Measuring public opinion via digital footprints

Abstract

Do digital traces accurately reflect individual preferences? Can signals from social media be used to measure public opinion? This paper provides evidence in favour of these hypotheses. We test a regression and post-stratification strategy that combines samples of digital traces with a stratification frame containing individual-level socio-economic data, in order to generate area forecasts of the outcome social phenomena of interest. In our example, we forecast the two-party vote of Democrats and Republicans in the 2018 Texas congressional district and Senate election. Our implementation assumes we can observe, and sample, individuals signaling their preference by favoring one virtual location over another; in our case, visiting Democrat versus Republican Facebook pages during the election campaign. Over the course of seven weeks preceding the mid-term elections we generate vote share forecasts which do not use any traditional survey data as input. Our results indicate that individuals leave digital traces that reflect their preferences.