Online Program Home
My Program

Abstract Details

Activity Number: 138 - Modeling Applications for Backcasting, Nowcasting and Forecasting
Type: Contributed
Date/Time: Monday, July 29, 2019 : 8:30 AM to 10:20 AM
Sponsor: Survey Research Methods Section
Abstract #304775
Title: Variable Selection for Multinomial Logistic Regression Modeling to Assign One of Six Census Mindsets Using Big Data
Author(s): Mary H. Mulry* and Yazmín A. García Trejo and Nancy Bates
Companies: U.S. Census Bureau and U.S. Census Bureau and U.S. Census Bureau
Keywords: 2020 Census; 2020 Census Communications Campaign; 2020 Census Barriers, Attitudes and Motivators Study; self-response

The U.S. Census Bureau is preparing to field the 2020 Census Communications Campaign to encourage participation in the 2020 Census. Similar campaigns aided in maintaining high self-response rates for the 2000 and 2010 Censuses. To prepare, the U.S. Census Bureau fielded the 2020 Census Barriers, Attitudes and Motivators Study (CBAMS) sample survey to collect data on attitudes and knowledge about the U.S. Census. Data from over 17,000 respondents was used to classify individuals into one of six psychographic profiles referred to as Census “mindsets”. In social marketing campaigns, mindsets are constructed to reflect an individual’s knowledge, attitudes and opinions toward a topic. The mindsets are then used in developing messages with a call to action. In our case, the requested action is a response to the 2020 Census. Our research examines the feasibility of assigning a mindset to each record in a Big Data file, which is a third-party dataset containing over 250 million adult records and ultimately to households. The 2020 CBAMS variables used in determining the mindsets are not present on the third-party dataset although the dataset does contain over 500 variables that reflect demographics, socioeconomic status, attitudes and behavior. Our approach links the 2020 CBAMS survey records to the third-party dataset and then uses multinomial logistic regression with independent variables from the third-party dataset to predict the probabilities of the mindsets.

Authors who are presenting talks have a * after their name.

Back to the full JSM 2019 program