance with the revised IAT scoring algorithm developed by Banaji, Greenwald, & Nosek (2003). This IAT was developed in accordance with the guidelines established in that article. Mean scores for the GAM constituted a second, explicit dependent variable. Detailed information on the programming aspect of this project, including annotated Java and PHP code, is featured in Appendix D. Results The results of IAT scores for all participants were analyzed using a 3(priming co ndition) x 2 (gender) mixed factorial ANOVA. Analysis showed a strong significant main effect of gender, F ( 1 ) =21.03, p =<.0001, but no significant effect of priming condition, F ( 2 ) =0.04, p =0.96, and no significant interaction between the two, F ( 2 ) =1.03, p = 0.36. Men tended to score substantially higher than women on implicit measures of sexism ( M= 0.34), though women also showed implicit associations favoring male to female authority ( M= 0.13; see Table 5). The effect of gender on explicit measures of preferen ce for male authority was also quite strong. Scores on the 15 item GAM, were averaged and treated as a second dependent variable. A 3(priming condition) x 2(gender), mixed factorial ANOVA on GAM means, revealed significant differences between
The Feeling of Power and System Justification | 52 female ( M= 1.7 0 ) and male ( M =2.08) scores, F ( 1 ) = 12.21, p =0.0006. However, neither prime, F ( 2 ) = 0.01, p= 0.99, nor interactions between gender and prime, F (2) = 1.07, p= 0.35, could account for significant differences in measures of explicit sexism (see Table 6). Early ana lyses of data collected from Facebook participants exclusively, revealed no effect of gender on self report measures of sexism. In a 3(priming condition) x 2(gender) mixed factorial ANOVA performed on data from the initial Facebook sample, a significant ef fect of gender on implicit sexism was revealed, F ( 1 ) =10.78, p =0.0014, while no significant effect was shown for gender on GAM scores, F ( 1 ) =1.83, p= 0.18. Neither were significant results demonstrated among these participants for implicit effects of priming condition, F (2) =0.16, p= 0.85, implicit effects of gender prime interaction, F ( 2 ) =0.2, p= 0.81, explicit priming effects, F ( 2 ) =0.29, p= 0.75, or explicit gender prime interaction effects, F ( 2 ) =0.2, p= 0.81 (see Tables 7, and 8). In light of the discrepancy be tween self report data gathered in the original Facebook sample, and that represented in the final cumulative sample, it seemed likely that the two sources of volunteers online represented slightly different populations. This was supported in a t test whic h revealed a considerable effect of online sampling method (Facebook vs. Reddit) on explicit sexism scores. That is, participants who were recruited from Reddit scored an average of 2.20 out of a possible 5, on the GAM survey, while those recruited from Fa cebook scored an average of 1.63. Results of this t test demonstrated strong statistical reliability, T ( 1)= 5.66, p=<0.0001 A separate Analysis of Variance was performed on the second participant pool, composed entirely of volunteers recruited on Reddit. This analysis was also based on a
The Feeling of Power and System Justification | 53 3(priming condition) x 2(gender), mixed factorial model. The Reddit sample failed to show any unique trends, revealing no significant effect of prime for both implicit, F ( 2 ) =0.54, p= 0.59, and explicit, F ( 2 ) =0.55, p= 0.55, scores, as well as significant effects of gender on both implicit and explicit measures of sexism: F ( 1 ) =8.90, p= 0.0038, and F ( 1 ) =4.7, p= 0.03, respectively. Nonetheless, a non significant gender prime interaction on implicit measures, F ( 2 ) =2.08, p= 0.13, was noteworthy for being substantially closer to significance than any of the other samples analyzed. This, in conjunction with the exceptionally low cell mean expressed by female participants in the written prime condition for IAT scores, prompted deeper ana lysis of the second sample (see Tables 9, and 10). Though Analysis on the three samples (Facebook, Reddit, and All) has demonstrated consistently a general failure of the various priming tasks to produce any significant effect on either implicit or explic it measures, a final ANOVA, restricted to female participants from the second sampling population (Reddit) revealed a marginally significant effect of priming on implicit attitudes, F ( 2 ) = 2.55, p =0.09, in roughly the direction predicted, such that those f emale participants who had received a written power prime, in fact scored negatively (in favor of female authority) on implicit sexism ( M= 0.03 SD =.34 ), where all other cell means for the IAT were positive. Furthermore, the low significance of this effec t appears to be due in large part to the limited sample size of this cell. Finally, a series of Pearson correlations were run on the two dependent variables (explicit and implicit sexism) for all cells. None of these yielded a significant correlation.
