Here is the scenario: You are trying to predict some outcome, , and some of your predictors have missing data. Will doing a complete case analysis give you unbiased results? What additional information do you need before deciding? For some reason, when I tried to answer this question, my first instinct was to try to decide whether the data were missing at random, but it turns out, this might not be the right first question! Why? Complete case analysis will give us unbiased estimates even if t...
