Meta-analysis of Randomized Response Research: 35 Years of Validation

The Randomized Response Technique (RRT) is a survey method especially developed to improve the accuracy of answers to questions on threatening topics. This paper reports the results of two meta-analyses on RRT studies, the first on the results of six individual validation studies, and the second on 31 comparative studies. The focus of these metaanalyses was on the performance of RRTs compared to conventional question-answer methods. A multilevel analysis with three levels (studies, conditions, and sensitive items) was used to compare the effects of data collection methods across different studies. As effect size for the individual validation studies we used the percentage of incorrect answers and as effect size for the comparative studies we used the standardized difference. Results indicated an overall positive effect for RRT across studies compared to the other methods. This positive effect is stronger as the topic under investigation becomes more socially sensitive by nature. This makes it opportune to use RRT when sensitive topics are studied. However, both metaanalyses showed a large variance component between studies. We interpret this as an indication that RRT is not under complete control of the researchers using it, and advise careful pretesting of the chosen RRT procedure. meta-analysis of randomized response research 3