Authors
Abstract
This paper investigates the impacts of follow-ups of web non-respondents with CATI (Computer Assisted Telephone Interviewing) on cross-sectional (all respondents to the survey wave) and longitudinal (only respondents to the wave and all waves previous) dataset quality, and of refinements to the survey design on datasets. The analysis uses data from the UK Household Longitudinal Study (UKHLS) COVID-19 Study. The impacts of follow-ups of non-respondents on response rates, representativeness and weighted dataset quality are quantified, along with those of not web sampling non-regular internet users and instead expanding use of CATI. Implications of findings for survey practice are then discussed.