This evaluation studied attitudes and opinions of respondents to the 2020 Census and assessed how privacy and confidentiality concerns affect response behavior in terms of mode and item nonresponse. It also analyzed whether respondents that could not be matched with administrative records have different attitudes. 

The study took a mixed-methods approach. It included a quantitative survey and a qualitative component. The 2020 Census Opinion Survey was designed to test privacy and confidentiality concerns overall and by demographic panels of interest. The qualitative component included interviews, focus groups, and event observations (e.g., community or public school events). It was designed to capture privacy and confidentiality concerns of people who were less likely to respond to the 2020 Census and less likely to respond to the 2020 Census Opinion Survey. Participants in this component of the evaluation allowed insight into the concerns of people who were less willing to respond to decennial censuses and act as proxies for those who are often missed in similar studies.

One of the major concerns going into the first decennial census that offered an online response mode to the majority of the nation was that respondents would avoid submitting their answers online because of privacy and confidentiality concerns. The 2020 Census Opinion Survey results yielded no strong link between privacy concerns and how members of the public chose to respond. Furthermore, there was no evidence that privacy concerns stopped people from responding online. However, respondents who provided their data to a 2020 Census enumerator differed from those who responded on their own in their overall views about the census and their beliefs about potential harm that could result from answering it.  

Privacy concerns were also not strongly related to mismatches between self-reported data and administrative records. However, most respondents wanted the Census Bureau to collect information directly from the public instead of relying on data from other government agencies. Nevertheless, a majority of respondents preferred the use of information from other government agencies over asking neighbors for household information. People less worried about the use of administrative records were more likely to leave at least one question blank on a paper questionnaire. Those who saw more benefits in the census were less likely to mail back a partially completed paper questionnaire. 

The results from this evaluation show areas to explore for 2030 Census research such as research into attitudes about the use of administrative data that provides added context to participants. Results also suggest possible strategies for communications campaigns to increase self-response. Examples include messaging about specific benefits of response in a particular area or community, and messaging tailored to address beliefs about harms of census response. This research not only provides insight into attitudes around the decennial census but also into attitudes about the use of administrative data.