Missing data mechanisms
In order to get more insight in the missing data problem, the reasons for missing data are differentiated in several missing data mechanisms. These mechanisms describe the underlying cause of missing data and were first described by Rubin (1976). Rubin distinguished three missing data mechanisms: missing not at random (MNAR), missing at random (MAR), and missing completely at random (MCAR). These missing data mechanisms are important since they are assumptions for the missing data methods. It must be noted that the missing data mechanisms are not characteristics of an entire data set, but the mechanisms are merely assumptions which apply to specific analyses methods.
The button below links to an application that demonstrates the missing data mechanisms with two variabeles.