gender inequality

Interaction and the Conservation of Gender Inequality: Considering Employment
Author(s): Cecilia L. Ridgeway
Source: American Sociological Review, Vol. 62, No. 2 (Apr., 1997), pp. 218-235
Published by: American Sociological Association
Stable URL: http://www.jstor.org/stable/2657301 .
Accessed: 28/09/2011 23:22
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

American Sociological Association is collaborating with JSTOR to digitize, preserve and extend access to
American Sociological Review.

http://www.jstor.org

INTERACTION AND THE CONSERVATIONOF
GENDER INEQUALITY:
CONSIDERING EMPLOYMENT*
Cecilia L. Ridgeway
Stanford University
How can we explain the persistence of gender hierarchy over transformnations in its socioeconomic base? Part of the answer lies in the mediation of gender inequality by taken-for-granted interactional processes that rewrite inequality into new institutional arrangements. The problems of interacting cause actors to automatically sex-categorize others and, thus, to cue gender stereotypes that have various effects on interactional outcomes, usually by modifying the performance of other, more salient identities. Because changes in the status dimension of gender stereotypes lag behind changes in resource inequalities, interactional status processes can reestablish gender inequalities in new structural forms. Interactional sex categorization also biases the choice of comparison others, causing men and women to judge differently the rewards available to them. Operating in workplace relations, these processes conserve inequality by driving the gender-labeling of jobs, constructing people as gender-interested actors, contributing to employers ' discriminatory preferences, and mediating men 's and women 's perceptions of alternatives and their willingness to settle for given job outcomes.

of ow can we explainthe persistence

Although there is no single answer, part of the solution may lie in the way gender stratification is mediated by interactional processes that are largely taken for granted. In this paper I argue that interactional gender mechanisms can operate as an "invisible hand" that rewrites gender inequality into new socioeconomic arrangementsas they replace the prior socioeconomic bases for gender hierarchy.
I focus on interactional mechanisms that mediate gender inequality in paid employment. Employment is one of two interdependent structural foundations on which our present system of gender hierarchy appears to rest; the other is the household division of labor. Some efforts have been made to understand the interactional mediation of the
*Direct communication to Cecilia Ridgeway, latter (Berk 1985; Risman 1987), but few for
Department of Sociology, Stanford University, the former.
Stanford, CA 94305-2047 (ridgeway@leland.
A substantial research industry has sought stanford.edu). I thank Karen Hegtvedt, Edward the persistence of wage inequality
Lawler, Judith Lorber, and Dafna Izraeli for com- to explain and sex segregatedjobs. Key processes idenmenting on earlier versions of this paper, as well as several helpful ASR reviewers. I also acknowl- tified include statistical discrimination, interedge the support of National Science Foundation nal labor markets, and the rendering of labor
(Grant SES 9210171). queues into gender queues, but explanations
H

gender hierarchy in our society over major historical transformationsin its socioeconomic base? A system that advantages men over women in material resources, power, status, and authority (i.e., gender hierarchy)has continued in one form or another despite profound structuralchanges such as industrialization and the movement of production out of the household, women 's accelerated movement into the labor force after
World War II, and, most recently, women 's entry into male-dominated occupations
(Hartmann 1976; Reskin and Roos 1990).
What accounts for the chameleon-like ability of gender hierarchy to reassert itself in new forms when its old structuralforms erode?

218

American Sociological Review, 1997, Vol. 62 (April:218-235)

THE CONSERVATION OF GENDER INEQUALITY

remain incomplete (England 1992; Reskin and Roos 1990). An analysis of mediating interactional mechanisms may improve our answers to several stubbornquestions including the reasons for unrelenting gender-labeling of jobs despite occupational change, how employers ' apparent preferences for male workers persist even under competitive market pressures, why women 's work is devalued, whether and how people act in their gender interests in employment matters, and why women workers accept lower wages than equivalent men.
Like race or class, gender is a multilevel system of differences and disadvantages that includes socioeconomic arrangements and widely held cultural beliefs at the macro level, ways of behaving in relation to others at the interactional level, and acquired traits and identities at the individual level. Interactional processes contributeto all forms of inequality, but there are several reasons for suspecting that they are especially important in gender inequality. First, our system of sex categorization divides the population into two groups of roughly equal size, creating the maximum structurallikelihood of a high rate of interaction between men and women
(Blau and Schwarz 1984). Sex categorization crosscuts almost all other divisions in the population, including kin and households, and forces regular cross-sex interaction on virtually everyone. In addition, there is growing evidence that our cultural system of gender difference relies heavily on interaction.
What Deaux and Kite (1987) call the "now you see them, now you don 't" nature of sex differences in behavior suggests that they are situationally and thus interactionally based, as many gender theorists now argue (Deaux and Major 1987; Eagly 1987; West and Zimmerman 1987).
I argue that gender becomes an important component of interactional processes because the problems of organizing interaction evoke cultural schemas that reinforce continual sex categorization. Sex categorization is the process by which actors classify one another as male or female, supposedly on the basis of physical sex criteria, but more commonly on the basis of personal presentation
(e.g., clothing, hairstyles) that the audience presumes stands for these sex criteria (West and Zimmerman 1987). As ethnometho-

219

dologists have demonstrated, this process is almost entirely socially constructed despite its apparent "naturalness" (Goffman 1977;
Kessler and McKenna 1978). Sex categorization in interaction, in turn, can activate a number of gender processes that may recreate gender hierarchy in the organizational and resource-distributingprocesses that the interaction mediates. I focus on two of these processes-status processes and biased referential processes-that are especially relevant for employment inequality. After describing the interactional gender mechanisms, I discuss the role they play in mediating the persistence of gender inequality in employment. INTERACTION, GENDER, AND
INEQUALITY
Interaction and Sex Categorization
It is striking that people are nearly incapable of interacting with one another when they cannot guess the other 's sex. The television program Saturday Night Live evoked this

situation in its comedy sequence about "Pat," an androgynous person who wreaked interactional havoc even in the most mundane encounters because others couldn 't place her/ him as a woman or man. Although people usually can interact with others whom they can 't place on other major dimensions of inequality, such as class or race, they seem to have difficulty completing even trivial, routine exchanges with someone they can 't classify by sex. This suggests that sex categorization is deeply rooted in the cultural rules that organize interaction (West and Zimmerman 1987).
Interaction requires coordinating your behavior with that of another. To act yourself, you need some way of making sense of and anticipating the other 's behavior. As symbolic interactionists have long noted, this requires that you develop at least a minimal definition, some initial beginning of "who" you and the other are in this situation (Alexander and Wiley 1981; Stryker 1980). Something can be "seen" only in explicit or implicit contrast to something else; therefore defining self and other requires one to find dimensions by which to categorize the other as similar to or different from self in various

220

ways, as social identity research has demonstrated (Turner 1987).
The process of situating self and other through categorization is a nested process that must begin so that interaction can start but it continues throughout the interaction episode, as documented in recent models of person perception (Brewer 1988; Fiske and
Neuberg 1990). Over time, more and more crosscutting classifications are introduced, yielding increasingly complex and nuanced situated identities for self and other. The relevance to action of any given classification waxes and wanes with events, but at least one initial categorization of self and other as similar or different on some dimension is necessary if any interaction is to take place.
This argumentimplies that sets of interacting individuals are likely to actively construct shared cultural schemas for readily categorizing self and other. Some of these schemas must be so simplified and so apparently obvious that they provide an easy means of initially situating self and almost any other so that interactionmay begin at all.
Such prior categorization systems in effect are cultural "superschemas"defining a few fundamentalcategories that can be applied to make sense of any person. They need not be relevant to the specific focus of interaction.
They merely renderactors sufficiently meaningful to one another to be able to address each other in relation to the focal goals and, by doing so, to introduce more relevant categorizations. Yet, by providing a cognitive starting point from which the rest proceeds, these superschemas can subtly influence the course of interaction even when they are irrelevant to its focus.
Supporting this analysis, research in cognition demonstrates that person perception is hierarchical: It begins with an initial, automatic classification according to a very small number of primary social categories and moves on to more detailed typing depending on the circumstances (Brewer 1988; Fiske and Neuberg 1990). Empirical evidence demonstrates that sex functions as one of these primarycategorizationsystems in Western society (Fiske 1992). Studies show that we automatically and unconsciously sex-categorize any specific other to whom we must relate (Brewer and Lui 1989; Stangor et al.
1992).