The Feeling of Power and System Justification | 54 D iscussion The results reported above point to at least two important findings. First, and most reliable, is the effect of gender on explicit and especially implicit attitudes about gender roles and authority. While data yielded by the explicit Gender and A uthority Measure (GAM) consistently supported the assumptions of what have previously been referred to as "group justification theories," revealing greater preferences for male authority among men than women, these data, as predicted, showed no substantial evidence of direct system justification tendencies among women. This was hypothesized to be due to a general social desirability effect common in explicit measures. The effect of gender on implicit sexism was shown to be surprisingly uniform across sampl es. In each sample analyzed, mean scores of male participants on the IAT were .34, while those for female participants were .13. (S ee Tables 5, 7, and 10). As predicted, these results show clear evidence of implicit system justification tendencies among wo men, as well as a strong "group justification" effect such that male participants implicitly favored male authority more than women. Nonetheless, women still held implicit biases favoring male authority to female authority. Differences between samples on explicit measures could be a consequence of the anonymous nature of the Reddit forums and the highly public nature of Facebook. Another possibility is that those participants recruited on Facebook reflected in large part the original recruitment pool, of w hich at least 90% of students included were enrolled at a particular liberal arts school in southern Florida. Though it can only be speculated, the level of attention given to Feminist discourse at this institution may be abnormally high, and this may enco urage social desirability effects favoring female to male authority
The Feeling of Power and System Justification | 55 among men. Given the absence of the appropriate demographics in the data, however, it is impossible to validate such speculations. The second major finding in this study, though statisti cally modest, was that among a certain population, the recollection, or feeling of power may indeed, play some role in mitigating system justifying effects. The more general failure of priming exercises in significantly influencing implicit gender associat ions may have been due in part to the quality of the primes utilized, and the context of the experiment. Because participants were not actually monitored during the priming phase, it is possible that some, especially those in the embodied power priming con dition, failed to complete primes altogether. A second issue concerns the written power priming condition. Given that participants in this condition were actually required to write and submit something, it is plausible that they were more likely than those in the embodied condition to actively engage in the desired power recognition process. The level of engagement, indeed, may account for the pronounced cell means of participants in this condition. This requirement, however, may also have deterred many fro m completing the experiment. Because the assignment of participants to their respective conditions was sequential, the relatively low frequency of participants in the written power condition can point only to a relatively high attrition rate among partici pants assigned to this condition. Importantly, the low frequency of written power primed participants necessarily contributes to the statistical weakness of this condition. The research presented here supports the hypothesis that explicit measures of sexi sm tend to reflect social desirability biases primarily, and that these will accordingly reflect the assumptions of "group justification theories" such as SIT/SCT. Additionally,
The Feeling of Power and System Justification | 56 data collected from the IAT supported a powerful system justification effect a mong women and men alike, as well as a moderate group justification effect. Though data did not adequately support the hypothesis that power priming would mitigate system justification tendencies in women, some evidence for this was found. Given the issue s presented above, future research may try to ensure equal frequencies across all priming conditions. Substantial revision of priming methods should also be considered.
