Randomized response is a research method to measure the behavior, opinions or attitudes of respondents towards sensitive topics, like breaking the law, the use of drugs and alcohol or sexuality. In research on these kind of topics the respondent often have a tendency to give desirable answers and randomized response is a research method that is especially designed to obtain more valid answers.

In randomized response research the tendency towards social desirability is minimized by having the respondents answer according to some chance mechanism. An example is of such a mecanism the dice method. The respondent throws two dice and answers the questions according to the total that is thrown with both dice. Is the total 2, 3 or 4 the respondent is forced to answer ‘yes’, and with a total of 11 or 12 the respondent is forced to answer ‘no’. In all other cases the respondent is required to answer truthfully. Since only the respondent can see the dice, nobody knows the real answer and is the anonimity of the respondent protected.

Data collected with randomized response contain errors that result from the chance mechanism, because in some cases the respondent is forced to answer untruthfully. Statistical methods exist to correct for this kind of errors and obtain an unbiased estimate of the prevalence of the sentive behavior or attitude. However, for the results to be reliable larger sample are needed than for instance with direct questioning.

This web site offers an introduction of different randomized-responsemethods in terms data collection and data analysis. Underneath these introductions downloads of relevant literature are offered. The section Research also offers downloads of reports and data sets of randomized-response research that has been conducted in the Netherlands.

To start you there is an interactive example of a randomized-response survey you can fill out. In this survey you will be asked four sensitive questions on an imaginary subject which you will have to answer according to the throw of virtual dice. When finished you will be asked for your opninion of the randomized-response method.

The following introductions on randomised response can be downloaded:

Lensvelt-Mulders, G., Hox, J. en van der Heijden, P. Meta-analysis of Randomized Response Research: 35 years of validation
Description: comparison randomized response and direct questions.

Lensvelt-Mulders, G. (2000).  A comparison of randomized response, computer-assisted self-interview, and face-to-face direct questioning.Sociological Methods & Research, 28, 505-537.
Description: comparison randomized response and other survey methods.
Abstract, full text can be ordered at: Sage Publications