AMERICAN SOCIOLOGICAL REVIEW

As a dimension of variation among individuals, sex may be especially susceptible to social construction as one of a culture 's primary systems of self-other categorization.
Brewer and Lui (1989) argue that although cultures vary in the specific dimensions of human variation that serve as their few primary person categories, sex is always among them. Once sex is constructed as a simple, roughly dichotomous distinction, its constant use in interaction keeps it always accessible in people 's minds (Bargh 1989) and discourages its differentiation into more than two sexes, which would reduce its usefulness as a quick, prior way of classifying self and other.
The social problems of organizing interaction over a wide range of actors and circumstances may facilitate the cultural construction of sex as a simple, prior categorization system. Once this occurs, however, sex categorization becomes a habitual, automatic partof person perception. In institutional settings, including workplaces, clear social scripts may define self and other (e.g., supervisor and worker). Yet sex categorization continues because the actual process of enacting an institutional script with a concrete other evokes habitual person perception, and with it, the culture 's superschemas that define the basic attributes necessary to make sense of any person. Cognition research shows that when institutional identities and occupational roles are activated in the process of perceiving a specific person, they become nested within the prior, automatic categorization of that person as male or female and take on slightly different meanings as a result (Brewer 1988; Brewer and Lui 1989).
Thus, although we may be able to imagine an ungendered institutional script whereby
"the student talks to the teacher,"we cannot interact with any actual student except as a male or female student. The sex categorization of self and others, even in institutionally scripted settings, is a fundamental process that injects a variety of gender effects into the activities and institutional contexts that people enact.
Gender Stereotypes, Salience, and Behavior
If the cultural construction of sex as a simplified, prior categorization system is related

THE CONSERVATION OF GENDER INEQUALITY

to its uses in interaction, then the cultural development of gender stereotypes is likely; these describe what behaviors can be expected from a person of a given category.
Given the basis of automatic sex categorization in interactional contrasts, it is likely that whatever specific content is attached to a sex category, it will be organized around polarized traits that differentiatemen from women
(Deaux and Kite 1987).1
Actors ' gender stereotypes are cued by sex categorization, which makes them implicitly accessible (Fiske and Neuberg 1990). But the extent to which these stereotypes shape actors ' behaviors in the setting (e.g., their performance scripted roles) depends on the salience of gender in the situation compared with other identities on which they have also categorized self and others (Berger et al.
1977; Deaux and Major 1987; Eagly 1987;
Fiske and Taylor 1991). Although sex categorization provides an all-purpose way to begin, its very generality as a social category usually necessitates subsequent, more specific categorizations (Brewer 1991; Turner
1987). As multiple categorizations occur, the cognitive implications of each, weighted by its relevance to the situation and its utility for making sense of the other, are combined by actors into an ongoing impression (Fiske and
Neuberg 1990). Therefore stereotypes cued by sex categorization can vary from vague cognitive backgrounds, whose implications for behavior are virtually overcome by more immediate identities in the situation, to powerful determinantsof actors ' expectations and behavior. In work settings institutional identities are likely to reside in the foreground for actors.
Evidence indicates, however, that even when other identities are the most powerful determinants of behavior in a situation, cultural gender stereotypes become effectively salient
(i.e., sufficiently salient to measurably modify actors ' expectations and behavior) under at least two conditions: when the interactantsdiffer in sex category, and when gender is relevant to the purposes or the social context of the interaction (Berger et al.

221

1977; Cota and Dion 1986; Deaux and Major 1987). Indeed, gender may shape behavior most commonly as an effectively salient backgroundidentity that acts in combination with more situationally salient foreground identities and modifies their performance.
Even when initially they are not effectively salient, gender stereotypes are primed by actors ' sex categorizationof one another so that they are easily triggered, or made salient, by events in interaction(Bargh 1989; Deaux and
Major 1987). This is especially likely because of the diffuse nature of gender stereotypes, which allows them to be construed as relevantto many situations. For these reasons and because of the high rate of mixed-sex interaction, the conditions in which gender stereotypes become salient enough to perceptively modify behavior and judgments are a large subset of all situations.
Gender Status and Behavior

By continually reinforcing sex/gender as a system of presumed difference, interaction creates a salient distinction that can easily become a basis for inequality. Gender status beliefs are one form of inequality: These are
-widely held cultural beliefs that evaluate one sex as generally superior and diffusely more competent than the other.When status beliefs form, they become an important component of gender stereotypes that is also effectively salient (affecting expectations and behavior) in mixed-sex and gender-relevant situations
(Carli 1991; Ridgeway 1993). It is well documented that currently accepted gender stereotypes incorporate assumptions of men 's greater status value; that is, men 's traits are generally viewed as more valuable than women 's, and men are diffusely judged as more competent (Broverman et al. 1972;
Deaux and Kite 1987; Eagly 1987).
Although other elements of gender stereotypes probably are also important, I focus here on status beliefs because they are directly relevant to inequality. Gender status beliefs have three types of effects on goaloriented interaction that affect employment inequality. First, when effectively salient,
I While sex category remainsdichotomous, they cause both men and women to implicare genderstereotypes morecomplex,containing itly expect (or expect that others will expect) womanor greater competence from men than from multiplesubtypessuchas professional women, all other things being equal. These woman. traditional

222

expectations tend to become self-fulfilling, shaping men 's and women 's assertiveness and confidence, their judgments of each other 's competence, their actual performance, and their influence in the situation
(Carli 1991; Miller and Turnbull 1986; Pugh and Wahrman1983; Ridgeway 1993).
Second, activated gender status beliefs create expectations for rewards that reflect an actor 's relative status and expected performance and thus favor men over equivalent women (Berger, Fisek, Norman, and Wagner
1985). These reward expectations often acquire the normative, moral quality of a
"right" to rewards corresponding to one 's status relative to others who are different in status-relevant ways (Berger et al. 1985;
Cook 1975). When gender status is effectively salient, men may react negatively if they are placed on the same rewardlevel as a similar woman and may experience this situation as an implicit status threat.
Third, because gender status beliefs advantage men over women who are otherwise their equals, men, on average, have less interest in attending to information that undermines expectations based on gender status.
Cognition research suggests that people are
"good enough" perceivers; the extent to which they move beyond initial categorizations, incorporate inconsistent information, and develop complex, individuated impressions of the other is mediated by their motives in the situation (Fiske 1992; Fiske and
Neuberg 1990). In interaction, men are less likely to notice, and more likely to discount if they do notice, information about self or other that might diminish or eliminate the effects of gender status beliefs on expectations for competence and rewards.2 As a result, women may find it difficult to alter the lower expectations held for them.

AMERICAN SOCIOLOGICAL REVIEW

source (e.g., access to material resources or coercive power), it drives the social construction of gender status beliefs (Ridgeway
1991). Most important, it does so in a manner that helps maintain these beliefs in spite of changes in the structural conditions that support them.
When people who differ in resources engage in goal-oriented interaction, they usually develop hierarchies of influence and respect in the situation that correspond to their resource differences (Harrod 1980; Stewart and Moore 1992). Experimental evidence shows that when this happens, and when the actors also differ on a distinguishing personal attribute,they form the belief that people in the resource- and influence-advantaged category of the attribute are more highly respected, more competent, and more powerful in most people 's eyes than are individuals in the disadvantaged category (Ridgeway et al.
1995). In other words, they form status beliefs about the distinguishing attribute.
This point suggests that because interactional categorization makes sex a salient distinction in mixed-sex encounters, goal-oriented encounters between men and women who differ in resources should foster gender status beliefs. With a gender inequality in the distribution of resources there will be more of these mixed-sex encounters in which the man is resource-advantaged; thus the encounters will produce a predominance of status beliefs favoring men, which diffuse widely (Ridgeway and Balkwell 1997).
If interactional processes, when they occur, are sufficient (if not necessary) to create gender status beliefs in the context of gender inequalities in the distributionof a valued resource, then interaction will ensure the continuance of such beliefs as long as some such distributionalinequalities exist. Interactionis also likely to conserve gender status beliefs over changes in the original distributionalinInteraction and Gender Status Beliefs equalities that supported them. Because staContinual sex categorization in interaction tus beliefs create expectations that have selfhas an especially potent consequence: Under fulfilling effects, they resist change and canconditions of distributional inequality be- not be eroded except by repeated disconfirmtween the sexes in some valued asset or re- ing experiences (Harris and Rosenthal 1985;
Miller and Turnbull1986; Rothbartand John
2 Some women also may have an interestin 1985). Multiple experiences are required,eswhich pecially for people who benefit from gender traditional maintaining genderstereotypes, makesthem similarlyresistantto disconfirming status beliefs because their self interest makes them more cognitively resistant to information. THE CONSERVATION OF GENDER INEQUALITY

