2020 Time Series: Updates & Errata

Current Dataset Release: February 10, 2022 version

February 10, 2022
The February 10, 2022 version of the ANES 2020 Time Series study includes corrected variables from the MEDSRC submodule. The variables that have been updated include V201630a-V201632s (TVPROGS). All other variables in the MEDSRC submodule were reviewed and no additional errors were found. No other changes have been made to the data.
January 20, 2022
We have become aware of an issue involving a subset of variables from the MEDSRC submodule. The randomization was incorrectly applied to the TVPROGS variables ( V201630a to V201632s). We are currently revising this set of variables and will re-release the 2020 Time Series data in the coming weeks. We are also reviewing the following variables from the MEDSRC submodule: V201633a-V201633r (RADPROGS) and V201634a-V201634s (WEBSITES). Thanks to L.J. Zigerell of Illinois State University for helping us to identify this issue.
July 19, 2021
This is a full release of the ANES 2020 Time Series study dataset and the third release overall. It contains both the pre-election and post-election study. The previous two releases were preliminary and did not contain some of the variables, processing, cleaning, and documentation that are a part of this current release. In particular, this release includes the following additions or changes to the data file: updated weights, additional variables (state, district, group consciousness, group identity, screener information, candidate name/gender/party, randomizations, interviewer observations, interviewer characteristics, etc.), additional and updated summary variables (including vote summaries), coding of Other Specify responses, data corrections, improved documentation and metadata, and removal of four post-election cases. As a result of these and other updates, users are recommended to check their preliminary release results against this current release, to assess whether the changes have any effect on their analyses.
April 22, 2021
In the processing of the 2020 Time Series post-election data, a potential issue has been detected involving the variables V202498x – V202504. For that reason, most of the variables from this submodule (IDENT) were coded to -2 with our intent being to research the issue and include those variables in the full release of the 2020 Time Series data. However, it was pointed out that the variable V202504 (How important is being American to your identity?) was not included in that set of recodes and contains valid data. The data, somewhat predictably, look off. We are in the process of trying to fully resolve the issues with all of the variables from this submodule and will re-release them as soon as possible.