On the use of inclusive strategy when some participants fail to provide data on all studied variables

Behavioural research scientists have become increasingly aware of the importance of missing data methods. Including auxiliary variables in data analysis can increase the plausibility of meeting the missing at random assumption, leading to increased parameter estimation accuracy and a more trustworthy goodness-of-fit evaluation. This study addresses a missing data pattern typically mishandled by using listwise deletion. The missing data pattern echoes a common research scenario in which some participants fail to respond to all the studied variables but provide information on auxiliary variables. Researchers commonly delete these participants from further data analyses in practice. Using confirmatory factor analysis models, this study shows that including effective auxiliary variables to analyse data with this missing data pattern can substantially improve the estimation accuracy, particularly when auxiliary variables correlate with latent factors.