disconfirming information. When structural changes (e.g., economic, technological, or widespread social organizational changes) cause a decline in the original distributional inequalities, people will have more frequent disconfirming interactionalexperiences-for example, interactions in which a woman has resource advantages large enough to override gender status so that she becomes men 's actual superior in situational power and prestige (Pugh and Wahrman 1983). Yet, unless structural change produces a rapid outright reversal in the inequalities (i.e., from favoring most men to favoring most women), the rate at which change produces such reversal interactions may not provide enough people with enough disconfirmations to permanently erode their status beliefs except over a long period. As a result, change in the evaluative content and consensuality of gender status beliefs across the population will be slow and will lag substantially behind the changes in the distributionalinequalities that support them.
The lagged effect of gender status beliefs creates a "window" of time during which, even as societal changes mitigate the former distributional inequality, the continued operation of gender status in interaction biases the interactionally mediated allocation of other resources, opportunities, or positions of power. As a result, men will retain their advantage in power and resources within newly emerging organizational forms, although their degree of advantage may change. Sex Categorization, Interaction, and
Comparison Others
Interactionally driven sex categorization activates a second process that also is important for gender inequality in many situations, but particularly in employment. The categorization of self and other establishes a referential set of those who are similar to oneself and are therefore appropriate comparison others for evaluating one 's rewards or other outcomes in a situation (Suls and Wills
1991). As Major (1989) observes, when valued rewards are distributedunequally among men and women in a population, the biased selection of comparison others can result in sex differences in the levels of rewards that

223

people feel they are entitled to receive in a given situation.
Research shows that people define the level of rewards they are entitled to receive in a work situation (e.g., pay, promotions, working conditions) in comparison with others who are similar to them in attributesrelevant to the situation (Major 1989; Major and Forcey 1985). Information on others ' outcomes is acquired primarily by searches
(Major 1989). Searches involve talking to others-to coworkers, friends, family, and associates-asking around and evaluating written or observed evidence of others ' outcomes. Basically they are interactive, involving a definition of self in relation to a concrete other, and thus searches evoke sex categorization. In turn, however, sex categorization during searches creates a dimension of implicit similarity that biases the search toward same-sex others. Because of the often unconscious bias introduced by sex categorization, people seek out more same-sex than other-sex comparison others and weigh more heavily the evidence of same-sex others with similar job qualifications in establishing the standard of rewards to which they feel entitled in a given situation (Crosby 1982; Major 1989; Moore 1991).
Sex-biased searches for comparison others both encourage and are facilitated by people 's tendency to form sex-homophilous social networks (McPherson and SmithLovin 1987). They are also encouraged by sex segregation in employment and other social contexts. The strength of the bias in a given actor 's search will depend on the availability of proximate same-sex others and on the assumed relevance, to the searcher, of sex category for reward outcomes. Even small biases can result in systematic differences in the comparative reward information acquired by men and by women if rewards are distributed differentially by sex in the population. Evidence in fact shows that women have lower pay expectations than similarly qualified men and that a major determinant of the discrepancy is the difference between the sexes ' estimates of what others earn (Major and Konar
1984; Major and Testa 1989).
Information from comparison others is useful in defining two types of referential standardsfor rewardoutcomes in a job. First,

224

what is the going rate for "people like me" with the same training, skills, and experience? Second, what range of outcomes do people in jobs like this one receive? Each question implies a search of slightly different others, but both searches are likely to show bias toward same-sex others, particularly the "people like me" search, which selects on attributesof the individual.
Referential standards for both "people like me" and "people in jobs like this" are beliefs about what is typical. From these beliefs people form expectations about the rewards to which they are entitled; these expectations in turn affect their willingness to settle for a given reward in a job or to press for more (see Major, Vanderslice, and
McFarlin 1984). Thus expectations for rewards, like performance expectations, tend to become self-fulfilling.
In sum, interaction makes gender a stubbornly available, if often implicit, distinction in the workplace and elsewhere by pushing actors to continually sex-categorize one another. Continual sex categorization, in turn, encourages the formation and use of gender status beliefs, and biases the choice of comparison others toward the same sex.
The task now is to consider how these processes help sustain gender inequality in employment.
GENDER AND OCCUPATIONAL
INEQUALITY
In the 1950s and 1960s, gender inequality in the United States seemed to rest heavily on two aspects of women 's relationship to paid employment: their lower rate of participation in the labor force and the concentration of employed women in a few low-paying, female-labeled jobs (Oppenheimer 1970).
Since that time, the first has changed profoundly as women have flooded into the labor force, and the second has changed considerably as they have moved into several formerly "male" occupations (Reskin and
Hartmann 1986:4; Reskin and Roos 1990:
17-18). Yet a significant degree of wage inequality and sex segregation in occupations and jobs has persisted in the face of these profound changes; this reality suggests that other processes are slowing their impact and conserving gender inequality.

AMERICAN SOCIOLOGICAL REVIEW

A major research industry has attempted to explain continuing wage inequality and job segregation. Differences in male and female workers ' work experience (i.e., human capital differences) explain only one-quarter to one-half of the sex gap in pay and account for little of the job segregation (England 1984, 1992; Kilbourne et al. 1994). If the problem is a "taste" for discrimination on the part of employers, competitive market conditions should wipe these out, as economists observe (Becker 1957). England
(1992) argues that employment inequality persists despite the flattening effects of the market because it is continually being created anew, even if it is worn down slightly over time.
What mechanisms continually recreate gender inequality in paid employment? England (1992; England and Browne 1992) points to the household division of labor and socialized internal constraints, as well as to employers ' prejudice, which devalues women and the activities associated with them.
Reskin and Roos (1990) argue that labor queues become gender queues because employers rank males as more valuable workers than females. Strober (1984; Strober and
Arnold 1987) points to a cultural system of patriarchyin which employers give men the first pick of the best jobs. Jacobs (1989) shows that socialized tracking affects the sex typing of initial jobs but that there are surprising rates of subsequentmobility; this suggests a "revolving door" process by which occupations remain predominantly the territory of one sex despite a great deal of individual movement. Jacobs argues that the culprit is a diffuse system of gender social control involving socialization and employment practices in the workplace (Jacobs 1989;
Jacobs and Steinberg 1995).
Thus, researchers maintain that gender arrangementsin employment result from structural and economic factors (e.g., the supply of certain types of workers, the growth or decline of certain occupations), on the one hand, and some type of gender status effect, on the other. Two sorts of institutional processes are importantin this regard. First, organizational structures, such as job ladders and internal labor markets, and institutionalized practices, such as job evaluation systems, incorporate assumptions about gender

THE CONSERVATION OF GENDER INEQUALITY

status at their inception and then persist through bureaucratic inertia (Baron, Jennings, and Dobbin 1988; Kim 1989; Reskin and Roos 1990; Steinberg 1995). Second, bureaucraticpolitics within employing organizations help to maintain inequality because actors in advantaged positions, often men, represent their own interests in salary-setting and job-evaluation processes more strongly than the interests of those in disadvantaged positions (Bridges and Nelson 1989). As
Stone (1995) comments, however, current theories and research "explain how gender works ratherthan why gender is such a major force in the organization of work" (p. 415).
A systematic incorporationof interactionally driven sex categorization,status, and referential reward processes can begin to explain why the work process is so relentlessly gendered and why this gendering persists in spite of ongoing economic and organizational change. Perhaps this perspective 's most distinctive contribution is its ability to answer the "why" question-although it can further specify answers to the "how" questions as well.
Most work-related interaction takes place in organizational contexts with established job structures and institutionalized practices that heavily constrain what occurs. Under business-as-usual conditions, interactional gender status and referential processes are part of the means by which existing genderbiased job structures and practices are enacted, reinforced, and maintained. Interactional gender processes, however, become important in themselves, rather than merely agents of higher-level structuresand rules, at the interstices of organizational structures and under conditions that force change on organizational structures and practices. In these transition zones where organizational structures are less clearly defined, sex-categorization, status, and referential processes play a part in shaping the interactions through which actors create new organizational rules and structural forms, and map gender hierarchy into them as they do so.
These transition zones are precisely where bureaucratic politics also have great effect
(Bridges and Nelson 1989); but interactional gender processes can help explain how and why bureaucratic politics become gender politics. 225

