The 2020 Census was accompanied by a diverse array of paid advertisements, partnership outreach programs, news coverage, and more designed to increase knowledge about the census and motivate households to self-respond, specifically through the internet mode.

This evaluation estimates the effect of these programs, collectively referred to as the decennial environment. By matching 2019 Census Test data to 2020 Census data, we compared self-response rates between the 2020 Census and the 2019 Census Test, which was void of advertisements.

This methodology allowed us to analyze self-response behavior with and without the decennial environment. Through comparisons, we determined if a significant percentage of households, by mode, did not respond in 2019 but responded in 2020, and whether respondents that used mail or Census Questionnaire Assistance (CQA) in 2019 switched to internet in 2020.

To understand if participating in the 2019 Census Test affected the likelihood of responding to the 2020 Census, a post-hoc sample was drawn from the unused portion of the American Community Survey (ACS) 2019 sample universe, and followed the same sampling procedures as the 2019 Census Test. This sample contained 240,000 households, which matched to 235,000 households in the 2020 Census sample. 

This report presents the results of this evaluation and evidence that the decennial environment encourage more people to respond online. It also outlines recommendations for the 2030 Census testing and advertising campaign strategies used to promote self-response for the 2030 Census.