Examining the accuracy of students’ self-reported academic grades from a correlational and a discrepancy perspective: Evidence from a longitudinal study

Missing data in medical research is a common problem that has long been recognised by statisticians and medical researchers alike. In general, if the effect of missing data is not taken into account the results of the statistical analyses will be biased and the amount of variability in the data will not be correctly estimated. There are three main types of missing data pattern: Missing Completely At Random (MCAR), Missing At Random (MAR) and Not Missing At Random (NMAR). The type of missing data that a researcher has in their dataset determines the appropriate method to use in handling the missing data before a formal statistical analysis begins. The aim of this practice note is to describe these patterns of missing data and how they can occur, as well describing the methods of handling them. Simple and more complex methods are described, including the advantages and disadvantages of each method as well as their availability in routine software. It is good practice to perform a sensitivity analysis employing different missing data techniques in order to assess the robustness of the conclusions drawn from each approach.