U.S. Equal Employment Opportunity Commission
Meeting of February 15, 2012 - Unlawful Discrimination Against Pregnant Workers and Workers with Caregiving Responsibilities
Thank you, Madam Chair and the Commissioners, for the invitation to present my work and related research to the commission. My name is Stephen Benard, and I am an assistant professor of sociology at Indiana University. In my work, I use social science experiments to understand how demographic characteristics of workers – such as their gender or parent status – affect the way they are evaluated in the workplace.
Because social science experiments may be unfamiliar to some audience members, I will briefly explain why social scientists use experiments and what kind of information experiments provide. Experiments are useful for understanding whether demographic characteristics affect how workers are evaluated, independently of their education, experience, productivity, or other workrelevant factors. This is because experiments allow us to make comparisons across identical workers, in a way that would not be possible with other types of data.
To give a simple example, an experiment on gender and hiring might ask participants to evaluate the resumes of job applicants. These resumes would be kept identical, with one exception: they would vary on whether the applicant has a typically male name or a typically female name. Because the resumes are identical, we know that any difference in the way the two resumes are evaluated is due to their gender, rather than their qualifications.
The reason why we may see differences is because stereotypes can affect the way workers are evaluated, independently of their actual productivity as workers. I do not have the space here to summarize the many decades of research on stereotyping, but a few important points are worth noting.
First, stereotypes tend to be widespread within a particular culture, although the content of stereotypes varies across cultures. Second, stereotypes can operate consciously, but also unconsciously. This means that sometimes our behaviors are affected by stereotypes even without us being aware of it, and even if we do not endorse the stereotype or believe it to be true. Third, individuals can hold stereotypes – even negative stereotypes – about groups to which they belong.1
In this statement, I will briefly describe research on the motherhood wage penalty, which motivated the work my collaborators and I have conducted. I will then describe some of the work I have been involved in, as well as relevant work by others.
The Motherhood Penalty in Wages
The work my collaborators and I have conducted was motivated in part by earlier work showing a wage gap between mothers and non-mothers. Economists and sociologists who analyze large, nationally representative samples have consistently found that mothers experience worse labor market outcomes than fathers or women without children (Anderson, Binder and Krause 2003; Budig and England 2001; Glauber 2007; Waldfogel 1997; Waldfogel 1998). For example, Budig and England (2001) find a wage gap for mothers of about 5% per child, after statistically controlling for education, work experience, full or part-time status, and a variety other factors known to affect wages.
This wage penalty cumulates over the life course, producing a substantial lifetime disadvantage for mothers (Crittenden 2001; Sigle-Rushton and Waldfogel 2004). In fact, motherhood constitutes a significant risk factor for poverty (Misra, Budig, and Moller 2005). Given that a majority of women with children under 18 work in the paid labor market (68% in 2008, U.S. Bureau of Labor Statistics 2008), this wage gap affects a broad segment of the U.S. population. Furthermore, mothers’ wages have grown more slowly than the wages of non-mothers, leading some scholars to suggest that the gender gap in wages may be primarily a motherhood gap (Glass 2004).
However, these findings on the wage gap cannot tell us whether mothers earn less because they experience discrimination, or because they are less productive at work. Analyses of survey data can show that a wage gap remains, after controlling for other factors. However, they cannot tell us whether the remaining gap is due to discrimination, or because we have failed to control for some other important factor. In order to know if parent status is a basis for discrimination, we need to be able to compare identical sets of individuals who differ only on parent status. In particular, we need to compare parents and non-parents who are equally productive at work.
Experiments offer one way to compare otherwise identical individuals by holding all other workrelevant characteristics of employees or job applicants constant. These studies can take a variety of forms, but in general ask participants to evaluate individuals in a work-like setting. For example, they may ask participants to rate resumes, watch a video of an interview, or role-play a meeting. The resumes, interviews, or other materials used in an experimental study are generally created by the experimenter to ensure they are identical on all traits other than those under study.
In the next section, I will discuss experimental work on the motherhood penalty conducted with my collaborators.