INTERACTIONAL MECHANISMS AND
OCCUPATIONAL GENDER
INEQUALITY
It is useful, first, to recognize the extent to which occupational arrangements and wage outcomes are interactionally mediated.
Workers gain information about jobs and evaluate them through contact with others.
Employers hire workers through direct interaction (e.g., interviews) or indirect interaction (e.g., reviewing resumes, records, references). On the job, as Kanter (1977) pointed out, performance, evaluations, task assignments, and promotions are mediated in complex ways by interaction. All of these mediating interactions are potential sites where interactionalmechanisms may help map gender hierarchy into the occupational patterns that result.3
The Sex Labeling of Workersand Jobs
Reskin and Roos (1990) argue that gender inequality in employment is maintained through the transformation of labor queues into gender queues. The necessary first step in this transformationis the sex labeling of workers. This point seems so obvious and so naturalthat we generally do not bother to explain it. But why should sex be such a primary and salient descriptor of workers? The answer lies in the way interactionevokes primary person perception, infusing sex categorization into the hiring process as it mediates employers ' recruitment and placement of workers. Because interaction evokes sex categorization, employers can never interview or read the resume of a sex-neutral worker.
Similarly, workerscannot interact with a sexunclassified coworker, boss, or subordinate; thus they create the conditions for writing gender inequality into workplace relations.
This situation begins a process that also leads to the sex labeling of jobs. The difficulty of interactingwith workers without cat3 Studies show that actors also sex-categorize others in computer-mediated interaction (Nass and Steuer 1993; Quist and Wisely 1991). In electronic communication, gender stereotypes affect actors ' judgments of others, but because the sense of audience is diminished, they constrain actors ' own behavior less strongly than in face-to-face contexts (Sproull and Kiesler 1991).

226

egorizing them by sex primes workers and employers alike to infuse gender into the institutional scripts by which the job is enacted, comprehended, and representedto others, effectively constructing it as a "man 's" or a
"woman 's" job. Employers often begin the process by implicitly or explicitly seeking workers of a given sex on the basis of assumptions about labor costs, which themselves are suffused with gender status effects
(Milkman 1987; Strober 1984). As the hirees interact with each other, bosses, clients, or customers, automatic sex categorization of self and others causes the employees ' enactment of work activities to be perceived as implicitly nested within their prior identities as men or as women and tinge those activities with gender. When hiring creates a predominance of one sex in the job, the gendered connotation of individualjob enactments spreads to the shared institutional scripts that represent the job and its activities to actors and others. Because interactional sex categorization primes gender stereotypes to become effectively salient on the job, even in sex-segregated contexts, workersand employers may come to justify in gender stereotypic terms those sex-segregated job activities that originally seemed gender irrelevant (e.g., electronic assembly or selling securities). This reaction furtherconsolidates the sex labeling of the job in the eyes of its participantsand of those who deal with them, and in representations in the media.
As the scripts that represent the job come to be labeled male or female, in either a given organizational culture or a wider culture, the differential status value attached to the sexes and their stereotypic traits spreads to the job as well. Continually reinforced by sex categorization in workplace interaction, the spread of status value affects the performance and reward expectations associated with the job. Experiments show that a job or task, when labeled feminine, is viewed by both job evaluators and job incumbents as requiring less ability and effort and as worth less compensation than the identical job or task when labeled masculine (Major and
Forcey 1985; McArthur and Obrant 1986).
Other evidence shows that the gender composition of a job alone has a significant impact on wages (Baron and Newman 1990;
England 1992), as does the association of the

AMERICAN SOCIOLOGICAL REVIEW

job with stereotypically female tasks such as nurturance(Kilbourne et al. 1994).
Although this labeling process involves shared cultural constructions of a job, it is governed primarily by the situational constructions of workplace interaction; thus it reacts to changes in the context of interaction (e.g., the sex of workers, managers, or trainees). Over time or between organizations, a given activity or job can be relabeled very flexibly from one sex to the other. What does not change, however, is the tendency to apply sex labels; interaction injects sex categorization into the work process and brings in status evaluations as well.
Men and Womenas Interested Actors
Employment inequality is also preserved, some observers have concluded, by men acting to maintain their advantages over women
(Acker 1989; Bridges and Nelson 1989;
Reskin 1988; Stone 1995). How does this actually play out? Part of the answer is structural: The interests of those in more powerful positions in employment organizations are represented more forcefully than the interests of the less powerful, who are more likely to be women (Bridges and Nelson
1989). But writers suggest that more is involved.
Because sex is such an all-encompassing category and crosscuts other differences, it has always been difficult to explain how or in what sense either men or women act in the interests of their gender. This question may be clarified by understanding how sex categorization in interaction tinges work identities with gender stereotypes in various degrees, thus evoking status interests and biasing perceptions.
When gender status is effectively salient in workplace interactions, because of the sextyped or mixed-sex context, it creates a number of apparentlygender-interestedbehaviors on the part of men, whether as employers, workers, or customers. People (including male actors themselves) will tend to judge male actors as more competent and more worthy of reward than equivalent women, to miss or discount information in the situation that undermines gender stereotypes, and to perceive an implicit status threat in the equal rewardingof equivalent men and women.

THE CONSERVATION OF GENDER INEQUALITY

All of these effects, however, usually occur as a modification, a biasing, of behavior and judgments during the enactment of a more situationally salient occupational and institutional identity. Thus a man acting in his role as an electrical engineer or a union representativemay slightly bias his treatment of other men and of women, usually in an implicit way that he himself does not recognize. Only occasionally will gender be so salient in the situation that men will act selfconsciously as men to preserve their interest.
Yet, the repeated background activation of gender status over many workplace interactions, biasing behaviorin subtle or more substantial degrees, produces the effect of men acting in their gender interest, even when many men feel no special loyalty to their sex.
This behavior-biasing process weights the encounters through which bureaucraticpolitics are enacted; it brings an implicit gender dimension to the outcomes in addition to that produced by the differential power of male and female actors.
What about women as interested actors?
As indicated by the entrance of women into male occupations and managementpositions, women pursue their interests in employment settings despite barriers (Jacobs 1992;
Reskin and Roos 1990). Even so, they are handicapped by the lower power attached to their positions and by interactional gender mechanisms. Where gender status is effectively salient, it is in women 's interest to introduce added job-relevant information that undermines its effects on perceptions of the competence and reward-worthiness of self and others in the situation. Doing so is difficult, however, precisely because gender usually operates as a background identity in workplace interactions; the participantsdo not define it explicitly as part of "what is going on here."Its implicitness complicates the task of recognizing its effects and introducing countervailing information in the real time of interaction.
The process is difficult as well because men 's own status interests tend to make them more cognitively resistant to countervailing information. As a result, women periodically may sense that something prejudicial is happening to them, but they may be frustratedin their efforts to act effectively against it. They will

