The Ethnic, Race, and Gender Gaps in Workplace Authority

Changes over Time in the United States

Bibliographic Data

ID2277342
AuthorsBeth Mintz (University of Vermont, corresponding author), Daniel H Krymkowski (University of Vermont)
Year2010
Volume51
Issue1
Pages20-45
Publication date2010-02-01
Peer ReviewedYes
Open AccessYes
TypeARTICLE
VenueSociological Quarterly (JOURNAL)
Journal identifiersISSN: 0038-0253 • E-ISSN: 1533-8525
PublisherInforma UK Limited (PUBLISHER • GB)
DOI10.1111/j.1533-8525.2009.01165.x
OpenAlexW2108427123
LanguageEN
Citations received15
References cited46

We analyze factors explaining differences in hierarchical authority between men and women within and across categories of race and ethnicity in two time periods, finding that the processes leading to authority within the workplace operate differently by gender than by race or ethnicity. The demand-side factor, percentage of women in an occupation, helps explain authority differences between men and women in most groups. Supply-side factors, and, in white–black comparisons, occupational location, contribute to differences by race and ethnicity within genders. In the later period, education is particularly important for Hispanic men reflecting, we believe, the recent surge in immigration rates. ACKNOWLEDGMENTSAn earlier version of this article was presented at the Meeting of the Research Committee on Social Stratification and Mobility of the International Sociological Association in Oslo, Norway, May 2005. The authors contributed equally to this article. They would like to thank three anonymous TSQ reviewers and Steven Rytina for extremely helpful comments.NOTESNotes1 Research utilizing a common, single, occupational classification scheme has documented that these changes are not an artifact of changing classifications (e.g., CitationJacobs 1992).2 Because of limitations in sample sizes created by examining both race/ethnicity and gender, we consider Hispanics as one group. Although there is much precedent for doing this (cf. CitationSmith 2001; CitationSmith and Elliott 2002), the little research available that has distinguished between Hispanics by country of origin has found important differences between groups (CitationBrowne 1999; CitationEngland et al. 2004), suggesting that a more detailed analysis is a fruitful avenue for future research.3 Details on how we did this, as well as the reasons for our time period selection, are included in the methods section.4 Other racial and ethnic groups have even smaller numbers in the GSS. For instance, there are only 83 individuals of Asian origin in our entire cumulative data set, a number far too low for analytic purposes. The GSS data were weighted to adjust for nonresponse and to reflect the fact that the GSS is a sample of households, not individuals. Also, the interviews conducted in Spanish in 2006 were excluded in order to enhance comparability with earlier years.5 Percentage female in an occupation did not change much during this time frame, however. The correlation between percentage female in an occupation in 1983 and 2002 was .95 in the CPS data.6 An important idea behind the ordinal probit (or logit) model is that the measured dependent variable can be viewed as a discrete realization of a latent continuous variable whose metric is unknown. Thus, in order to estimate the model, identifying restrictions must be imposed. In the ordered probit model, one restriction is that the error variance is equal to 1; it is set to π2/3 in the ordered logit model. This implies that the variance of the latent variable changes when additional independent variables are added to the equation, making it, strictly speaking, incorrect to compare coefficients across models with different sets of independent variables. To deal with this problem, a number of authors have recommended the computation of "y*-standardized" coefficients (see, e.g., CitationLong 1997:128–9). We computed these coefficients and found the comparisons between them across models to be qualitatively identical and quantitatively very similar to the results reported herein. These analyses are available from the authors upon request.7 We also examined the effect of age of children and an interaction effect between having children and gender. We found no statistically significant effects for these factors.8 An anonymous reviewer noted that such changes over time might reflect nonlinearities in the relationship between authority and percentage female. If, for instance, the relationship was nonlinear for Hispanics in the late period, then the size of the linear effect we estimated might be too low; this in turn might impact upon the percentage of the gap that is explained. We examined a number of different nonlinear specifications but still found no effect of percentage female on authority for Hispanics in the late period.9 We again would like to point out that these results are tentative given small sample sizes.10 An anonymous reviewer suggested that perhaps women are more likely to be employed in large organizations, which have recently witnessed declines in opportunities for exercising authority owing to restructuring of various kinds. This would explain why the increase in women's representation in the managerial ranks has not been accompanied by an increase in access to authority. We looked at firm size by gender in both the CPS and the GSS and found little association between these variables.11 We also conducted a test for an interaction effect in this context and found that the effect of being married on authority is the same for black and white women

Demographic economics · Economics · Ethnic group · Immigration · Political science · Race (biology) · Sociology · Demography · Employment and Welfare Studies · Gender Studies · Gender, Labor, and Family Dynamics · Law · Psychology · Work-Family Balance Challenges

Tools

Open DOISci-Hub