Missing data may occur in surveys, clinical trials, and even in planned experiments using appropriate design. Initially this was realized in experimental designs and Yates (1933) and Bartlett (1937a) provided methods for analyzing the data. In these problems the analysis is carried out using only the available data. Missing values result in having nonorthogonal setting and the analysis can easily be made using computer programs. Before computers were introduced, Yates had substituted estimates for missing data which made the residuals small. Bartlett provided covariance analysis by using covariates for the missing observations where the ith covariate takes 0 values for all responses except for the ith missing response and 1 or −1 was given for the ith missing observation. These methods allowed the researcher to get the analysis on the observed responses only using the available statistical methods.