227

be vulnerable to "role encapsulation," whereby others define them in their work identities in implicitly gendered terms that limit their effectiveness as actors in their own interests (Kanter 1977). In their analysis of the Washington State pay system, for instance, Bridges and Nelson (1989:645) found that women employees were disadvantaged not only because they had fewer representatives in pay-setting processes, but also because the actors and groups that traditionally represented women (e.g., the Nurses Association) were viewed as "passive and ineffective" on pay issues. Interactional gender mechanisms contributeto the situational construction of women in the workplace as stereotypically more "passive and ineffective" than many men in pursuing their interests.
Preferences for Male Workers
Labor queues become gender queues not simply through the sex labeling of workers but also through employer preferences that rank male workers higher in the queue
(Reskin and Roos 1990). The persistence of such preferences is problematic, because many women 's wage rates are lower than men 's, even when their qualifications are similar. Again, part of the explanation may lie in the interactional mediation of workplace relations and the opportunity this mediation provides for the operation of gender status beliefs.4 At least four types of sex discrimination have been suggested to account for employers ' preferences: tastes, error, statistical discrimination, and group collusion
(England 1992:54-68). The first three can be understood as straightforwardresults of gender status processes. Such processes also create the conditions for collusion.
When an applicant pool contains at least some members of both sexes or when the job has been sex-labeled (as in most hiring situations), sex categorization of applicants acti4 The agentsof employers 'preferences not are a alwaysthe personsauthorizing hire but lowerlevel functionarieswho actually constructads, screen applicants,recommendplacements,and
These individuals evaluateperformances. engage with proin moredirectand indirectinteraction spectiveandactualemployees. 228

vates status beliefs as employers assess applications, interview candidates, and talk to others in the hiring process. If the employer holds these activated gender status beliefs explicitly as a matterof ideology, or believes them to be so held by other workers or customers, status beliefs can function as "tastes"
(outright preferences for not hiring a given group) in the hiring process.
For the large majority of jobs, especially
"good" jobs involving either stereotypically male or gender-neutral tasks, gender status beliefs create a preference for male workers.
At the same time, they create only a weak taste for female workers for jobs defined as involving stereotypically female tasks, such as nursery school teacher. Gender status beliefs contain both general assumptions that men are more competent than women and specific assumptions that men are superior at stereotypically male tasks, while women are better at stereotypically female tasks.
For the "female" tasks, however, people appear to combine the general assumptions of male superiority with specific assumptions of female superiority to form expectations that women will be somewhat, but not greatly, better than men at female tasks
(Ridgeway 1993).
Probably more common than explicit tastes are the discriminatory effects exerted by activated gender status beliefs through their impact on employers ' judgments of workers ' potential productivity.Performance expectations based on gender status cause a male worker to appear "better" than an equally qualified woman (see Lott 1985).
Furthermore, an equally competent performance by the two appears more indicative of skill and ability in the man than in the woman (Deaux and Emswiller 1974; Foschi,
Lai, and Siegerson 1994). On the surface, then, not gender but merit is involved, but worker 's sex is connected with merit by the interactional mediation of employers ' evaluations of workers, and by the way this mediation injects gender status into the process.
In this way interactionally activated status processes create errordiscrimination, whereby two workers who would perform equally are judged to be different and are paid accordingly.
The operation of gender status beliefs in the workplace can also create the basis for a

AMERICAN SOCIOLOGICAL REVIEW

type of statistical discrimination. Performance expectations created by employers ' status beliefs tend to become self-fulfilling, and this tendency often produces employer experiences with male and female workers that confirm such judgments. A competent performanceby a female worker appears less competent than by a male worker.Also, and more insidious, the pressure of low expectations actually can interferewith some women workers 'performance.Similarly, high expectations of others can improve the performances of some men workers (Harris and
Rosenthal 1985). Thus the effect of statusbased expectations on some men and some women can create "real" differences in the average performance and productivity for groups of similar male and female workers.
When interactional sex categorization makes gender salient in the hiring process, the employer 's experience of these average differences also becomes salient. The employer may react by preferringmale workers across the board, thus creating statistical discrimination. Statistical discrimination is especially important for gender inequality because it resists the flattening effects of market forces more strongly than do other types of discrimination (England 1992:61-8).
Gender status processes in interaction also consolidate the conditions for interest-based collusion, a fourth type of discrimination suggested by several writers (Hartmann
1976; Reskin 1988; Strober 1984). Both workers and employers have a wide variety of crosscutting interests. But few are enacted and reinforced in interaction so repeatedly
(albeit implicitly) as are gender status interests; therefore these interests are especially fertile ground for explicit collusion. To occur, collusion may need to be triggered by contingent events that threaten gender hierarchy in a work setting, such as organizational or technological change.
As studies of organizationalpractices show
(Steinberg 1995), interaction is not the only source of discriminatory processes or their persistence. But the knowledge that interactional processes are sufficient to create most of the observed forms of discrimination clarifies an important(if subtle and even insidious) means by which discrimination is continually recreated in the face of leveling market forces. Especially important here is

THE CONSERVATION OF GENDER INEQUALITY

the capacity of gender status beliefs to lag behindchanges in the distributionalinequalities that support them; they give interaction a chameleon-like ability to reestablish discrimination in new forms in the face of market forces and structuralchanges in the work process. Why Do Women WorkersAccept
Lower Wages?
Employers ' ability to attract and retain women workers for lower wages is also critical for maintaining occupational gender inequality. Why do women settle for less than similarly qualified men? Although some argue that women place less value on money, women step forward when good-paying jobs open up for them, even if these have been labeled men 's jobs (Jacobs 1989; Reskin and
Roos 1990).
At this juncture, too, interaction plays a role by shaping different senses of entitlement on the part of similarly qualified male and female workers. Although many women work in sex-segregated jobs, their performance is often evaluated through direct or indirect interaction with male supervisors, clients, or customers. Also, their work may be typed as a women 's job. In any of these situations, interactionally determined sex categorization will activate status beliefs, affecting women workers ' own performance and reward expectations as well as their employers ' and fellow workers ' expectations for them. Evidence suggests that women underestimate the quality of their performances in comparison with men, and thus are susceptible to argumentsthat they deserve less pay
(Deaux and Kite 1987; Lenny 1977).
Status beliefs create expectations for the relative rewards that male and female workers deserve. Referential standards for rewards, established (like evaluation) through interaction, anchor those relative expectations around a specific reward level. If sexbiased searches of comparison others cause women to estimate the going rate at lower levels than do similar men for given work by people with given qualifications, this effect is a second reason why women judge the compensation they deserve as less than men do. Sex-biased referential standards reduce women 's reward expectations even in work

229

situations where status beliefs are not effectively salient. They also reduce the estimates of deserved rewards among women who try to resist the pressures of status expectations by developing very high skill levels about which they can be confident (Major 1989).
If women workers inadvertentlyunderestimate the rewards to which they are entitled, employers can more easily force them to settle for lower wages (Major et al. 1984). If correspondingstatus- and sex-biased referential processes cause male workers to overestimate what they deserve, employers find it harder to force lower wages on them. This unequal and self-fulfilling entitlement process, which operates within a work organization, is bolstered further by workers ' comparisons of their rewards with those in other employment settings (Reskin and Roos
1990). Searches among comparison others in different firms and different jobs are also biased by the interactional sex categorization and sex labeling of jobs. Thus, here, too, women generally will compare themselves to lower paid others than will men.
As a result of these entitlement and comparison processes, which are activated by interactional sex categorization, women settle
-for lesser rewards than do similar men. Although most women find it unfair that men have higher wages, they are no more likely than men to be dissatisfied with their own rewards and job (Crobsy 1982). This inadvertent acceptance of lower compensation helps sustain the system of gender hierarchyin pay over time by moderating women 's resistance
(Major 1989). By this analysis, it is a product of interactional gender mechanisms.
Women 's Entrance into Male Occupations
Interactionalmediation can also help explain why the movement of women into male occupations sometimes results in feminization of the occupation or resegregation by specialty, which reduces the wage benefits to women and moderates the impact on wage inequality (Reskin and Roos 1990). Given employers ' general preferences for male workers, it is usually a structuralchange that opens male occupations to women workers.
When the demand for employees in a given occupation outstripsthe pool of qualified and interested male workers at the acceptable

230

price, women begin to be hired (Reskin and
Roos 1990; Strober and Arnold 1987). Yet aspects of this structuralchange are also mediated interactionally, just as employers ' preferences are, with significant gendering effects that maintain inequality.
Male workers ' preferences are a majordeterminantof the available pool of male workers for a job. These are formed through the same sex-biased referential processes that shape women workers ' sense of entitlement.
With some exceptions (Wright and Jacobs
1994), case studies suggest that male occupations commonly are opened to women when their pay and working conditions start to deteriorate, often because of technological or organizational change (Reskin and
Roos 1990). I suggest that when men in these jobs experience a decline in their work outcomes, this decline triggers a search of comparison others through whom they evaluate their situation. Male workers will start to leave if the search, which is biased by interactional sex categorization, yields a standard for what is available to "people (men) like me" that is higher than the declining outcomes currently available in the job. If the search does not yield such a standard, they will stay with the job and may resist efforts to bring in lower-paid female workers.
As the shortage of male workers brings women into the job, gender-based status interests become increasingly salient in the workplace; sometimes they create tensions that appear greatest at the balance point in the gender mix (Wharton and Baron 1987).
Activated gender status beliefs cause women 's presence to subtly devalue the status and reward-worthiness of the job in the eyes of both workers and employers. Male workers may react to the perceived threat to status and rewardsby hostility towardwomen in the job. Tensions from male coworkers increase the costs of the job for women, but for many women, given their sex-biased referential standards, the job still will be relatively attractive in both status and pay. Men 's sexbiased referential standards, on the other hand, suggest that the job is increasingly less attractivethan alternatives. The men 's flight from the job will accelerate, and even fewer males will apply for the openings.
As women become more numerous in the job, supervisors ' gender status beliefs and

