Inspecting Regression Trees of Subjective Burden Perception with Objective Burden Measures in a Household Survey

• Activity Number: 54 - Record Linkage, Data Integration, and Improving Survey Measurement
• Type: Contributed
• Date/Time: Sunday, August 8, 2021 : 3:30 PM to 5:20 PM
• Sponsor: Survey Research Methods Section
• Author(s): Daniel Yang*
• Companies: Bureau of Labor Statistics
• Keywords: Respondent burden; Survey design; Nonparametric; Regression Tree model

Abstract

Ambiguous responses to a questionnaire or refusals to continue participation in a survey could introduce bias and downgrade overall survey data quality. This reduction in data quality could potentially be induced by higher respondent subjective (perceived) burden. Hence, it is essential to study what could affect the respondent’s subjective burden when improving the design of a survey. In order to reduce respondent’s subjective burden along with its potential bias, interventions may be required during data collection to support data quality. Between April 2017 and March 2018, subjective burden data were collected from the Consumer Expenditure Surveys Quarterly Interview respondents. In this study, we construct a subjective burden outcome by using principal component technique, and investigate its association with objective burden measures, household demographics and other explanatory variables using the nonparametric recursive partitioning models under complex survey design.