The effect of imputation of missing values on selected statistical analyses

This bachelor’s thesis concerns the problem of missing values in data sets, not only in termsof the reasons and mechanisms that may lead to missing values but mainly in terms ofthe methods that can be used to deal with them. The aim of the thesis is to describe theaforementioned methods and examine the effect they have on selected statistical analyses. Thethesis is divided into theoretical and practical part. The theoretical part firstly deals with thereasons behind missing values and introduces missing data types and mechanisms. The mainsubject of the theoretical part is a detailed description of eight selected methods for handlingmissing data. These methods are divided into two main groups – the deletion methods andthe imputation methods. Among the imputation methods, we can also distinguish betweendeterministic and stochastic imputation methods. A brief introduction to the methods andtools of exploratory data analysis and cluster analysis, which are used in the practical part, isalso a part of the theoretical part of this thesis. The practical part demonstrates the effect ofdeleting or imputing missing values in a real data set – with the use of all methods describedin the theoretical part – on the outputs of exploratory data analysis and cluster analysis. TheR programming language is used in the practical part of this thesis.