Gender Diversity in Tech
An analysis of the implicit biases in the hiring processes of tech companies
Understanding Implicit Bias in Tech Hiring
White males have been historically overrepresented in science, technology, engineering and mathematics (STEM); despite efforts to increase gender and racial diversity in these fields, the pattern continues today.[i] Diversity advocates have criticized the homogeneity of STEM in academia as well as the workplace, and researchers have attempted to measure the impact of implicit biases on diversity in these fields.[ii] But now, with Tinder co-founder Justin Mateen’s sexist comments to Snapchat CEO Evan Spiegel’s old fraternity emails coming to the foreground, the specific issue of gender diversity in the technology industry has become one of increasing interest.[iii],[iv]
In recent years, women have earned a greater proportion of STEM degrees. Since 2008, women have earned more undergraduate degrees in biology than men. The gender balance of 2011 chemistry graduates was roughly equal.[v] Meanwhile, the trend in computer science and engineering looks bleak. In 1985, 37% of people who graduated with a bachelor’s degree in computer science were female but by 2011, that number had dropped to only 18%.[vi] While the reduced proportion of female computer science graduates seems to imply that the lack of gender diversity in the industry might resolve itself by moving more women through the academic pipeline, there also exists a persistent disparity in number of computer and information sciences PhD degrees earned by women and the number hired full time, which suggests that the women’s advancement in the industry may also be impeded.[vii]
In the summer of 2014, many of the largest technology companies released their diversity statistics, detailing the proportion of women on the technology side (often defined as jobs related to programming, product development, and user experience). The numbers are alarmingly low: women represent only 10% of Twitter’s tech employees, 15% of Yahoo’s, 17% of Google’s, and 24% of Ebay’s.[viii] Dropbox, a startup valued at over $10 billion, had 9 women on its engineering team of 143 people in 2013.[ix]
Why are there so few women in tech? There appears to be little biological difference in the science and math capabilities of men and women, and thus female lifestyle choices (such as child and family care) are often touted as the primary cause of the gender disparity.[x],[xi] In the tech industry specifically, the “brogrammer” culture is frequently cited as what deters women from the field.[xii] However, given the research on prejudice and stereotypes, it seems likely that implicit structural biases affect technology companies and could represent part of the problem.
Effect of Implicit Bias on Hiring Decisions
In the earlier stages of the social science research, prejudice was seen as abnormal: most people were good and unprejudiced, but there existed a small and evil subset of the population that was prejudiced, and if only they could be eliminated or treated, there would be no more prejudice.[xiii] This approach allowed for a lot of finger pointing but explained very little of the observed world. A more recent approach finds the opposite – that prejudice is rooted in normalcy.[xiv] Everyone is biased.
And it’s natural. Humans are social animals, and we evolved in coalitional groups. We rely on each other in order to survive, which means that it is advantageous for us to be able to distinguish between who is likely to help and who is not. Social space is divided into the in-group, an assortment of actors like us, and the out-group, a collection of dissimilar others.[xv] Moral obligations hinge on social membership, so much so that even young children understand social categories as marking people who are intrinsically obligated to one another.[xvi] Prejudice stems from this categorization: we have a systematic preference for the in-group. Antipathy towards the out-group is not necessary for prejudice to exist.[xvii] Simply put, the out-group might be very nice but the in-group is still a bit better.
