The thesis analyses the differences that result from using of specific data collection method, computer assisted telephone interviewing (CATI) and computer assisted personal interviewing (CAPI). The main findings are based on election model built by research agency TNS Aisa for elections to the Chamber of Deputies of the Parliament of the Czech Republic in 2013. The aim of the thesis is to determine whether the chosen method of data collection influences the results of the estimated electoral preferences, to determine in which sociodemographic categories it is happening and whether the differences in the obtained data are related to the ideological orientation of the political parties. The dependence of the results on the data collection method is assessed by the chi-quadrate independence test. Further, through personal interviews with researchers, it is determined at what stage of the research process the data may be distorted and what its possible causes are. The benefit of the thesis is the finding that the chosen method of collection influences the established preferences among voters aged 30-44, university graduates and voters living in Prague. In these groups, the left-hand side ČSSD is preferred more by personal interviewing and right-handed TOP 09 through telephone interviewing. The collection of data itself was evaluated by the most risky phase of the research process, in which possible distortion could occur. While the accuracy of personal interviewing depends largely on the interviewer's personality, the sources of distortion of the telephone inquiry result more from the nature of the use of the phone itself.