AMERICAN SOCIOLOGICAL REVIEW

women workers ' lower sense of entitlement exert self-fulfilling effects on women 's reward outcomes, and these effects increasingly spread to the job itself. This situation facilitates employers ' introduction of more organizational and technological changes that reduce the status and reward outcomes of the job. Although this scenario is not inevitable, the likely result when it occurs is feminization of the job or resegregation by specialty, with the female jobs and specialties declining in rewards and status (Reskin and Roos 1990).
The point here is that this transition, which maintains gender hierarchy over a change in the structuralorganization of jobs, is mediated by interactional processes. The combination of interactionally activated status processes and biased referential standards, also a product of interaction, creates a complex mix of discrimination, status-based interest competition, differences in entitlement, and differential perceptions of alternatives. The result is a system of interdependent effects that are everywhere and yet nowhere because they develop through multiple workplace interactions, often in taken-for-granted ways.
Their aggregate result is structural:the preservation of wage inequality and the sex segregation of jobs.
An Empirical Prediction
If the interactional perspective presented here is to be more than illustrative, it should provide testable empirical predictions. I have argued that under business-as-usual conditions interactional gender mechanisms are part of the process by which gender-biased organizational structures and institutional practices are implemented. But under more organizationally chaotic conditions, such as those at organizational interstices or those produced by economic change, interactional mechanisms are sufficient in themselves to create gender inequality in pay and power among the participants and to generate sextyping of work; as a result, any new organizational structures or practices that emerge from actions under these conditions will themselves embody gender hierarchy.Thus a general prediction is: Wage inequality and sex labeling of work will be present even in employmentsettings where the usual organi-

THE CONSERVATION OF GENDER INEQUALITY

national structures and practices that produce them are relatively absent, such as internal labor markets and biased job evaluation systems.
Evidence exists to support this prediction.
Television writers are employed on shortterm contracts through an organizationally unstructured, interpersonally mediated process whereby a few successful writer-producers serve as "brokers" (Bielby and Bielby
1995). Gendering structuressuch as internal labor markets and job systems are lacking; in addition, humancapital effects are blunted because competence is difficult to judge from the products themselves. Yet despite the absence of these usual sources of wage inequality, Bielby and Bielby (1995:224) found that between 1982 and 1990 women television writers had a net earnings disadvantage of 22 to 25 percent less pay than men of similar age, experience, and work histories. Male writers were better known and were perceived as better risks than equally successful female writers (a classic status effect) and women writers tended to be typecast in gender-stereotypicalways as situation comedy writers ratherthan action writers.
The degree of gender inequality in an organizationally unstructuredoccupation such as television writing is as large as in bureaucratically organized work. Yet there are differences in the primary mechanisms sustaining the inequality (interpersonalprocesses or organizationalstructures)(Bielby and Bielby
1995).5 This point underscoresthe multilevel nature of the processes by which gender hierarchy is enacted. As these data show, the power of interactional gender processes is that they are sufficient to maintaingender hierarchy in employment in the absence of the usual gendered organizational structures.
The organizational circumstances of television writers are atypical, but other situations where interactional gender processes should have testable effects are not. These include start-upcompanies and newly form-

231

ing professions (e.g., personal financial planners) that draw people from diverse occupational backgrounds. Within organizations, they include interdepartmental and interagency teams charged with change or innovation. In each of these organizationally less highly structuredsettings, interactional gender mechanisms can be predicted to measurably shape the interpersonaland power politics from which new organizational structures and practices emerge.
CONCLUSION

Adding an interactional account to labor market and organizational accounts of employment inequality helps explain why gender is such a major force in the labor process. Hiring,job searches, placement, performance evaluation, task assignment, promotion, and dealing with customers, clients, bosses, coworkers, and subordinates all involve direct or indirect (e.g., the evaluation of resumes) interaction. Interacting with a concrete other evokes primary cultural rules for making sense of self and of other, pushing actors to sex categorize one another in each of these situations. Sex categorization pumps gender into the interactionally mediated work process by cueing gender stereotypes, including status beliefs, and by biasing the choice of comparison others. The process is insidious because gender is usually an implicit, background identity whose effective salience varies situationally, acting in combination with more salient work identities and modifying their performance. In bureaucraticallywell-ordered work contexts, interactional gender mechanisms become part of the process for enacting more formal structures that embody gender bias, such as job ladders and evaluation systems. Interactional processes contribute to the sex labeling of jobs, to the devaluation of women 's jobs, to forms of sex discrimination, to the construction of men as effectively gender-interested actors, to the control of women 's interests, to differences between men 's and
5 The increasing rationalization and explicitwomen 's rewardexpectations, and to the proness of procedures that attend bureaucraticorgacesses by which women 's entrance into male nization may make the implicit gender inequality of interpersonallyorganized work difficult to sus- occupations sometimes leads to feminization tain without justification, and thus may encour- or resegregation by specialty. In less bureauage the differentiation of job titles by gender cratically ordered settings, such as those at organizational interstices, in start-upcompa(Baron and Pfeffer 1994).

232

nlies, in newly forming professions, or in some types of work, interpersonalprocesses come to the fore and are sufficient in themselves to create gender inequality in wages and sex typing of work. As they do so, interactional processes conserve gender inequality over significant changes in the organization of work, writing it into new work structures and practices as they develop.
If this inequality is to be reduced, it is vital to understand the multilevel nature of gender processes and the role of interactional processes in maintaining gender inequality.
Structural changes such as the implementation of comparable worth policies, for instance, would change men 's and women 's referential rewardexpectations. Yet, changes in gender status beliefs lag behind changes in the distributional inequalities that support them; thus the degree of equality achieved is likely to be substantially undermined by interactional processes mediating the decision making through which comparable worth policies would be adopted and implemented.
Concerted interventionis requiredat both the structurallevel (e.g., comparable worth) and the interactional level, through policies such as affirmativeaction that change the interpersonal configuration of actors, and create more stereotype-disconfirming experiences for all. Insofar as commitment to affirmative action creates greater accountability among workplace decision makers, social cognition research suggests that it also will reduce the impact of stereotypes, like gender, on their judgments and evaluations (Fiske and
Neuberg 1990).
Gender inequality in employment is maintained not only by the work processes discussed here but by its interdependence with the household division of labor. One of the promises of an interactional approachis that it may clarify how this interdependence works. Accomplishing this goal will necessitate the incorporation of additional gender mechanisms that affect interaction in enduring, intimate relations. It is a promising project for the future.
Cecilia L. Ridgeway is Professor of Sociology at
Stanford University. Her research focuses on interactional status processes and the role these processes play in multilevel processes of stratification, particularly gender stratification. A current project addresses the social construction of

