Recent Events and the Coding of Cross-National Indicators

Much research in political science relies on datasets produced by human coders. Many variables included these are not based observable facts but rather require a considerable level of judgment. This project studies the extent to which this judgment is affected availability bias and how it influences retrospective coding historic cases. The analysis uses coder-level data from V-Dem project, one few collecting releasing codings tagged with timestamps when they were produced. results show that recent dramatic events country just prior have small, visible impact coder ratings, primarily for those directly related observed events. magnitude effect, however, small. alleviates concerns prominent world politics around time significantly affect reliability cross-national indicators.