This chapter treats how the model can handle three different ways in which data about tie variables may be absent. First, some actors might not be members of the network at some of the waves; second, information about whether there is a tie might be missing; third, the existence of a tie between two actors might be impossible at some or all of the waves – or it might be impossible for the tie not to exist. These three situations are conceptually totally different, although in practice there may be confusion between them. The first is regarded as change in composition of the network and modeled by a concept of “joiners” and “leavers.” The second is the classical case of “missing data.” For the third, the concept of “structural zeros” is used, and complemented by the concept of “structural ones.”