How to Get Better Survey Data More Efficiently

A key challenge facing many large, in-person public opinion surveys is ensuring that enumerators follow fieldwork protocols. Implementing quality control processes can improve data quality and help ensure the representativeness of the final sample. Yet while public opinion researchers have demonstrated the utility of quality control procedures such as audio capture and geo-tracking, there is little research assessing the relative merits of such tools. In this paper, we present new evidence on this question using data from the 2016/17 wave of the AmericasBarometer study. Results from a large classification task demonstrate that a small set of automated and human-coded variables, available across popular survey platforms, can recover the final sample of interviews that results when a full suite of quality control procedures is implemented. Taken as a whole, our results indicate that implementing and automating just a few of the many quality control procedures available can streamline survey researchers' quality control processes while substantially improving the quality of their data.

Automated summary

The paper presents new evidence on the effectiveness of quality control procedures in public opinion surveys, using data from the AmericasBarometer study. The results show that automating a few quality control procedures can streamline researchers' processes while significantly improving data quality.