Interaction of Factors Causing Missing Data at Random

Obtaining qualitative empirical information is very significant for any researcher. Missing are one of the barriers to high-quality reliable data. Many scientists have conducted sociological experiments to identify causes of respondent’s desire to leave the question unanswered. Nevertheless, all of them have studied one-dimensional impact of concrete factor on missing. An objective of this research is studying of multidimensional missing data at random and factors causing it in questions of relation to the political decisions made in the Russian Federation. Multidimensionality allows defining missing data at random which may not be such a one-dimensional consideration. Also multidimensional studying of missing data at random allows to find such a combination of factors, in which will show up the largest share of missing, compared with one-dimensional communication. In this research we propose an algorithm that allows detecting the combination of the characteristic values, leading to considerable changes in the share of non-responses. This algorithm helps not to miss major regularities and missing in categories of respondents, which are significant for research problems.