The Feeling of Power and System Justification | 57 References Adorno, T.W., Frenkel Brunswick, E., Levinson, D.J, & Sanford, R. N. (1 950). The Authoritarian personality. New York: Harper. Allport, Gordon W. The Nature of Prejudice. Reading, MA: Addison Wesley, 1954. Altemeyer, B. (2004). "The Other Authoritarian Personality." Pp. 85 107 in J. T. Jost & J. Sidanius, (eds.), Political Psychology. New York: Psychology Press. Arndt, J, Solomon, S, Kasser, T., & Sheldon, K. (2004). The Urge to Splurge: A Terror Management Account of Materialism and Consumer Behavior. Journal of Consumer Psychology 14 (3), 198 212. doi: 10.1207/s15327663jc p1403_2. Bem, S. L., & Bem, D. J. (1970) Case study of a nonconscious id eology: Training the woman to know her place. In D. J. Bem (Ed.), Beliefs, attitudes, and human affairs (pp. 89 99). Belmont, CA: Brooks/Cole Publishing Co. Billig, M. & Tajfel, H. ( 1973). Social categorization and similarity in intergroup behavior. European Journal of Social Psychology, 3 27 52. Brewer, M. B., (1991). The social self: On being the same and different at the same time. Personality and Social Psychology Bulletin 17 475 482. Buwalda, B., Kole, M. H.P., Veenema, A.H., Huininga, M., de Boer, S.F., Korte, S.M (2005). Long term effects of social stress on brain and behavior: a focus on hippocampal functioning. Neuroscience and Biobehavioral Reviews, 29 ,83 97 Cantor, N. & Mischel, W. (1977). Prototypes in person perception. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 12, pp. 3 52). New York: Academic Press. Chaigneau, S.E., Barsalou, L.W., & Zamani, M. (2009). Situational information
The Feeling of Power and System Justification | 58 contrib utes to object categorization and inference. Acta Psychologica, 130 81 94 Chatard, A., Arndt, Jamie, & Pyszczynski, T. (2010). Loss Shapes Political Views? Terror Management, Political Ideology, and the Death of Close Others. Basic and Applied Social Psy chology 32 (1), 2 7. Cheryan, S., & Bodenhausen, G. V. (2000). When positive stereotypes threaten intellectual performance: The psychological hazards of "model minority" status Psychological Science 11 399 402. Correll, J., Park, B., Judd, C. M., & Wittenbrink, B. (2002). The police officer's dilemma: Using ethnicity to disambiguate potentially threatening individuals. Journal of Personality and Social Psychology, 83 1314 1329. Damasio, a R. (1998). Emotion in the perspective of an integrated nervo us system. Brain research. Brain research reviews 26 (2 3), 83 6. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/9651488. Eagly, A. (1987) Sex differences in social behavior: A social role interpretation Hillsdale, NJ: Erlbaum. Eagly, A. H., & Mladinic A. (1993). Are people prejudiced against women? Some answers from research on attitudes, gender stereotypes, and judgments of competence. In W. Stroebe & M. Hewstone (Eds.), European review of social psychology (Vol. 5, pp. 1 35). New York: Wiley. Ecke s, T., Trautner, H. M., & Behrendt, R. (2005). Gender subgroups and intergroup perception: adolescents! views of own gender and other gender groups. The Journal of social psychology 145 (1), 85 111. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/15739780.
The Feeling of Power and System Justification | 59 Engels, Friedrich, Karl Marx, and Robert Cinnamond Tucker. The Marx Engels R eader New York: Norton, 1978. Print. Feather, N. T. (1979). Value correlates of conservatism: Journal of Personality and Social Psychology Vol 37(9) Sep 1979, 1617 1630. Fiedler,K., Asbeck, J., & Nickel, S. (1991). Mood and Constructive Memory Effects on Social Judgment. Cognition and Emotion, 5, 363 378. Fiske, Susan T., & Shelley E. Taylor. Social Cognition: from Brains to Culture Boston: McGraw Hill Higher Education, 2008. Print. Fong, C. (2001). Social preferences, self interest, and the demand for re distribution. Journal of Public Economics, 82, 225 246. Forgas, J.P., & Bower, G. H. (1987). Mood effects on person perception judgments. Journal of Personality and Social Psychology, 53, 53 60. Fox News (Producer). (2010) Glenn Beck on Illegal Immigration [video]. Retrieved April 15, 2011 from http://www.youtube.com/watch?v=r79 Za5tj9E. French, J., & Raven, B. H. (1959). The bases of social power. In D. Cartwright (Ed.), Studies in social power (pp. 150 167). Ann Arbor, MI: Institute for Social Research Gilens, M. (1999). Why Americans hate welfare: Race, media, and the politics of antipoverty policy. Chicago: University of Chicago Press. Glick, P., & Fiske, S. T. (2001). An Ambivalent Alliance. American Psychologist 56 (2), 109 118. Goldenberg, J. L., Pyszczynski, T., Greenberg, J., Solomon, Sheldon, Kluck, B., & Cornwell, R. (2001). I am not an animal: Mortality salience, disgust, and the denial of
The Feeling of Power and System Justification | 60 human creatureliness. Journal of Experimental Psychology: General 130 (3), 427 435. Greenwald, A. G., B anaji, M. R., Rudman, L. A., Farnham, S. D., Nosek, B. A., & Mellott, D. S. (2002). A unified theory of implicit attitudes, stereotypes, self esteem, and self concept. Psychological Review, 109 (1), 3 25. Greenwald, A. G., McGhee, D. E., & Schwartz, J. L. ( 1998) Measuring individual differences in implicit cognition: The implicit association test. Journal of personality and Social Psychology, 74(6), 1464 1480. Greenwald, A. G., Nosek, B. a, & Banaji, M. R. (2003). Understanding and using the Implicit Associa tion Test: I. An improved scoring algorithm. Journal of Personality and Social Psychology 85 (2), 197 216. Guinote, A. (2007a). Power and goal pursuit. Personality and Social Psychology Bulletin, 33 (8), 1076 1087 Guinote, A. (2007b). Power affects basic cognition: Increased attentional inhibition and flexibility. Journal of Experimental Social Psychology, 43, 685 697. Guinote, A., Willis, G. B., & Martellotta, C. (2010). Social power increases implicit prejudice. Journal of Experimental Social Psyc hology 46 (2), 299 307. Elsevier Inc. doi: 10.1016/j.jesp.2009.11.012. Hogg, Michael A. (2000). Subjective Uncertainty Reduction through Self categorization: A Motivational Theory of Social Identity Processes. European Review of Social Psychology, 11, 22 3 255. Jost, J. T., & Banaji, M. R. (1994). The Role of stereotyping in system justification and the production of false consciousness. British Journal of Social Psychology, 33, 1
The Feeling of Power and System Justification | 61 27. Jost, J., Banaji, M. & Nosek, B. (2004). A Decade of System Justificat ion Theory: Accumulated Evidence of Conscious and Unconscious Bolstering of the Status Quo. Political Psychology, 25(6) 881 919. Kahn, K., Ho, a K., Sidanius, J., & Pratto, F. (2009). The Space between Us and Them: Perceptions of Status Differences. Gro up Processes & Intergroup Relations 12 (5), 591 604. Kay, A. C., Gaucher, D., Peach, J. M., Friesen, J., Laurin, K., Zanna, M. P., et al. (2009). Inequality, discrimination, and the power of the status quo: Direct evidence for a motivation to view what i s as what should be. Journal of Personality and Social Psychology, 97, 421 434. Lewin, Kurt. "Psychoanalysis and Topological Psychology." The Complete Social Scientist: a Kurt Lewin Reader Washington, DC: American Psychological Association, 1999. 67 74. Print. Lewis, R. J. (2002). Beyond dominance: The importance of leverage. Quarterly Review of Biology, 77, 149 164. Livingston, K. R., Andrews, J. K., & Harnad, S. (1998). Categorical perception effects induced by category learning. Journal of experiment al psychology. Learning, memory, and cognition 24 (3), 732 53. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/9606933. Macrae, N.C., Milne, A.B., Bodenhausen, G.V., (1994) Stereotypes as energy saving devices: A peek inside the cognitive toolbox. Jour nal of Personality and Social Psychology. 66(1 ) 37 47.
The Feeling of Power and System Justification | 62 Macrae, C. N. & Bodenhausen, G. V. (2002a). Social cognition: Categorical person perception. British Journal of Psychology, 92, 239 255. Macrae C.N., Alnwick K.A., Milne A.B., Schloerscheidt A.M. (20 02b). Person perception across the menstrual cycle: Hormonal influences on social cognitive functioning. Psychological Science, 13, 532 536. Michaels, C.F., & Carello, C. (1981). Direct Perception Englewood Cliffs, NJ: Prentice Hall. Markus, H. R., & Ki tayama, S. (1991). Culture and the self: Implications for cognition, emotion, and motivation. Psychological Review 98 (2), 224 253. Niedenthal, P.M., Setterlund, M.B., & Jones, D.E. (1994). Emotional organization of perceptual memory. In P.M. Niedenthal & S. Kitayama (Eds.) The heart's eye: Emotional influences in perception and attention. NY: Academic Press. Pomerantz, E. M., Chaiken, S., & Tordesillas, R. (1995). Attitude strength and resistance processes. Journal of Personality and Social Psychology 69 408 419. Pyszczynski, T., Greenberg, J., & Solomon, Sheldon. (1999). A dual process model of defense against conscious and unconscious death related thoughts: An extension of terror management theory. Psychological Review 106 (4), 835 845. Rubin, M., & H ewstone, M. (1998). Social Identity Theory's Self esteem Hypothesis: A Review and some questions for clarification. Personality and Social Psychology Review, 4(2), 40 62. Rudman, L. A., Feinberg, J. M., & Fairchild, K. (2002). Minority members' implicit attitudes: Ingroup bias as a function of group status. Social Cognition, 20, 294 320.
The Feeling of Power and System Justification | 63 Rudman, L. A., & Kilianski, S. E. (2000). Implicit and Explicit Attitudes Toward Female Authority. Personality and Social Psychology Bulletin 26 (11), 1315 1328. Rudman L. a, Greenwald, a. G., & McGhee, D. E. (2001). Implicit Self Concept and Evaluative Implicit Gender Stereotypes: Self and Ingroup Share Desirable Traits. Personality and Social Psychology Bulletin 27 (9), 1164 1178. Sapolsky, R.M. (2004). Social status and health in humans and other animals. Annual Review of Anthropology, 33, 393 418. Sheridan, J.F., Stark, J.L., Avitsur, R., & Padgett, D.A. (2000). Social disruption immunity, and susceptibility to viral infection: Role of glucocorticoid insensiti vity and NGF. Annals of the New York Academy of Sciences, 917, 894 905. Sherif, M., Harvey, O. J., White, B. J., H ood, W. R., & Sherif, C. W. (1954). Intergroup conflict and cooperation: the Robbers Cave experiment Norman: University of Oklahoma Book Exc hange. Shih, M., Pittinsky, T.L., & Ambady, N. (1999). Stereotype susceptibility: Identity salience and shifts in quantitative performance. Psychological Science 10 80 83. Sidanius, J., Pratto, F., & Bobo, L. (1994). Social dominance orientation and the political psychology of gender: A case of invariance? Journal of Personality and Social Psychology, 70, 476 490. Smith, P.K. & Trope, Y. (2006). You focus on the forest when you're in charge of the trees: Power priming and abstract information processin g. Journal of Personality and Social Psychology, 90, 578 596. Sniderman, P., & Piazza, T. (1993). The Scar of Race Cambridge, Mass.: Harvard University Press.
The Feeling of Power and System Justification | 64 Strack, F., Martin, L.L., & Stepper, S. (1988). Inhibiting and facilitating conditions of the human smile: A nonobtrusive test of the facial feedback hypothesis. Journal of Personality and Social Psychology 54 768 777. Tajfel, H. (1970). Experiments in intergroup discrimination. Scientific American, 223 96 102. Tajfel, h., Billig, M., Bundy, R. P., & Flament, C. (1971). Social categorization and intergroup behavior. European Journal of Social Psychology, 27, 27 36. Tajfel, H., & Turner, J. C. (1979). An integrative theory of intergroup conflict. In W. G. Austin & S. Worchel (Eds.), The Socia l Psychology of intergroup relations. Monterey, CA: Brooks/Cole. Uchikawa, K. & Boynton, R. (1987).Categorical color perception of Japanese observers: Comparison with that of American participants. Vision Research, 20(10), 1825 1833. Uleman, J. S., Adil S aribay, S., & Gonzalez, C. M. (2008). Spontaneous inferences, implicit impressions, and implicit theories. Annual review of psychology 59 329 60. Vail, K. E., Rothschild, Z. K., Weise, D. R., Solomon, Sheldon, Pyszczynski, T., & Greenberg, J. (2010). A terror management analysis of the psychological functions of religion. Personality and social psychology review : an official journal of the Society for Personality and Social Psychology 14 (1), 84 94. Vescio, T. K., Gervais, S.J., Heidenreich, S., & Snyd er, M. (2006). The effects of prejudice level and social influence strategy on powerful people's responding to racial out group members. European Journal of Social Psychology, 36, 435 450.
The Feeling of Power and System Justification | 65 Weick, M., & Guinote, A. (2008). When subjective experiences mat ter: power increases reliance on the ease of retrieval. Journal of personality and social psychology 94 (6), 956 70. Wells, G. L., & Petty, R. E. (1980). The effects of head movement on persuasion: Compatibility and incompatibility of responses. Basic and Applied Social Psychology, 1, 219 230. Zanna, M.P., Kay A.C., Jost, J.T., Mandisodza, A.N., Sherman, S.J., Petrocelli, J.V., Johnson, A.L. (2007). Panglossian ideology in the service of system justification: How complementary stereotypes help us to ratio nalize inequality. In Zanna M.P. (Ed.), Advances in experimental social psychology (Vol. 38, pp. 305 358). San Diego, CA: Academic Press. !i"ek, Slavoj. Mapping Ideology London: Verso, 1999. Print.
The Feeling of Power and System Justification | 66 Tables Table 1 Total Frequency According to Gender and Prime Priming Condition Gender Written Embodied None overall Female 29 43 35 10 7 Male 18 38 30 86 overall 43 81 69 193 Table 2 Frequency According to Gender and Prime of Facebook Participants Priming Condition Gender Written Embodied None overall Female 16 29 28 73 Male 9 17 13 39 overall 46 41 25 112 Table 3 Total Frequency According to Gender and Prime of Reddit Participants Priming Condition Gender Written Embodied None overall Female 9 14 11 34 Male 9 21 17 37
The Feeling of Power and System Justification | 67 overall 18 35 28 81 Table 4 Sequence of Blocks in the Implicit Association Test Block N o. of trials Function Items assigned to left key Items assigned to right key 1 20 Practice Male Female 2 20 Practice Weak Strong 3 20 Practice Male + Weak Female + Strong 4 40 Test Male + Weak Female + Strong 5 20 Practice Female Male 6 20 Practice Female + Weak Male + Strong 7 40 Test Female + Weak Male + Strong Table 5 Cell mean s of IAT score according to gender and prime f or all participants Priming Condition Gender Written( SD ) Embodied( SD ) None( SD ) overall Female .07( .3 2 ) .15( .34 ) .14( .32 ) .13 Male .41(. 35 ) .33( .32 ) .31( .26 ) .34 overall .22 .24 .2 1 Table 6 Cell means of mean self report (GAM) scores according to gender and prime for all participants
The Feeling of Power and System Justification | 68 Priming Condition Gender Written( SD ) Embodied( SD ) None( SD ) overall Female 1.60( 65 ) 1.76( .78 ) 1.70( .69 ) 1.70 Male 2.24(. 35 ) 2.00( .67 ) 2.06( .74 ) 2.08 overall 1.87 1.88 1.86 Table 7 Cell mean s of IAT score according to gender and prime for the original Facebook sample Priming Condition Gender Written( SD ) Embodied( SD ) None( SD ) overall Female .14 ( .3 4 ) .10( .34 ) .14( .33 ) .13 Male .41(. 28 ) .34( .34 ) .31( .36 ) .34 overall .24 .19 .19 Table 8 Cell means of mean self report (GAM) scores according to gender and prime for Facebook sample Priming Condition Gender Written( SD ) Embodied( SD ) None( SD ) overall Female 1.43( 60 ) 1.60( .83 ) 1.60( .07 ) 1.56 Male 1.76(. 58 ) 1.8 1( .60 ) 1.65( .50 ) 1.75 overall 1.55 1.68 1.61 Table 9
