Affirmative Action Increases Female Leadership in Scientific Research

Labor Economics Public Policy No Comments
A group of four diverse scientists in white lab coats work in a modern, well-lit chemistry laboratory. In the foreground, a female lead researcher stands on the left, gesturing as she gives instructions. Across from her, an older male scientist looks up from a notebook, a younger male researcher pours a blue liquid into a flask, and a younger blonde woman looks toward the leader while holding a tablet. High windows in the background reveal a city skyline.

Addressing the underrepresentation of women in science, particularly in leadership positions, is important for several reasons. For instance, having more women leading projects may help diversify the topics that are researched. It may also reduce biases that hinder efficiency in the scientific profession. 

In the Barcelona School of Economics Working Paper 1478, “Promoting Female Talent in Research: Evidence from an Affirmative Action Policy”, Lídia Farré and Judit Vall Castelló evaluate the impact of an affirmative action (AA) policy that granted additional evaluation points to research groups led by women in Catalonia, Spain, on the promotion of female talent in science.

The authors find that the AA policy increased the share of research groups led by women and eliminated gender penalties in the evaluation process. 

Read BSE Working Paper 1478 on bse.eu

Affirmative Action Increased Female Representation

The AA policy awarded extra evaluation points to research groups led by female principal investigators (PIs). The policy was implemented in 2021, and the researchers use data from the universe of all applications to the funding calls in 2014, 2017, and 2021 to compare outcomes before and after its implementation. 

The authors find that the policy led to a significant increase in the number of research groups led by women. This change was driven by two key factors: 

  1. A notable shift in the gender of the PI in existing groups, with 19% of the groups in 2021 transitioning from a male to a female PI
  2. The creation of new groups led by women, which account for 9% of all the groups in 2021 

These shifts reflect the successful promotion of female leadership in research prompted by the policy. The positive effects of the policy are larger in STEM fields, where the pre-existing gender gaps are also larger.

Affirmative Action Corrected Gender Penalties

Prior to the introduction of the AA policy, research groups led by women consistently received lower evaluation scores than those led by men, with an average gap of 0.32 points in 2017. 

By 2021, this gap had narrowed by approximately 50% before accounting for the additional AA points (see Figure 1 below – Raw Score w/o AA). 

Controlling for group and PI characteristics (all the variables that are available to evaluators) narrows the gender gap in scores (Figure 1 – Score (controls) w AA – dark red bar). 

When the additional AA points are considered, the policy successfully offset pre-existing, unexplained penalties faced by female-led groups in the funding evaluation process (Figure 1 – Score (controls) w AA – light red bar). 

Finally, the figure also shows that the AA policy equalized the probability of receiving funding for female PIs, bringing it in line with that of comparable male PIs (Figure 1 – Granted AA).

Red bars represent the coefficient on the Female indicator, while light red bars represent the coefficient on the interaction between Female and 2021, from regressions where the dependent variable corresponds to the outcome indicated on the x-axis. Raw Score refers to a regression without controls, while Score (controls) includes controls for group and PI characteristics. Labels marked w/o AA refer to scores before adding the additional points from the AA policy, whereas w AA includes those additional points.
Figure 1. The Effect of the Policy on Gender Gaps in Scores

Note: Red bars represent the coefficient on the Female indicator, while light red bars represent the coefficient on the interaction between Female and 2021, from regressions where the dependent variable corresponds to the outcome indicated on the x-axis. Raw Score refers to a regression without controls, while Score (controls) includes controls for group and PI characteristics. Labels marked w/o AA refer to scores before adding the additional points from the AA policy, whereas w AA includes those additional points.

Affirmative Action is More Effective than Gender Quotas in Evaluation Panels

The researchers also compare the AA policy to other potential strategies for promoting gender equity in research environments, such as gender quotas in review boards.  

They show that the share of female reviewers would have to increase from 0 to 50% to have the same effect on the promotion of female talent as the AA policy.

Given the cost that gender quotas impose on female researchers that act as reviewers, and the limited effects on the promotion of female researchers associated with these policies, the authors conclude that the AA policy is more cost-effective in promoting female talent.

A Strong Case for Affirmative Action Policies

In sum, Farré and Vall Castelló provide evidence that AA policies can be effective in addressing gender imbalances in competitive settings and highlights the potential of AA to counteract biases and promote fairer resource allocation. 

Their research also suggests that AA policies are more cost effective, as they directly compensate for gender biases in the evaluation process without imposing significant additional costs on female researchers.

Read BSE Working Paper 1478 on bse.eu