Dealing with data that is missing not at random
What happens when you have missing data? Many people recommend using multiple imputation (see Briggs et al. 2003; Faria et al. 2014; Gomes et al. 2013; Hunter et al., 2015; Hughes et al., 2016). This approach, however, generally assumes that data are missing at random. Missing at random could be interpreted as the probability that health…