Social categorization is a very pervasive cognitive framework. While our voiced and endorsed explicit attitudes might be egalitarian, we are often susceptible to subtle unconscious biases that manifest themselves in our actions.[xx] These implicit biases might not even reflect what we actively think or believe, but instead could simply stem from understanding of pervasive cultural stereotypes.[xxi]
Regardless of the root cause for their existence, implicit biases affect our behavior. In many scenarios, aversive racist biases result in unintended and negative consequences. For example, when Hodson, Dovidio, and Gaertner (2002) asked white participants to judge a white or black candidate with either very strong (high SAT scores and high GPA), moderate (either high SAT and low GPA or low SAT and high GPA), or very weak (low SAT and GPA) qualifications, there was no discrimination in both the very strong and very weak credentials cases. However, when the candidate had moderate qualifications, the white candidate was chosen significantly more than the black candidate. And, when asked which qualification was more important -- a high GPA or good SAT score -- participants were more likely to answer the qualification that backed up their biased judgment.[xxii] Decisions about applicants are easily susceptible to unconscious biases and manifest themselves in times when it is possible to justify the decisions on different criteria.
How might implicit biases affect gender hiring? Whereas racial biases are often accompanied by negative racial stereotypes, there is not a comparable negative stereotype for women. Both sexes tend to rate their attitudes towards women as higher than their attitude towards men, a finding dubbed the “women-are-wonderful” effect.[xxiii] Moreover, the Goldberg paradigm argues that people do not necessarily devalue women’s work either. In one study, for example, subjects were asked to judge the quality of an essay, and while the wording was identical, the author’s name was either a typically male or female name. In most cases, the gender of the author had little or no effect.[xxiv]
In ‘masculine’ contexts though (in an essay about football or war), gender matters. Subjects devalue women’s work. When hiring for male-dominated jobs such as police officer, people tend to prefer a resume with a man’s name to one with a woman’s name, despite the resumes being otherwise identical.[xxv] Some evidence suggests that raters judge women as less competent than men with identical backgrounds.[xxvi]
A recent study, Moss-Racusin et al. (2012) had science faculty rate application materials of a student, randomly assigned a male or female name, for a laboratory manager position. The only difference between the applications was the gender of the applicant’s name, yet both male and female faculty rated the male applicant as significantly more competent and hireable than the female applicant, and were more likely to offer mentorship to that applicant.[xxvii]
The effect size of that study was not huge, so one could argue that implicit bias has little effect. However, I would argue the opposite. Even with a small effect size, a gender bias would have cascading effects. Martell (1996) ran a computer simulation of a company with a pyramidal management structure with an equal gender ratio at the lowest-level positions. The company filled its higher-level positions with people from lower level positions depending on performance ratings. Even when sex differences explained only 1% of the variance of performance evaluations, only 35% of the highest-level positions were filled by women[xxviii]. “The effects of male-female differences are best determined not by the magnitude of the effect but its consequences in natural settings,” Martell argues. And in this case, taking into account the relative scarcity of women in the tech side, bias hurts women a lot.
If seeing a name can make such a difference, having raters judge resumes name-blind could be one possible way to circumvent this implicit bias. The National Institute of Health (NIH) is taking that approach in response to a similar situation. The NIH found that African Americans were denied grants more often, controlling for educational background, country of origin, training, previous research awards, publication record, etc. [xxix]. In response, they began to look into the implicit biases that affect raters, deciding that they would have raters judge applications name-blind. Maybe it's time that we too (in the tech industry) make some structural changes in the way we hire.
[i] National Science Foundation (2014). Women, Minorities, and Persons with Disabilities in Science and Engineering (National Science Foundation, Arlington).
[ii] Vescio et al. (1999). The stereotypic behaviors of the powerful and their effect on the relatively powerless. Handbook of Prejudice, Stereotyping, and Discrimination.
[iii] Waxman, O. “Snapchat CEO apologizes for explicit frat emails.” Time Magazine, 29 May 2014. Web.
[iv] Wortham, J. “Tinder is target of sexual harassment lawsuit.” New York Times, 01 July 2014. Web.
[v] Matson, J. (2013). Women are earning greater share of STEM degrees, but doctorates remain gender-skewed. Scientific American. Vol. 308(5). Retrieved from http://www.scientificamerican.com/article/women-earning-greater-share-stem-degrees-doctorates-remain-gender-skewed/
[vi] National Center for Women and Information Technology. “Women and information technology by the numbers.” NCWIT, 2013. Web.
