ABSTRACT
Missing data There are numerous reasons why data might be missing: you failed to collect any data from some people you wanted to include but they weren’t available, someone failed to complete a question in a survey, a person dropped out of a longitudinal study or a person dropped out of one phase in a longitudinal study but reappeared later. Missing data are commonly seen to fall into three types, after a taxonomy which is usually attributed to Rubin (1976): missing completely at random (MCAR), missing at random (MAR) and missing not at random (MNAR, sometimes shown as NMAR).
