Skip to main content

Measuring the price of hidden biases

Illustration of a man standing next to a height chart
Timothy Cook

Do hidden biases about appearance affect income? In recent research published in PLOS ONE, UVA associate professor of data science Stephen Baek and economist Suyong Song of the University of Iowa drew from a large database that included a wealth of demographic information, including family income, along with full 3D body scans. More comprehensive than simple measurements of height or weight, the scans were also more accurate than self-reported measures often relied upon in previous research. (In fact, one of the researchers’ findings was a “prominent discrepancy” between the scans and people’s self-reported height and other body measures; both men and women tended to report themselves as taller than their actual height, and heavier people underreported their weight while lighter people overreported it.) Controlling for family income affecting weight, the researchers found that men had about $1,000 more per year in family income per one centimeter increase in height, while women had a $600-$700 decrease in family income for every one unit increase in body mass index (BMI), says Baek. Interestingly, researchers did not find a similar relationship between weight and income for men, suggesting a stronger weight bias against women.  The researchers hope that this work will provide further evidence that such biases exist, says Baek, “and that we need to make changes to improve the situation.”