[vii] National Science Foundation (2014). Women, Minorities, and Persons with Disabilities in Science and Engineering (National Science Foundation, Arlington).
[viii] Lafrance, A. “Tallying female workers isn’t enough to make tech more diverse.” The Atlantic, 11 Aug 2014. Web.
[ix] Wadhwa, V. “Dropbox’s hiring practices explain its disappointing lack of female employees.” Washington Post, 14 February 2014. Web.
[x] Spelke, E.S. (2005). Sex differences in intrinsic aptitude for mathematics and science?” A critical review. Am Psychol. Vol. 60: 950-958.
[xi] Ceci et al. (2014). Women in academic science: a changing landscape. Psychological Science in the Public Interest. Vol. 15(3): 75-141.
[xii] Weisul, K. (2014). It’s the culture, bro: why women leave tech. Inc. Magazine. Retrieved from http://www.inc.com/kimberly-weisul/its-the-culture-bro-why-women-leave-tech.html
[xiii] Dovidio, J.F., Hewstone, M., Glick, P., & Esses, V.M. (2010). Prejudice, stereotyping and discrimination: theoretical and empirical overview. In J. Dovidio, M. Hewstone, P. Glick, & V. Esses. The Handbook of Prejudice, Stereotyping, and Discrimination. 328-344. Thousand Oaks, CA: SAGE Publications Ltd.
[xiv] See 13
[xv] Dunham, Y. (Spring 2013). Research Methods in Cognitive Development. Lecture conducted from Yale University, New Haven, CT.
[xvi] Rhodes, M., & Chalik, L. (2013). Social categories as markers of intrinsic interpersonal obligations. Psychological Science. doi:10.1177/0956797612466267.
[xvii] See 13
[xviii] Debruine, L. (2002). Facial resemblance enhances trust. Proc. R. Soc. Lond. Vol. 269: 1307-1312.
[xix] Tajfel, H., Turner, J.C. (2004) The social identity theory of intergroup behavior. In: Jost, J.T.; Sidanius, J., Political psychology: Key readings. Psychology Press; p. 276-293.
[xx] Greenwald AG, Banaji MR (1995) Implicit social cognition: Attitudes, self-esteem, and stereotypes. Psychological Review. Vol. 102: 4—27.
[xxi] Glick, P., & Rudman, L. (2010). Sexism. In J. Dovidio, M. Hewstone, P. Glick, & V. Esses. The Handbook of Prejudice, Stereotyping, and Discrimination. 328-344. Thousand Oaks, CA: SAGE Publications Ltd.
[xxii] Dovidio et al. (2002). Why can’t we just get along? Interpersonal biases and interracial distrust. Cultural Diversity and Ethinic Minority Psychology, Vol. 8(2): 88-102.
[xxiii] Eagly, A., & Mladinic, A. (2011). Are people prejudiced against women? Some answers from research on attitudes, gender, stereotypes, and judgments of competence. European Review of Social Psychology, Vol 5(1): 1-35.
[xxiv] Glick, P., & Rudman, L. (2010). Sexism. In J. Dovidio, M. Hewstone, P. Glick, & V. Esses. The Handbook of Prejudice, Stereotyping, and Discrimination. 328-344. Thousand Oaks, CA: SAGE Publications Ltd.
[xxv] See 24.
[xxvi] Foschi, M. (2000). Double standards for competence: theory and research. Annual Rev Sociol. Vol. 26(1): 21-42.
[xxvii] Moss-Racusin et al. (2012). Science faculty’s subtle gender bias favor male students.
[xxviii] Martell, R.F., Lane, D.M., Emrich, C. (1996). Male-female differences: a computer simulation. American Psychologist. Vol. 51(2):157-158.
[xxix] Ginther et al. (2011). Race, Ethnicity, and NIH Research Awards. Science. Vol. 333(6045): 1015-1019.