Gender Decoder is a web application that helps users locate gender bias in a given job ad and analyzes the biases found in over 500 scientist and engineer job descriptions.

This application uses the Python package, genderdecoder by Richard Pope, which uses the analysis code from Gender Decoder by Kat Matfield, which is based on the paper, Evidence That Gendered Wording in Job Advertisements Exists and Sustains Gender Inequality by Danielle Gaucher, Justin Friesen, and Aaron C. Kay.

From Kat Matfield's Gender Decoder About:

Below are the full lists of words that Danielle Gaucher's, Justin Friesen's, and Aaron C. Kay's research considered masculine- and feminine-coded. This tool checks job adverts for the appearance of any of these words, then calculates the relative proportion of masculine-coded and feminine-coded words to reach an overall verdict on the gender-coding of the advert.

Some words have been reduced to a 'stem' to cover a range of noun, verb and adjective variants; for instance, "compet" covers "compete", "competitive" and "competition".

• Streamlit
• genderdecoder
• SerpApi