The Feeling of Power and System Justification | 69 Cell mean s of IAT score according to gender and prime for Reddit sample Priming Condition Gender Written( SD ) Embodied( SD ) None( SD ) overall Female 0.03( .3 4 ) .24( .32 ) .13( .30 ) .13 Male .41(. 35 ) .33( .32 ) .31( .26 ) .34 overall .19 .30 .24 Table 10 Cell means of mean self report (GAM) scores according to gender and prime for Reddit sample Priming Condition Gender Wri tten( SD ) Embodied( SD ) None( SD ) overall Female 1.90( 66 ) 2.10( .58 ) 1.96( .63 ) 2.00 Male 2.73(. 58 ) 2.15( .70 ) 2.38( .74 ) 2.35 overall 2.32 2.13 2.20
The Feeling of Power and System Justification | 70 Appendix A Figure 1. Initial demographics survey used to determine whet her or not to store data. Figure 2. Prompt for the Embodied power priming condition.
The Feeling of Power and System Justification | 71 Figure 3. Prompt for the written power priming condition.
The Feeling of Power and System Justification | 72 Figure 4. Default prompt shown in each condition
The Feeling of Power and System Justification | 73 Appendix B Figure 1. Instruction s for Block 1
The Feeling of Power and System Justification | 74 Figure 2 Example of Block 1 trial
The Feeling of Power and System Justification | 75 Figure 3 Instructions for Block 2 Figure 4 Example of Block 2 trial
The Feeling of Power and System Justification | 76 Figure 5 Instructions for Block 3
The Feeling of Power and System Justification | 77 Figure 6 Instructions for Block 4 Figure 7 Blocks 3 and 4. Example 1 Fig ure 8 Blocks 3 and 4. Example 2
The Feeling of Power and System Justification | 78 Figure 9 Instructions for Block 5 Figure 10 Example of Block 5 trial
The Feeling of Power and System Justification | 79 Figure 11 Instructions for Blocks 6 & 7
The Feeling of Power and System Justification | 80 Figure 12 Blocks 6 & 7. Example 1 Figure 13 Blocks 6 and 7. Example 2
The Feeling of Power and System Justification | 81 Appendix C The Gender and Authority Measure
The Feeling of Power and System Justification | 82 *Items require reverse scoring
The Feeling of Power and System Justification | 84 purposes, and referred to a page informing them that their data would not be saved and asking if they would like to try the experiment anyway. Those responding in the affirmative were subsequently shown a drop down menu from which they cou ld choose a priming condition to view, and were passed through the appropriate experimental sequence. Each PHP page, importantly, was filtered by an if' (isset') clause such that only data from the prescribed population would be passed forward through su bsequent experimental stages. One drawback of using this method was related to a fairly high, and unexpected attrition rate among participants. Since the preliminary MySQL database responsible for regulating prime and ordering conditions was updated with subsequent prime and order as soon as it was determined that participants met experimental criteria, those participants who dropped out prematurely before completing the IAT (implicit measure), and GAM (explicit self report measure) skewed data away f rom those priming and ordering conditions they were assigned. While the effect of this on ordering was fortunately minor, significantly smaller frequencies in a written priming condition did reflect a higher attrition rate for these, than for other partici pants. This trend was reflected in each participant pool analyzed. A potentially smarter approach would have been to pass data from those participants meeting experimental criteria through all stages of the experiment, and to update data in the regulative database only upon completion of all necessary experimental stages. Such a method, if potentially better, would still not ensure equivalent frequencies, as many data were collected almost simultaneously (esp. in early hours of experimental distribution). T hus those participants, who began the experiment during another experimental session, would all receive the same priming condition. Since
The Feeling of Power and System Justification | 85 a single experimental session was roughly 20 minutes long, and many data were gathered in close temporal proximity, th is would not have been ideal. A second MySQL database served the sole function of data storage. This database was updated for each participant at the very end of the experimental session, and by means of PHP code contained in: http://aufbees.com/experimen t/complete.php. Insertion of data took the form, $query = ("INSERT INTO data (sid, gender, score, ord, prime, s1s15) VALUES ( $_SESSION[sid] $_SESSION[gnd ]' ), $_SESSION[score ], $_SESSION[or] $_SESSION[prm ], $_SESSION[s1] $_SESSION[s15] )") ; mysql_query($query) or die(mysql_error()); in