Predicting individual-level income from Facebook profiles

Abstract

Information about a person's income can be useful in several business-related contexts, such as personalized advertising or salary negotiations. However, many people consider this information private and are reluctant to share it. In this paper, we show that income is predictable from the digital footprints people leave on Facebook. Applying an established machine learning method to an income-representative sample of 2,623 U.S. Americans, we found that (i) Facebook Likes and Status Updates alone predicted a person's income with an accuracy of up to r = 0.43, and (ii) Facebook Likes and Status Updates added incremental predictive power above and beyond a range of socio-demographic variables (ΔR2 = 6-16%, with a correlation of up to r = 0.49). Our findings highlight both opportunities for businesses and legitimate privacy concerns that such prediction models pose to individuals and society when applied without individual consent.

Fig 1. Density distributions of annual income.

The distribution of the study sample is displayed in red, the distribution of the US Census data in blue.

Fig 2. Pearson Product-Moment correlations between predicted and actual income values.

Red bars indicate the predictive power of the socio-demographic variables used as baseline comparisons. The ‘Demographics” model includes age, gender, and ethnicity. Blue bars indicate the predictive accuracy of Facebook data, separated by Likes, Facebook Status updates, and a combination of the two. The purple bars display the results of the comprehensive models, which include both socio-demographic variables, personality and Facebook data.

Fig 3. Low and high income word clouds.

Words and phrases most positively correlated with income (top) and most negatively correlated with income (bottom), after controlling for age and gender.