Characterizing individual and methodological risk factors for survey non-completion using machine learning: findings from the U.S. Millennium Cohort Study.

2 , if we inspect the frequency of missingness by item location among those classified as completers versus non-completers, we can see that these groups successfully decompose missingness into MAR and MNAR, respectively; non-completers show an increasing trend as missingness aggregates from those who have quit the survey, whereas no trend is present among completers who instead exhibit increased missingness on a few well-distributed items (Fig. 3 ).

also in 63,971 other papers