Intrastate applicability of National Census Response Rate Models: The case of the 2020 Census in the state of Illinois, United States. Issue 8 (21st August 2022)

• Record Type: Journal Article
• Authors: Shin, Jaeyong; Zandiatashbar, Ahoura; Massey, Dean; Rozhkov, Anton; Smith, Janet; Zellner, Moira
• Other Names: Ulceluse Magdalena guestEditor.; Bock Bettina guestEditor.; Haartsen Tialda guestEditor.
• Is Part Of: Population space and place. Volume 28:Issue 8(2022)
• Page Start: n/a
• Page End: n/a
• Publication Date: 2022-08-21
• Subjects: 2020 U.S. Census -- geographic difference -- hard‐to‐count -- self‐response rates; Population geography -- Periodicals; Population -- Periodicals; 304.6
• DOI: 10.1002/psp.2608
• Languages: English
• ISSNs: 1544-8444

Abstract

Abstract: For the 2020 Census, the U.S. Census Bureau developed the Low Response Score (LRS) to predict census self‐response participation. Based on a previous hard‐to‐count metric, which accounts for people who are difficult to survey or count, the LRS was undoubtedly an improvement over previous census effort. However, census response behaviour can be sensitive to inter‐ and intrastate locational and socioeconomic differences; in other words, geographic differences should be accounted for prediction. In this study, we strive to assess the assumption that geographic difference influences census participation and is one of the key factors in predictive models. The framework presented in this study is composed of three steps: (i) detecting the geographic differences in census participation; (ii) building a statistical model to estimate impact of geographic difference on response rates; and (iii) comparing the model results with actual final self‐response rates of the 2020 Census to suggest policy implications. Our findings confirmed that geographic differences in census participation matter, independent of socioeconomic and demographic characteristics. The results show that the combined impacts of sociodemographic and geographic factors are important in estimating census participation. Our findings also show high accuracy in predicting census responses with a parsimonious model that includes variables for geographic differences. The results suggest a localized approach with consideration of the geographic differences, as well as different populations and neighbourhoods within a state to tailor outreach campaigns for greater effectiveness and impact.