The Mythical Swing Voter

TL;DR

This paper demonstrates that with proper statistical adjustments, non-representative polling can accurately predict election outcomes and reveal that apparent public opinion swings are often due to sampling biases rather than real shifts.

Contribution

It introduces the use of multilevel regression and post-stratification (MRP) on non-representative data to produce reliable election predictions and insights into polling biases.

Findings

Non-representative polling can be accurate with statistical adjustments.

Reported swings in polls often result from sampling bias, not actual opinion shifts.

Non-representative data can reveal new political insights.

Abstract

The only acceptable form of polling in the multi-billion dollar survey research field utilizes representative samples. We argue that with proper statistical adjustment, non-representative polling can provide accurate predictions, and often in a much more timely and cost-effective fashion. We demonstrate this by applying multilevel regression and post-stratification (MRP) to a 2012 election survey on the Xbox gaming platform. Not only do the transformed top-line projections from this data closely trend standard indicators, but we use the unique nature of the data's size and panel to answer a meaningful political puzzle. We find that reported swings in public opinion polls are generally not due to actual shifts in vote intention, but rather are the result of temporary periods of relatively low response rates among supporters of the reportedly slumping candidate. This work shows great promise for using non-representative polling to measure public opinion and the first product of this new polling technique raises the possibility that decades of large, reported swings in public opinion-including the perennial "convention bounce"-are mostly artifacts of sampling bias.