Survey Transfer Learning: Recycling Data with Silicon Responses

TL;DR

This paper introduces Survey Transfer Learning (STL), a transfer learning approach that recycles existing survey data to generate accurate silicon responses, outperforming large language models in social science applications.

Contribution

STL applies transfer learning paradigms from computer science to survey research, enabling efficient and accurate generation of silicon responses using neural networks.

Findings

STL achieves up to 93% accuracy in silicon responses.

STL outperforms LLMs on sensitive measures like racial resentment.

STL reduces costs and opacity compared to LLM-based methods.

Abstract

As researchers increasingly turn to large language models (LLMs) to generate synthetic survey data, less attention has been paid to alternative AI paradigms given environmental costs of LLMs. This paper introduces Survey Transfer Learning (STL), which develops transfer learning paradigms from computer science for survey research to recycle existing survey data and generate empirically grounded silicon responses. Inspired by political behavior theory, STL leverages shared demographic variables with high predictive power in a polarized American context to transfer knowledge across surveys. Using a neural network pre-trained on the Cooperative Election Study (CES) 2020, freezing early layers to preserve learned structure, and fine-tuning top layers on the American National Election Studies (ANES) 2020, STL generates silicon responses CES 2022 and in held-out ANES 2020 data with accuracy rates of up to 93 percent. Results show that STL outperforms LLMs, especially on sensitive measures such as racial resentment. While LLMs silicon samples are costly and opaque, STL generates empirically grounded silicon responses with high individual-level accuracy, potentially helping to mitigate key challenges in social science and the polling industry.