PHP ; where all variables within the first set of parameters indicate columns in the table: data', and all variables of the form, $_SESSION[x]', invoke values stored within cookies on the user's mach ine as PHP "session variables." Session variables are a useful tool provided within PHP, which allow data to pass easily from one page to the next, linking series of web pages through common user specific data. Though PHP sessions make use of user stored cookies (unless cookies are disabled, in which case sessions will default to passing data through HTML GET calls), each session cookie contains an expiration date, at which date it will automatically self destruct. Furthermore, existing session cookies are automatically overridden, such that two users who take the experiment in succession will receive the session information appropriate to them. The MySQL database used to store user data is represented in PHP at: http://aufbees.com/experiment/testmysql.php.
The Feeling of Power and System Justification | 87 w hich would result from variability in connection speeds. Applets, in fact, are relatively low weight, and easy to run on any java enabled system. Applets are easily hosted by means of HTML, which allows for the passing of parameter values. The IAT applet developed here, took exactly one parameter, which determined display order (whether the pairing: male'/ weak' or female'/ weak' was displayed first). Initially the applet was called in HTML with a simple invocation of the PHP session variable for disp lay order. This method however, proved difficult. Instead, a PHP clause was included which simply called the IAT with a parameter value of 0', if the session variable's value for order was equal to 0, and called it with 1', if the session variable was equal to 1. This produced slightly more code, but was effective. Applets take two forms: signed, and unsigned. While signed applets have user permission to access and interact with the client machine, and therefore pose a potentially large security threa t to users, unsigned applets have the practical drawback of operating within what is called a security "sandbox." It is accordingly easier to pass data from a signed applet, than it is an unsigned applet. Given the infamy of signed applets, and the large w arning concerning security risks that accompanies a request for applet signature, it seemed that using a signed applet for the experiment would only deter otherwise willing participants from completing the IAT. Initially, it was my intention to send IAT s cores directly from the applet to the MySQL database, due to the necessity of applet signatures, and uncertainty about this method expressed in various forums online, I resolved to find an alternative solution. The alternative that was settled on initial ly was an HTML POST internal to the java applet. The java applet was to truncate the current URL upon completion of the test,
The Feeling of Power and System Justification | 88 and open the subsequent survey with POST variable for IAT score. This score was to be subsequently passed as a PHP session variabl e, along with survey results, priming condition, order, and participant identification number. Unfortunately, HTML POST ing was ineffective for reasons that are still unclear. To achieve the same results, the GET method was used. Passing data through GET i s essentially a less subtle version of POST ing which involved appending the variable and its assigned value to the desired URL. While this method of transporting data from page to page, can be problematic for those who wish to keep values hidden, GET ing was for these purposes a simple, and elegant solution to the sometimes tricky issue of moving data from an unsigned applet to MySQL(by means of PHP). The Applet Itself Without question, the bulk of programming efforts in this thesis were devoted to the de velopment of the Implicit Association Test in applet form. Because the code itself is quite complex, the discussion here will be fairly high level. The applet itself included three classes. The first, and primary class was an extension of Java AWT's Appl et library whose primary function was to house the other two classes and initialize the applet for use. This applet class also served the basic function of sending data out to a new PHP page through calls to "getCodeBase(),"which returns the URL on which t he applet is being hosted, and "getAppletContext().showdocument( x )", which truncates the current host page and refers the browser to a new URL, x In this case, x was the URL returned from getCodeBase() with the appended string, "survey" + ?score=" + sc ore ", constituting the GET call discussed above.
