Enhancing the efficiency of randomized response techniques using direct responses

The quality of the findings of a research study largely depends on data it is based on. Many research studies require data collection from a sample of human participants. Respondents’ refusals and untruthful responses are common issues in surveys based on human participants. Randomized response survey methods ensure the respondents’ privacy protection, motivating them for participation in the survey. The existing scrambling methods do not offer the respondents to report the direct response. In practice, survey statisticians face situations where some of the survey participants may be willing to report their direct responses, thus avoiding the complex process of scrambling the responses. Moreover, some recent models involve a complicated scrambling process which may put a burden on the respondents which makes it difficult to practically implement such models in sample surveys. We propose two novel randomized response techniques using both the direct and scrambling response options. The proposed techniques are found to achieve a significant boost in efficiency over the competitor techniques. The efficiency conditions have been derived and are found mathematically strong. The results suggest that the proposed techniques are suitable for implementation in practical sample surveys.