This chapter reviews research on the determinants and consequences of race and sex composition of organizations. Determinants include the composition of the qualified labor supply; employers' preferences, including the qualifications they require; the response of majority groups; and an establishment's attractiveness, size, and recruiting methods. The race and sex composition of an establishment affects workers' cross-group contact; stress, satisfaction, and turnover; cohesion; stereotyping; and evaluation. Composition also affects organizations themselves, including their performance, hiring and promotion practices, levels of job segregation, and wages and benefits. Theorydriven research is needed (a) on the causal mechanisms that underlie the relationships between organizational composition and its determinants and consequences and (b) on the form of the relationships between organizational composition and workers outcomes (e.g., cross-group contact, cohesion, turnover, etc). Research is needed on race and ethnic composition, with a special focus on the joint effects of race and sex.

We examine the effects of organizations' employment practices on sex-based ascription in managerial jobs. Given men's initial preponderance in management, we argue that inertia, sex labels, and power dynamics predispose organizations to use sex-based ascription when staffing managerial jobs, but that personnel practices can invite or curtail ascription. Our results-based on data from a national probability sample of 516 work organizations-show that specific personnel practices affect the sexual division of managerial labor. Net of controls for the composition of the labor supply, open recruitment methods are associated with women holding a greater share of management jobs, while recruitment through informal networks increases men's share. Formalizing personnel practices reduces men's share of management jobs, especially in large establishments, presumably because formalization checks ascription in job assignments, evaluation, and factors that affect attrition. Thus, through their personnel practices, establishments license or limit ascription. B ARON and Bielby (1980) encouraged researchers interested in labor market inequality to "bring the firm back in" because firms "link the 'macro' and 'micro' dimensions of work organization and inequality" (p. 738). Researchers who have taken up the call have demonstrated the importance of organizational structures, such as their nonprofit or government status or their size, for sustaining or eroding sex-based ascription (Baron 1991; Baron and Newman 1990; Nelson and Bridges 1999; Tomaskovic-Devey 1993). Relatively little is known, however, about the effects of establishments' personnel practices, and it is these practices that are the proximate causes of establishment-based gender inequality (Bielby 2000; Reskin 2000). We examine the effects of organizations' employment practices on sexbased ascription in managerial jobs for a national probability sample of establishments. Ascription exists when a status, position, or opportunity is allocated at least in part on the basis of an ascribed characteristic (Kemper 1974; Mayhew 1968; Parsons 1964). Baron (1991:143), for example, has contended that employers practice ascription when employees' ascribed characteristics directly influence their jobs or rewards. Thus, ascription involves differential treatment based on sex, race, ethnicity, and the like. Sex-based ascription does not necessarily involve invidious intent; it can occur when custom shapes employers' personnel practices, when sex stereotypes or jobs' sex labels affect allocation decisions, when decision-makers have the discretion to act on their biases, or when they use sex as a proxy for productivity or employment costs. Impulses toward ascription may originate in or-Direct all correspondence to Barbara Reskin

A central assumption of much of the previous research on race differences is that the process by which Blacks and Whites advance in the workplace is race blind so that if Blacks and Whites had the same amount of education and job experience and were located across the same bureaucratized structures, the gap in Black-White attainment in the workplace would disappear. The authors argue that to understand the systematic differences in Black-White outcomes in the workplace, we need to reexamine this assumption. The authors develop a theoretical argument for the existence of race-specific models of attainment that builds on Kanter's concept of “homosocial reproduction.” They then test this argument by estimating whether the determinants of promotion differ by race, using a nationally representative sample of U.S. workers and their employing organizations. Their findings indicate that the determinants of promotion systematically differ for Blacks and Whites.