Leah von der Heyde, Anna-Carolina Haensch, Alexander Wenz
Vox Populi, vox AI? Using language models to estimate German public opinion

SocArXiv preprint
64 S.

The recent development of large language models (LLMs) has spurred discussions about whether LLM-generated “synthetic samples” could complement or replace traditional surveys, considering their training data potentially reflects attitudes and behaviors prevalent in the population. A number of mostly US-based studies have prompted LLMs to mimic survey respondents, finding that the responses closely match the survey data. However, several contextual factors related to the relationship between the respective target population and LLM training data might affect the generalizability of such findings. In this study, we investigate the extent to which LLMs can estimate public opinion in Germany, using the example of vote choice as outcome of interest. To generate a synthetic sample of eligible voters in Germany, we create personas matching the individual characteristics of the 2017 German Longitudinal Election Study respondents. Prompting GPT-3 with each persona, we ask the LLM to predict each respondents’ vote choice in the 2017 German federal elections and compare these predictions to the survey-based estimates on the aggregate and subgroup levels. We find that GPT-3 does not predict citizens’ vote choice accurately, exhibiting a bias towards the Green and Left parties, and making better predictions for more “typical” voter subgroups. While the language model is able to capture broad-brush tendencies tied to partisanship, it tends to miss out on the multifaceted factors that sway individual voter choices. Furthermore, our results suggest that GPT-3 might not be reliable for estimating nuanced, subgroup-specific political attitudes. By examining the prediction of voting behavior using LLMs in a new context, our study contributes to the growing body of research about the conditions under which LLMs can be leveraged for studying public opinion. The findings point to disparities in opinion representation in LLMs and underscore the limitation of applying them for public opinion estimation without accounting for the biases in their training data.