Analysis of Incomplete Data Sets
Standard statistical procedures assume the availability of complete data sets. In sample surveys or censuses, some of the individuals may not respond to some or all items being asked. In such cases missing data may have a strong influence on the statistical analysis of the remaining data set. Rubin (1976, 1987) and Little and Rubin (1987) have discussed some concepts for handling missing data based on decision theory and models for mechanism of nonresponse.