AMERICAN SOCIOLOGICAL REVIEW cultural status beliefs about attributes on which people differ, proposing a powerful role for interactional contexts in creating and spreading such beliefs. Experimental tests of these arguments are underway, along with computer simulations exploring their implications.
REFERENCES
Acker, Joan. 1989. Doing Comparable Worth.
Gender, Class, and Pay Equity. Philadelphia,
PA: Temple University Press.
Alexander, C. Norman and Mary Glenn Wiley.
1981. "Situated Activity and Identity Formation." Pp. 269-89 in Social Psychology: Sociological Perspectives, edited by M. Rosenberg and R. Turner. New York: Basic Books.
Bargh, John. 1989. "Conditional Automaticity:
Varieties of Automatic Influence in Social Perception and Cognition." Pp. 3-51 in Unintended Thought, edited by J. Uleman and J.
Bargh. New York: Guilford.
Baron, James and Andrew Newman. 1990. "For
What It 's Worth: Organizations, Occupations, and the Value of Work." American Sociological Review 55:155-75.
Baron, James and Jeffrey Pfeffer. 1994. "The Social Psychology of Organizations and Inequality." Social Psychology Quarterly 57:190-209.
Baron, James, P. Devereaux Jennings, and Frank
Dobbin. 1988. "Mission Control? The Development of Personnel Systems in U.S. Industry." American Sociological Review 53:497514.
Becker, Gary. 1957. The Economics of Discrimination. Chicago, IL: University of Chicago.
Berger, Joseph, M. Hamit Fisek, Robert Norman, and David Wagner. 1985. "The Formation of
Reward Expectations in Status Situations." Pp.
215-61 in Status, Rewards, and Influence, edited by J. Berger and M. Zelditch. San Francisco, CA: Jossey-Bass.
Berger, Joseph, Hamit Fisek, Robert Norman, and
Morris Zelditch. 1977. Status Characteristics and Social Interaction. New York: Elsevier.
Berk, Sarah Fenstermaker. 1985. The Gender
Factory: TheApportionmentof Workin American Households. New York: Plenum.
Bielby, William and Denise Bielby. 1995. "Cumulative versus Continuous Disadvantage in an
Unstructured Labor Market: Gender Differences in the Careersof Television Writers."Pp.
209-30 in Gender Inequality at Work, edited by J. Jacobs. Thousand Oaks, CA: Sage.
Blau, Peter and Joseph Schwartz. 1984. Crosscutting Social Circles: Testing a Macrostructural
Theory of Intergroup Relations. New York:
Academic.
Brewer, Marilynn. 1988. "A Dual Process Model of Impression Formation." Pp. 1-36 in Ad-

THE CONSERVATION OF GENDER INEQUALITY vances in Social Cognition, vol. 1, edited by T.
Srull and R. Wyer. Hillsdale, NJ: Earlbaum.
. 1991. "The Social Self: On Being the
Same and Different at the Same Time." Personality and Social Psychology Bulletin 17:
475-82.
Brewer, Marilynn and Layton Lui. 1989. "The
Primacy of Age and Sex in the Structure of
Person Categories." Social Cognition 7:26274.
Bridges, William and Robert Nelson. 1989. "Markets in Hierarchies:Organizational and Market
Influences on Gender Inequality in a State Pay
System." American Journal of Sociology 95:
616-58.
Broverman, Inge, Susan Vogel, Donald Broverman, Frank Clarkson, and Paul Rosenkrantz.
1972. "Sex-Role Stereotypes: A Reappraisal."
Journal of Social Issues 28:59-78.
Carli, Linda. 1991. "Gender, Status, and Influence." Pp. 89-113 in Advances in Group Processes, vol. 8, edited by E. Lawler, B. Markovsky, C. Ridgeway, and H. Walker. Greenwich, CT: JAI.
Cook, Karen. 1975. "Expectations, Evaluations, and Equity." American Sociological Review
40:372-88.
Cota, Albert and Kenneth Dion. 1986. "Salience of Gender and Sex Composition of Ad Hoc
Groups: An Experimental Test of Distinctiveness Theory." Journal of Personality and Social Psychology 50:770-76.
Crosby, Faye. 1982. Relative Deprivation and
Working Women. New York: Oxford University Press.
Deaux, Kay and Tim Emswiller. 1974. "Explanation of Successful Performance on Sex-Linked
Tasks: What Is Skill for the Male Is Luck for the Female." Journal of Personality and Social
Psychology 29:80-85.
Deaux, Kay and Mary Kite. 1987. "Thinking about Gender." Pp. 92-117 in Analyzing Gender: A Handbook of Social Science Research, edited by B. Hess and M. Ferree. Newbury
Park, CA: Sage.
Deaux, Kay and Brenda Major. 1987. "Putting
Gender into Context: An Interactive Model of
Gender-Related Behavior." Psychological Review 94:369-89.
Eagly, Alice. 1987. Sex Differences in Social Behavior: A Social-Role Interpretation.Hillsdale,
NJ: Erlbaum.
England, Paula. 1984. "Wage Appreciation and
Depreciation: A Test of Neoclassical Economic
Explanations of Occupational Sex Segregation." Social Forces 62:726-49.
. 1992. Comparable Worth: Theories and
Evidence. New York: Aldine.
England, Paula and Irene Browne. 1992. "Internalization and Constraint in Women 's Subor-

233

dination." Current Perspectives in Social
Theory 12:97-123.
Fiske, Susan. 1992. "Thinking Is for Doing."
Journal of Personality and Social Psychology
63:877-89.
Fiske, Susan and Steven Neuberg. 1990. "A Continuum of Impression Formation, from Category-Based to Individuating Processes: Influences of Information and Motivation on Attention and Interpretation."Pp. 1-73 in Advances in Experimental Social Psychology, edited by
M. Zanna. New York: Academic.
Fiske, Susan and Shelley Taylor. 1991. Social
Cognition. New York: McGraw-Hill.
Foschi, Martha, Larissa Lai, and Kirsten Sigerson. 1994. "Gender and Double Standards in the Assessment of Job Applicants." Social Psychology Quarterly 57:326-39.
Goffman, Erving. 1977. "The Arrangement between the Sexes." Theory and Society 4:30131.
Harris,Monica and Robert Rosenthal. 1985. "Mediation of InterpersonalExpectancy Effects: 31
Meta-Analyses." Psychological Bulletin 97:
363-86.
Harrod, Wendy. 1980. "Expectations from Unequal Rewards." Social Psychology Quarterly
43:126-30.
Hartmann,Heidi. 1976. "Capitalism, Patriarchy, and Job Segregation by Sex." Pp. 137-70 in
Womenand the Workplace, edited by M. Blaxall and B. Reagan. Chicago, IL: University of
Chicago.
Jacobs, Jerry. 1989. Revolving Doors: Sex Segregation and Women 's Careers. Stanford, CA:
Stanford University Press.
.1992. "Women 's Entry into Management:
Trends in Earnings, Authority, and Values among Salaried Managers."AdministrativeScience Quarterly 372:282-301.
Jacobs, Jerry and Ronnie Steinberg. 1995. "Further Evidence on Compensating Differentials and the Gender Gap in Wages." Pp. 93-124 in
Gender Inequality at Work,edited by J. Jacobs.
Thousand Oaks, CA: Sage.
Kanter, Rosabeth. 1977. Men and Women of the
Corporation. New York: Basic Books.
Kessler, Suzanne and Wendy McKenna. 1978.
Gender: An Ethnomethodological Approach.
New York: Wiley.
Kilbourne, Barbara, Paula England, George
Farkas, Kurt Beron, and Dorothea Weir. 1994.
"Returnsto Skill, Compensating Differentials, and Gender Bias: Effects of Occupational
Characteristicson the Wages of White Women and Men." American Journal of Sociology
100:689-719.
Kim, Marlene. 1989. "GenderBias in Compensation Structures: A Case Study of Its Historical
Basis and Persistence." Social Issues 45:39-50.