An Experimental Study of the Motherhood Penalty
To test whether mothers are evaluated differently than non-mothers with identical qualifications, we conducted a laboratory experiment in which 192 (84 male and 108 female) paid undergraduate volunteers evaluated pairs of resumes (for a full description of the methods and results, see Correll, Benard, and Paik 2007). The applicants were ostensibly for a vice president of marketing position at the east coast branch of a California-based telecommunications start-up.
Each pair of applicants included resumes and other application materials for two same-gender (male or female), same-race (African American or white) candidates who differed on parent status. We signaled the gender and race of the employees using first names that are more common for members of those demographic groups (based on analyses of birth certificate data by Bertrand and Mullainathan , see figure 2). We signaled the parent status of the applicants in two ways. First, we included work for a parent-teacher association as an activity on the parent resume (see figure 1). For the non-parent, we listed a similar position in a neighborhood association. Second, we included a memo from an interview with the applicant, in which the author of the memo mentioned that the applicant did nor did not have children.
Figure 1: Signaling parent status on the resume.
Each participant thus rated resumes for either two white women, two African-American women, two white men, or two African-American men, one of whom was presented as a parent and one of whom was presented as a non-parent. The resumes were closely matched and pre-tested to ensure that participants viewed the applicants as equally qualified. In addition, to be sure that any differences in ratings of the applicants were not due to differences in the resumes, we created two different resumes, and each was presented as the parent’s resume for half of the study participants, and the non-parents’ resume for the other half of study participants. The conditions are shown in figure 2.
Figure 2: Experimental Conditions
After participants had time to examine the application materials, we asked them evaluate the resumes on three types of measures: the applicant’s competence and commitment, the standards to which the applicants should be held, and a set of “organizational rewards”, including whether they thought the applicant should be hired, and what starting salary they would propose for the applicants.
I will describe several key findings below. Please see the published article for the full findings. While I discuss a subset of the findings here, it is worth noting that the results were generally consistent across all of our ratings measures. In addition, the magnitude of the motherhood penalty did not vary by race, and so the results below pool across race of the applicants.
One of our measures was the perceived competence of the applicants. Participants rated the perceived competence of the applicants on a series of 7 point scales, evaluating the extent to which they viewed the applicant as capable, efficient, skilled, intelligent, independent, selfconfident, aggressive, and organized. These were averaged to form a single composite measure of competence. The findings from the experiment are shown in figure 3.
Figure 3: Perceived competence of the applicants by parent status.
Figure 3 shows that mothers were perceived as about 10% less competent than otherwise equivalent women without children. In contrast, fathers and men without children were seen as equally competent.
We also assessed whether mothers were held to different performance standards than other types of workers. For example, we asked participants how often the applicant could come in late or leave early and still be considered for hire. Figure 4 shows that participants stated that women without children could arrive late or leave early about a half day per month more often than otherwise identical mothers. This indicates that women with children appear to be held to stricter standards of punctuality than women without children.
Interestingly, we found the opposite pattern for men. Fathers were held to more lenient standards for lateness than men without children. In the same study, we found that fathers were viewed as significantly more committed than men without children. It may be the case that having children leads evaluators to view men as more dependable, and women as less dependable.
The third group of outcomes we examined were what are called “organizational reward” measures. This group of measures includes whether the applicant was recommended for hire, the recommended starting salary, whether the applicant should be put on a “fast track” to receive additional management training, and the applicant’s likelihood of being promoted if hired.
Figure 4: Days allowed late per month without penalty.
The results across these measures were consistent, with mothers penalized relative to women without children, and fathers advantaged relative to men without children. To illustrate, figure 5 shows the proportion of workers in each category recommended for hire.
Figure 5: Proportion of each type of applicant recommended for hire.
As shown in figure 5, women with children are recommended for higher at lower rates than all other workers, being recommended for hire about 47% of the time. In contrast, fathers were seen as more hirable than men without children.
