Answers are grounded in this paper and cite their sources.
Online surveys are increasingly used to investigate health conditions, especially among young people. However, this methodology presents some limitations including item nonresponse concerning sensitive and knowledge-related questions. This study aimed to suggest latent class analysis as the methodology to statistically deal with item nonresponse data, while describing item nonresponse in an online health survey addressed to young people. We used data from an e-cohort study on 8,663 French university students and identified six homogenous groups of students with similar use of nonresponse options (‘I don’t know’, ‘I’m not sure’, or ‘I don’t want to answer’): one of High Respondents (64.2%), four of Partial Respondents (34.3%), and one of High Nonrespondents (1.5%). Item nonresponse largely depended on the survey’s contextual features (type of nonresponse options, question domains), and was significantly associated with (non)respondents’ individual characteristics (gender, age, field of study). We conclude that item nonresponse can be very informative and that latent class analysis is a useful method to identify its patterns in online health surveys.
No takes yet. Share an insight, caveat, or question.
Montagni et al. (2019) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: