Meeting Minutes - Oct. 10 2012 - Biases in Science

RACIAL BIASES

• Minorities are underrepresented in STEM fields 
• Socioeconomic status and race are often related; the playing field is not level
• In other fields, a resume with a stereotypically black name will elicit lower hiring success than the very same resume with a stereotypically white name (Bertrand and Mullainathan 2003).

GENDER BIASES

• Women are underrepresented in STEM fields
• Women are paid less and are less successful at getting grants/getting larger grants
• Letters of recommendation for females vs. males
• PNAS Paper (Moss-Racusin et al. 2012)
• Controlled, randomized, experimental study
• Question: Does discrimination based solely on gender exist affect hiring decisions in academic 
• science?
• Problem: Most research statistically analyzes outcomes for women at different levels of academic science and correlates it with questionnaire data. These correlational studies are not adequately to reveal why there is a “leaky pipeline” for women in science. Randomly controlled studies are necessary to determine causation, but are difficult to execute.
• Methods: The authors distributed application materials for a lab manager position to faculty in Biology, Chemistry, and Physics departments at prominent research institutions. The gender of the names and pronouns on the application materials were varied (i.e. John and Jennifer).  The materials were identical in all other aspects. Faculty evaluated materials based on the applicant’s competence, hirability, deservingness of mentoring, starting salary offers, and likeability. Afterward, they took a short survey that evaluated their subtle gender biases based on the Modern Sexism Scale.
• Results: Male applicants were rated as significantly more competent, hirable, and deserving of mentoring by both male and female faculty (all p<0.01). They were offered hirer starting salaries ($30,238.10 vs $26,507.94, t(142) = 3.42, p<0.01). Females were rated as significantly more likable. There was a significant interaction between student gender and pre-existing subtle bias that predicted perceptions of composite competence (β = −0.39, P < 0.01), hireability (β =−0.31, P < 0.05), and mentoring (β = −0.55, P < 0.001). Female students were rated as more likable (p<0.05). 
• Online Response to PNAS Paper
• Some people think it's ok to be biased against women for their ability to get pregnant (due to maternity leave concerns)
• What are the economic costs of hiring a female candidate who takes maternity leave for a few months vs. hiring a less qualified male candidate in her stead? 
• One commenter removed her first name from her CV and used only initials and found that her hiring success improved

SOLUTIONS

• U. Wisconsin guidelines for minimizing bias when reviewing job applicants (http://wiseli.engr.wisc.edu/docs/BiasBrochure_2ndEd.pdf)
• Many times biases are sub-conscious and can be recognized and avoided with awareness; differences should be recognized and celebrated
• Allocate enough time to hiring so that biases don't seep into the process and unfair screening short-cuts aren't used
• Determine screening requirements before time (eg. GPA, previous research experience)
• Take names off of CVs (use first initial)?  Avoid pronouns in letters?
• Level the playing field
• Affirmative Effort programs; diverse communities are better for everyone
• Be proactive about seeking out networking opportunities (your CV won't be judged based on your name if you're already known)