The results of our experiment show that mothers are consistently stereotyped as less competent and less committed, and disadvantaged in hiring and promotion relative to (otherwise equivalent) applications from women without children. In addition, we found that fathers were sometimes (though not always) rated more positively than men without children.
A few other findings are worth noting. First, the magnitude of the penalty was similar whether the mothers were presented as African-American or white. However, there was evidence of general racial discrimination, with candidates presented as African American being offered significantly lower starting salaries than candidates presented as white. Second, the gender of the study participants did not affect their evaluations: both men and women penalized mothers to similar degrees.
An Audit Study of the Motherhood Penalty
One question that one might raise in response to our study is whether the evaluations of college students are similar to the evaluations that real managers would make. In fact, this question has been the topic of a number of studies. Such studies compare how college students and managers evaluate resumes, and have found that their evaluations tend to be very similar (Cleveland and Berman 1987; Cleveland 1991; Olian and Schwab 1988). Based on these direct comparisons, we expect our student sample and actual managers to make similar evaluations. To test this hypothesis directly, we conducted an audit study of actual employers.
In the audit study, we selected a sample of jobs from a newspaper in a large Northeastern city. The jobs were similar to the marketing position we used in the laboratory study. The jobs were randomly assigned to receive either a pair of male applications or a pair of female applications. One application was a parent, the other a non-parent. We applied to 638 employers over a period of 18 months. Our dependent measure in this study was whether a resume received a callback for an interview. The results are summarized in figure 6.
Figure 6: Percent of applicants receiving a callback in the audit study.
We found that, in our audit study of an actual labor market, women without children were called back more than twice as often as mothers (a statistically significant difference). Fathers also tended to be called back more often than men without children, though this difference was not statistically significant.
The audit study has the advantage of providing evidence from an actual labor market, while the laboratory study has the advantage of allowing us to ask participants many more questions, and thus get a better picture of why they rated mothers and non-mothers differently. Taken together, the two studies provide evidence that women experience a range of penalties when they become mothers, and the penalty exists in real labor markets.
Related Studies of the Motherhood Penalty
I would also like to mention that a number of other related studies exist and may be of interest to the committee. I recommend our review in the Hastings Law Journal for a fuller picture of past research (Benard, Paik, and Correll 2008), but point out a few particular findings below.
Additional laboratory studies of the motherhood penalty: The basic finding that mothers are evaluated more negatively than otherwise equivalent non-mothers has been found across a range of studies. These studies vary in the specific details of their design, sample, and dependent measures, but are consistent in showing a penalty for mothers (e.g. Cuddy et al. 2004; Fuegen et al. 2004). At least one article (Heilman and Okimoto 2008) finds a penalty for mothers using both a student sample and a sample of full-time workers in business organizations. We have also replicated our laboratory experiment findings in a follow-up study (Benard and Correll 2010). This study also included a condition designed to overcome evaluator’s doubts about mothers’ competence, by including extremely positive performance evaluations in the application materials. This did lead mothers to be seen as more competent and committed. However, it also led them to be seen as less likeable and warm than highly successful fathers, at least by female study participants.
Pregnancy discrimination: While many studies focus on discrimination against mothers, there is also a subset of studies examining discrimination against pregnant women (Corse 1990; Halpert et. al 1993; Halpert and Hickman Burg 1997; DeNicholas Bragger 2002; Cunningham and Macan 2007; Masser et al 2007). These studies are excellent for examining how an identical woman would be treated when pregnant versus not pregnant. This is because they use experimental stimuli – such as videos of a job interview – in which an actress enacts an identical script, but appears to be pregnant in one video, and not pregnant in another video. Like work on the motherhood penalty more generally, these studies find penalties for pregnant women relative to otherwise identical non-pregnant women.
