Methods for Increasing the Sincerity of Respondents' Answers to Sensitive Questions

Surveys with sensitive items often face the problem of low data quality. Nevertheless, in the last several decades sensitive questions have been increasingly used in sociological research. Consequently, the bias arising from the sensitivity of such questions has become a vital methodological issue. To overcome this problem, researchers have developed special questioning methods. In the theoretical part of this research we have analyzed three main ways for increasing the sincerity of respondents. The focus of the empirical part are non-randomized response (NRR) techniques as one of the most actively developing method. These, however, have rarely been applied in empirical studies because there is insufficient information about their efficiency. Based on this, the aim of this research is to conduct a comparative analysis of ways to increase sincerity to identify the most effective technique. We have conducted a survey using the crosswise model as one of the most used techniques, the parallel model as one of the least used techniques, and direct questioning. We have surveyed 310 students on a subject of legal awareness of Moscow youth. The main hypothesis of this study is that the NRR models reduce the questions` sensitivity (it has been confirmed).