Statistical Analysis With Missing Data In Logistic Model

Missing data is a common problem of the modern social survey and research work. The record contains missing data is also called incomplete data. Missing data or incomplete data has a great influence on the results of the research. For instance, cause unrealizable results, make the analysis much more difficult decrease the statistical program. Traditional techniques of replacing missing data may have some limitations. With the development of modern technology, it is possible to use more advanced methods to handle missing data.This paper introduces the methods of replacing missing data. Also, this paper compares the efficacy of three methods. The study contains seven levels of missing data (5%,10%,15%,20%,30%,40% and 50%) and comparisons the efficacy of expectation maximization, regression imputation and mean substitution.