Discrimination faced by men: An important frontier in work on the motherhood penalty is whether fathers also face discrimination. Our work found that fathers tended to be evaluated as positively or more positively than men without children. However, the evaluators in our studies may have assumed that the male applicants – but not the female applicants – had a partner serving as the primary caregiver for the children. This raises the question of whether men experience discrimination when they engage in caregiving activities. Increasingly, research has begun to explore this possibility. One study (Wayne and Cordeiro 2003) finds that men who took leave to care for a child or an elder parent were seen as worse “organizational citizens” than those who did not take leave. In addition, a study by Rudman and Mescher (forthcoming) finds that men who requested family leave were perceived as weaker, less masculine, and at greater risk for being demoted or downsized. This suggests that the motherhood penalty may be more accurately described as a caregiver penalty. This question is closely related to another issue: perceptions of individuals who use flexible work options.
The stigma of flexible work: One proposed solution to work-family conflict is for the workplace to become more flexible. This could include broadening the use of practices such as family leave, working part time, or working flexible hours. Both the Wayne and Cordeiro and Rudman and Mescher papers mentioned above suggest that there are work-related penalties attached to leave-taking. A forthcoming edition of the Journal of Social Issues, edited by Joan Williams, Shelley Correll, Jennifer Glass, and Jennifer Berdahl, will examine the flexibility stigma in depth.
The role of laws: Another proposed solution is that the existence of laws supporting caregivers in the workplace might reduce bias towards caregivers. Aside from enforcement, some scholars have proposed that laws are taken as a signal of what society considers appropriate behavior. Thus, when individuals see that the law supports a behavior, they are less likely to stigmatize that behavior. One experiment found that reminding evaluators about FMLA policies reduced biases towards those who used family leave (Albiston, Correll, and Tucker 2012).
In conclusion, our work and the work of others finds consistent evidence of a penalty for mothers. The evidence includes a number of experiments, across a range of samples and research designs. More recently, work has turned to investigating the possibility of a broader “caregiver penalty”, including the ways in which men might be penalized for caregiving. I appreciate the opportunity to address the commission, and hope you will find this statement useful.
Albiston, Catherine R., Shelley J. Correll, and Traci Tucker. 2012. “Law, Norms, And The Motherhood/Caretaker Penalty.” Working paper.
Anderson, Deborah J., Melissa Binder and Kate Krause. 2003. “The Motherhood Wage Penalty Revisited: Experience, Heterogeneity, Work Effort and Work-Schedule Flexibility.” Industrial and Labor Relations Review 56:273-294.
Bargh, John A., Mark Chen, and Lara Burrows. 1996. “Automaticity of Social Behavior: Direct Effects of Trait Construct and Stereotype Activation on Action.” Journal of Personality and Social Psychology 71: 230-44.
Benard, Stephen, In Paik, and Shelley J. Correll. 2008. “Cognitive Bias and the Motherhood Penalty.” Hastings Law Journal 59:101-129.
Bertrand, Marianne and Sendhil Mullainathan. 2003. “Are Emily and Greg more employable than Lakisha and Jamal? A field experiment on labor market discrimination.” National Bureau of Economic Research Working Paper Series: Working paper number 9873. Retrieved September 2003. (http://www.nber.org/papers/w9873).
Budig, Michelle and Paula England. 2001. “The wage penalty for motherhood.” American Sociological Review 66:204-225.
Cleveland, Jeanette N. and Andrew H. Berman. 1987. “Age Perceptions of Jobs: Agreement Between Samples of Students and Managers.” Psychological Reports 61: 565-566.
Cleveland, Jeanette N. 1991. “Using Hypothetical and Actual Applicants in Assessing Person- Organization Fit: A Methodological Note.” Journal of Applied Social Psychology 21:1004-1011.
Correll, Shelley J., Stephen Benard, and In Paik. 2007. “Getting a job: is there a motherhood penalty?” American Journal of Sociology 112:1297-1338.
Crittenden, Ann. 2001. The Price of Motherhood: Why the Most Important Job in the World is Still the Least Valued. New York: Metropolitan Books.
Devine, Patricia G. 1988. “Stereotypes and Prejudice: Their Automatic and Controlled Components.” Journal of Personality and Social Psychology 56: 5-18.
