Thesis (Ph.D.) - Cairo University - Institute of Statistical Studies and Research - Department of Mathematical Statistics

Missing data is a very common problem in most empirical research areas. Missing data is present if the researcher fails to get the information from the sample. The problem of missing data in survey sampling is called the problem of non-response. It causes possible bias in estimators of population characteristics if no treatment is made to compensate for the non-response. These biases occur when the non-respondents are different from the respondents with respect to the study variable. Also, these biases are difficult to eliminate since the precise reasons for non- response are usually undetermined. Moreover, the efficiency of parameter estimation will be reduced and the results can be misleading which violated the statistical inference about population parameters. Imputation is commonly used to treat item non-response. Imputation aimed to replace the missing values with a plausible value to get a complete data set valid for inference. Mean method of imputation, ratio method of imputation, product method of imputation and regression method of imputation are most commonly used methods of imputation

Issued also as CD