Skip to main content

IFS/STICERD/UCL Development Work in Progress Seminar

Understanding and Eliminating Biased Inference and the Mismeasure of Women

Peter Ward-Griffin (LSE)

Thursday 07 March 2024 13:00 - 14:00

Many of our seminars and public events this year will continue as in person or as hybrid (online and in person) events. Please check our website listings and Twitter feed @STICERD_LSE for updates.

Unless otherwise specified, in-person seminars are open to the public.

Those unable to join the seminars in-person are welcome to participate via zoom if the event is hybrid.


About this event

Women are systematically less likely to become managers, CEOs, and politicians. They are also systematically underestimated when hiring for these positions. In addition, women earn substantially less when working. I investigate the role of a novel behavioral and econometric bias linked with biased inference procedures in driving the underestimation of women. This project is intended to involve three stages. I first analyze the mismeasurement that results from the widespread use of Mincer Residuals under existing biased inference procedures. Through failure to account for differential selection, all women are underestimated. Through failing to account for distributional differences in gender disparity, the typical result is that high-ability women are underestimated, and low-ability women are overestimated. While less frequent, failure to include gender dummies produces even more bias. I have proposed several methods that address the second channel and will discuss some methods that produce unbiased estimates but result in loss of information. I am in the process of identifying methods to reduce information loss while retaining unbiasedness. I am actively working on the second stage of the project. In this stage, I will run an experiment through MTurk in India. I identify the extent of biased inference across groups when people are told they are evaluating women and men and are provided with biased signals. I compare this to when people receive biased signals about individuals from two unnamed groups. In each setting, I identify the extent to which biased inference is ameliorated by information and by actively debiased information. In addition, I identify the influence of two influential psychological biases on biased inference and decision-making through randomized timing of information provision, salience shocks, and active debiasing. This also allows me to identify how biased econometric and statistical procedures can worsen persistently incorrect inference strategies employed in hiring decisions.

This seminar series is jointly organized by the IFS, STICERD, and UCL.

IFS/STICERD/UCL Development Economics Work In Progress seminars are held on Thursdays in term time at 13:00-14:00, at the IFS, unless specified otherwise.

Seminar organisers: Oriana Bandiera (STICERD, LSE), Imran Rasul (UCL), Britta Augsburg (IFS) and Jonathan Weigel (LSE).

For further information please contact Britta Augsburg: britta_a@ifs.org.uk.

Registration is required, via our mailing list:

Please use this link to subscribe or unsubscribe to the Development Economics Work In Progress seminars mailing list (developmentwip).