Fuegen, Kathleen, Monica Biernat, Elizabeth Haines, and Kay Deaux. 2004. “Mothers and fathers in the workplace: how gender and parental status influence judgments of jobrelated competence.” Journal of Social Issues 60:737-754.
Glauber, Rebecca. 2007. “Marriage and the motherhood penalty among African Americans, Hispanics, and Whites.” Journal of Marriage and the Family 69:951-961.
Glass, Jennifer. 2004. “Blessing or curse? Work-family policies and mother’s wage growth over time.” Work and Occupations 31:367-394.
Greenwald, Anthony G. and Linda Hamilton Krieger. 2006. “Implicit Bias: Scientific Foundations.” California Law Review 94:945-967.
Halpert, Jane A., Midge L. Wilson and Julia Hickman. 1993. “Pregnancy as a source of bias in performance appraisals.” Journal of Organizational Behavior 14:649-663.
Harkness, Susan and Jane Waldfogel. 1999. “The family gap in pay: Evidence from seven industrialised countries.” CASE paper 29: Centre for Analysis of Social Exclusion.
Heilman, Madeline E. and Tyler G. Okimoto. 2008. “Motherhood: A Potential Source of Bias in Employment Decisions.” Journal of Applied Psychology 93: 189-198.
Kunda, Ziva and Stephen J. Spencer. 2003. “When Do Stereotypes Come to Mind and When Do They Color Judgment? A Goal-Based Theoretical Framework for Stereotype Activation and Application.” Psychological Bulletin 129: 522-44.
Misra, Joya, Michelle Budig, and Stephanie Moller. 2005. “Employment, wages, and poverty: reconciliation policies and gender equity.” Paper presented at the Annual Meeting of the American Sociological Association, Philadelphia, PA.
Olian, Judy D. and Donald P. Schwab. 1988. “The Impact of Applicant Gender Compared to Qualifications on Hiring Recommendations: A Meta-analysis of Experimental Studies.” Organizational Behavior and Human Decision Processes 41: 180-195.
Cuddy, Amy J.C., Susan T. Fiske, and Peter Glick, P. 2004. “When professionals become mothers, warmth doesn’t cut the ice.” Journal of Social Issues 60:701-718.
Corse, Sara J. 1990. “Pregnant managers and their subordinates: The effects of gender expectations on hierarchical relationships.” Journal of Applied Behavioral Science 26:25-48.
Halpert, Jane A. and Julia Hickman Burg. 1997. “Mixed Messages: Co-Worker Responses to the Pregnant Employee.” Journal of Business and Psychology 12: 241-253.
DeNicholas Bragger, Jennifer, Eugene Kutcher, John Morgan, and Patricia Firth. 2002. “The Effects of the Structured Interview on Reducing Biases Against Pregnant Job Applicants.” Sex Roles 46:215-226.
Cunningham, Jennifer and Therese Macan. 2007. “The Effects of Applicant Pregnancy on Hiring Decisions and Interview Ratings.” Sex Roles 57:497-508.
Masser, Barbara, Kirsten Grass and Michelle Nesic. 2007. “We Like You, But We Don’t Want You’—The Impact of Pregnancy in the Workplace.” Sex Roles 57:703-712.
Rudman, Laurie A. and Kris Mescher. Forthcoming. “Penalizing Men Who Request a Family Leave: Is Flexibility Stigma a Femininity Stigma?” Journal of Social Issues.
Sigle-Rushton, Wendy and Jane Waldfogel. 2004. “Motherhood and women’s earnings in Anglo-American, Continental European, and Nordic countries.” Paper presented at Conference on Cross-National Comparisons of Expenditures on Children, Princeton.
Waldfogel, Jane. 1997. “The effect of children on women’s wages.” American Sociological Review 62:209-17.
Waldfogel, Jane. 1998. “Understanding the ‘family gap’ in pay for women with children.” Journal of Economic Perspectives 12:137-156.
Wayne, Julie Holliday, & Cordeiro, Bryanne L. 2003. “Who is a good organizational citizen? Social perception of male and female employees who use family leave.” Sex Roles, 49:233-246.