The Feeling of Power and System Justification | 89 The first "Main" class, was related to a second "MainPanel" class via composition (vs. inheritance), meaning that both classes contained each other as instances. This allowed information from the latter t o be dynamically returned to the former, and vice versa. This was necessary because latencies were established, stored, and calculated via a fairly complex scoring algorithm, to yield a final score within the MainPanel class only. Since the getCodeBase(), and showdocument() methods could be called only from within an applet context, it was ultimately necessary to send the data from the Main class used to instantiate the MainPanel class. Thus new instances of MainPanel accepted not only the dynamically defi ned "order" as a variable, but also an instance of the class Applet. Instantiation of MainPanel within the class, Main, was as follows (in public void init()): int or = Integer.parseInt(getParameter("ORDER")); MainPanel pane = new Ma inPanel(or, this); A third class defined, was the "Experiment" class, perhaps a misnomer. This class, called by the Main class, was responsible primarily for interfacing with .txt files within the .jar, which contained the instructions to be displayed at the beginning of each testing block. The Experiment class was initially to be designated with the task of running the scoring algorithm, and interfacing between Main, and MainPanel, as the code was developed, however, this class assumed an increasingly mar ginal role. A cleaner code would surely exclude this class. The MainPanel' Class As indicated, much of the logic within the IAT was contained within the MainPanel' class, which was an extension of the AWT class, "Panel." While both Main and MainPanel classes initially used Java Swing as their basis, certain unforeseen
The Feeling of Power and System Justification | 90 interactions with UNIX based operating systems, prevented the Swing code from focusing. A handful of participants, mostly running Mac OSX, thus complained of seeing the initial instructi on slide, but being unable to progress beyond that point. This issue was resolved easily enough however, with the sole compromise that the IAT now displayed a slight flicker between trials. Among the earliest processes and methods implemented within the MainPanel class upon its instantiation are those devoted to (a). assembling stages, or "blocks" in accordance with the order' parameter discussed above; (b). populating a distinct list of target words at random for each trial; and (c). setting parameter v alues for the display methods, which assumed most of the conditionality and logic within the experiment. While assembly of stages was fairly simple, the population of target words for each stage involved some measure of complexity, as it involved running through the array of target words appropriate to a given block in sequence, and assigning each to a random position in a variable length array (length determined by the number of trials in each respective block). It was also necessary to ensure that targe t words were not assigned to positions containing the same word either immediately before or after that trial. Another complicating factor was that the array of target words had to be accompanied by an associated array establishing whether target words wer e correctly paired with categories on the right or left, and another array determining whether target words belonged to categories on the top (male/female) or bottom(weak/strong) word spectrum. This latter array, it is worth noting, was responsible for det ermining whether target words were to be displayed in white or green font. Methods related to the display of the IAT, included: (a). a regulative display
The Feeling of Power and System Justification | 91 method called initially and with each subsequent key event, (b). a default display method for all act ual trials, (c). a "displayX" method, identical to the default display except for the presence of a large red x in the center of the screen (this was to be called when a user pressed the wrong button), (d). a display method showing instructions, and (e). a n end display, indicating that the experiment was complete and calling on the scoring algorithm to calculate a final score. The role of the regulative display method was merely to decide based on current parameter values, which display method to call. P arameters were adjusted with each key event according to the current state, whether or not participants answered correctly, and the number of trials in the present block. During experimental trials (blocks 3,4,5, and 7), each key event also played the role of stopping and storing the millisecond value of a timer initialized inside the regulative display method, as well as checking the key to determine whether answers were correct against a two dimensional Boolean array. A two dimensional array held these re sponse latency values, while an associated array of Boolean values stored whether or not the initial response value was correct for each trial. The display and display x methods were identical, as explained above, and merely served to paint word strings in the appropriate locations and color on the screen. The method responsible for displaying instructions, however, was notable in that it effectively employed a miniature, and quite primitive word processing program. Instructions were displayed by measurin g the height and width of the applet and relating it to the current location. Individual words were painted variable distances from previous words depending on word length, to create the appearance of equal sized spaces. Because some words in the instructi ons demanded greater attention (i.e. "green", "white", "strong",
The Feeling of Power and System Justification | 92 "weak"), these were highlighted in green and white. Scoring Algorithm The scoring algorithm utilized here was developed in accordance with Grenwald, et al.(2003)'s research on the IAT and d evelopment of an improved scoring algorithm. Data from blocks 3, and 4, was compared with that of blocks 5 and 7. Participants with response latencies greater than 10 seconds were rejected, and assigned a static score of 6.0 (scores typically ranged from 0.5 to +1.2 ), as were those for whom more than 10% of trial latencies were less than 300ms. The mean of correct latencies were computed for each block, and each latency on a trial that was answered incorrectly was replaced with that block's mean (for correct latencies) + 600ms. Pooled standard deviations were also computed for blocks 3 and 6, an d blocks 4 and 7. Discrepancies between mean latencies for blocks 3 and 6, and blocks four and 7 were computed and divided by their associated pooled standard deviations. The final score was realized by averaging these two quotients.
The Feeling of Power and System Justification | ii