234

AMERICAN SOCIOLOGICAL REVIEW

Lenny, Ellen. 1977. "Women 's Self-Confidence the annual meeting of the American Sociologiin Achievement Settings." Psychological Bulcal Association, August 27, Cincinnati, OH.
Reskin, Barbara. 1988. "Bringing the Men Back letin 84: 1-13.
Lott, Bernice. 1985. "The Devaluation of Womin: Sex Differentiation and the Devaluation of en 's Competence." Journal of Social Issues
Women 's Work." Gender and Society 2:58-8 1.
41:43-60.
Reskin, Barbara and Heidi Hartmann. 1986.
Major, Brenda. 1989. "Gender Differences in
Women 's Work,Men 's Work. Sex Segregation
Comparisons and Entitlement: Implication for on the Job. Washington, DC: National AcadComparable Worth." Journal of Social Issues emy Press.
45:99-115.
Reskin, Barbara and Patricia Roos. 1990. Job
Major, Brenda and Blythe Forcey. 1985. "Social
Queues, Gender Queues: Explaining Women 's
Inroads into Male Occupations. Philadelphia,
Comparisons and Pay Evaluations: Preferences for Same-Sex and Same-Job Wage CompariPA: Temple University Press. sons." Journal of ExperimentalSocial Psychol- Ridgeway, Cecilia L. 1991. "The Social Construction of Status Value: Gender and Other ogy 21:393-405.
Nominal Characteristics." Social Forces 70:
Major, Brenda and Ellen Konar. 1984. "An Investigation of Sex Differences in Pay Expectations
367-86.
and Their Possible Causes." Academy of Man. 1993. "Gender, Status, and the Social agement Journal 27:777-92.
Psychology of Expectations." Pp. 175-98 in
Major, Brenda and Maria Testa. 1989. "Social
Theory on Gender/Feminism on Theory, edited
Comparison Processes and Judgments of Enby P. England. New York: Aldine. titlement and Satisfaction." Journal of Experi- Ridgeway, Cecilia L. and James Balkwell. 1997. mental Social Psychology 25:101-20.
"Groupsand the Diffusion of Status Value BeMajor, Brenda, Virginia Vanderslice, and Dean liefs." Social Psychology Quarterly 60:14-31.
McFarlin. 1984. "Effects of Pay Expected on Ridgeway, Cecilia, Elizabeth H. Boyle, Kathy
Pay Received: The ConfirmatoryNature of IniKuipers, and Dawn Robinson. 1995. "Interaction and the Construction of Status Value Betial Expectations." Journal of Applied Social
Psychology 14:399-412. liefs." Paper presented at the annual meeting of the American Sociological Association, August
McArthur, Leslie and Sarah Obrant. 1986. "Sex
Biases in Comparable Worth Analyses." Jour23, Washington, DC. nal of Applied Social Psychology 16:757-70.
Risman, Barbara. 1987. "Intimate Relationships from a MicrostructuralPerspective: Men Who
McPherson, Miller and Lynn Smith-Lovin. 1987.
Mother." Gender and Society 1:6-32.
"Homophily in Voluntary Organizations: Status Distance and the Consequences of Face-to- Rothbart, Myron and Oliver John. 1985. "Social
Face Groups." American Sociological Review
Categorization and Behavioral Episodes: A
52: 370-79.
Cognitive Analysis of the Effects of Intergroup
Milkman, Ruth. 1987. Gender at Work. Urbana,
Contact." Journal of Social Issues 41:81-104.
IL: University of Illinois Press.
Sproull, Lee and Sara Kiesler. 1991. Connections: New Ways of Workingin the Networked
Miller, Dale and W. Turnbull. 1986. Expectancies and Interpersonal Processes. Annual Review of
Organization. Cambridge, MA: MIT Press.
Psychology 37:233-56.
Stangor, Charles, Laure Lynch, Changming
Duan, and Beth Glass. 1992. "Categorization
Moore, Dahlia. 1991. "Entitlement and Justice of Individuals on the Basis of Multiple Social
Evaluation: Who Should Get More and Why?"
Features." Journal of Personality and Social
Social Psychology Quarterly 54:208-23.
Psychology 62:207-18.
Nass, Clifford and Jonathan Steuer. 1993.
"Voices, Boxes, and Sources of Messages:
Steinberg, Ronnie. 1995. "GenderedInstructions:
Cultural Lag and Gender Bias in the Hay SysComputers and Social Actors." Human Comtem of Job Evaluation." Pp. 57-92 in Gender munication Research 19:504-27.
Inequality at Work,edited by J. Jacobs. ThouOppenheimer, Valerie. 1970. The Female Labor sand Oaks, CA: Sage.
Force in the United States: Demographic and
Economic Factors Governing Its Growth and Stewart, Penni and James Moore. 1992. "Wage
Disparities and Performance Expectations."
Changing Composition. Westport, CT: GreenSocial Psychology Quarterly 55:78-85. wood. Pugh, Meredith and Ralph Wahrman. 1983. Stone, Pamela. 1995. "Assessing Gender at
Work: Evidence and Issues." Pp. 408-23 in
"Neutralizing Sexism in Mixed-Sex Groups:
Gender Inequality at Work,edited by J. Jacobs.
Do Women Have to Be Better than Men?"
Thousand Oaks, CA: Sage.
American Journal of Sociology 88:746-62.
Quist, Theron and Phillip Wisely. 1991. "Gender Strober, Myra. 1984. "Toward a General Theory of Occupation Sex Segregation: The Case of
Attribution:Structuraland Behavioral Power in an Experimental Setting." Paper presented at
Public School Teaching." Pp. 144-56 in Sex

THE CONSERVATION OF GENDER INEQUALITY
Segregation in the Workplace, edited by B.
Reskin. Washington, DC: National Academy
Press.
Strober, Myra and Carolyn Arnold. 1987. "The
Dynamics of Occupational Segregation among
Bank Tellers." Pp. 107-48 in Gender in the
Workplace, edited by C. Brown and J. Pechman. Washington, DC: Brookings Institution.
Stryker, Sheldon. 1980. Symbolic Interactionism.
Menlo Park, CA: Benjamin/Cummings.
Suls, Jerry and Thomas Wills. 1991. Social Comparison: Contemporary Theory and Research.
Hillsdale, NJ: Erlbaum.
Turner, John. 1987. Rediscovering the Social

235

Group: A Self-Categorization Theory. Oxford,
England: Blackwell.
West, Candance and Don Zimmerman. 1987.
"Doing Gender." Gender and Society 1:12551.

Wharton, Amy and James Baron. 1987. "So
Happy Together? The Impact of Gender Segregation on Men at Work." American Sociological Review 52:574-87.
Wright, Rosemary and Jerry Jacobs. 1994. "Male
Flight from Computer Work: A New Look at
Occupational Resegregation and Ghettoization." American Sociological Review 59:
511-36.

References: Acker, Joan. 1989. Doing Comparable Worth. Bargh, John. 1989. "Conditional Automaticity: Varieties of Automatic Influence in Social Perception and Cognition." Pp Baron, James and Andrew Newman. 1990. "For What It 's Worth: Organizations, Occupations, Baron, James and Jeffrey Pfeffer. 1994. "The Social Psychology of Organizations and Inequality." Social Psychology Quarterly 57:190-209. Becker, Gary. 1957. The Economics of Discrimination. Chicago, IL: University of Chicago. Berk, Sarah Fenstermaker. 1985. The Gender Factory: TheApportionmentof Workin American Households Bielby, William and Denise Bielby. 1995. "Cumulative versus Continuous Disadvantage in an Unstructured Labor Market: Gender Differences in the Careersof Television Writers."Pp. Blau, Peter and Joseph Schwartz. 1984. Crosscutting Social Circles: Testing a Macrostructural Theory of Intergroup Relations Brewer, Marilynn. 1988. "A Dual Process Model of Impression Formation." Pp Brewer, Marilynn and Layton Lui. 1989. "The Primacy of Age and Sex in the Structure of Bridges, William and Robert Nelson. 1989. "Markets in Hierarchies:Organizational and Market Influences on Gender Inequality in a State Pay Carli, Linda. 1991. "Gender, Status, and Influence." Pp. 89-113 in Advances in Group Processes, vol. 8, edited by E. Lawler, B. Markovsky, C. Ridgeway, and H. Walker. Greenwich, CT: JAI. Cook, Karen. 1975. "Expectations, Evaluations, and Equity." American Sociological Review Cota, Albert and Kenneth Dion. 1986. "Salience of Gender and Sex Composition of Ad Hoc Crosby, Faye. 1982. Relative Deprivation and Working Women Deaux, Kay and Tim Emswiller. 1974. "Explanation of Successful Performance on Sex-Linked Tasks: What Is Skill for the Male Is Luck for Deaux, Kay and Mary Kite. 1987. "Thinking about Gender." Pp Deaux, Kay and Brenda Major. 1987. "Putting Gender into Context: An Interactive Model of Eagly, Alice. 1987. Sex Differences in Social Behavior: A Social-Role Interpretation.Hillsdale, NJ: Erlbaum. England, Paula. 1984. "Wage Appreciation and Depreciation: A Test of Neoclassical Economic England, Paula and Irene Browne. 1992. "Internalization and Constraint in Women 's Subor- 233 Fiske, Susan. 1992. "Thinking Is for Doing." Journal of Personality and Social Psychology Fiske, Susan and Shelley Taylor. 1991. Social Cognition Foschi, Martha, Larissa Lai, and Kirsten Sigerson. 1994. "Gender and Double Standards in the Assessment of Job Applicants." Social Psychology Quarterly 57:326-39. Goffman, Erving. 1977. "The Arrangement between the Sexes." Theory and Society 4:30131. Harris,Monica and Robert Rosenthal. 1985. "Mediation of InterpersonalExpectancy Effects: 31 Meta-Analyses." Psychological Bulletin 97: