Indmedica Home | About Indmedica | Medical Jobs | Advertise On Indmedica
Search Indmedica Web
Indmedica - India's premier medical portal

Indian Journal of Community Medicine

CME: Randomized Response Technique - An Innovative Method to Measure Culturally Sensitive Variables: Result from a Piliot Study

Author(s): M.B. Soudarssanane, Balaji Naik, Ajit Sahai, Joy Bazroy

Vol. 28, No. 3 (2003-07 - 2003-09)

Deptt. of Preventive and Social Medicine,
Jawaharlal Institute of Postgraduate Medical Education & Research (JIPMER),
Pondicherry - 605 001 Abstract:

Research question: What is the advantage of Randomized Response Technique (RRT) over the conventional Direct Interview (DI) and Anonymous Questionnaire (AQ) in the assessment of culturally sensitive variables?

Objectives: To compare the efficacy of the three methods, namely RRT, DI and AQ in the measurement of prevalence of Pre/Extra marital sex.

Study design: Cross sectional study, using the three methods.

Setting: A pilot study in a given community in Pondicherry.

Statistical analysis: Probability equations.

Results: The prevalence of pre/extra marital sex in the study population by the DI, AQ and RRT methods were 0%, 6% and 10% respectively in this pilot study.

Conclusion: RRT improves validity of measurement of culturally sensitive variables both by ensuring a high participation in the study and by enabling a true response by assuring full confidentiality of information.

Key Words: Measurement of culturally sensitive variable, Direct interview method, Anonymous questionnaire, Randomized response technique

Introduction:

Measurement of culturally sensitive variables in the field of community health research is an important but difficult task. Information on such culturally sensitive parameters like (i) social aspects related to health, for example, pre or extra marital sex, homosexuality etc. (ii) RCH related information like induced abortion (iii) medical jurisprudence related information like child abuse, child marriage, child labour, female infanticide, juvenile delinquency etc, are some important examples - especially in the changing scenario with new emerging diseases and socio-cultural requirements. The established conventional and routine methods like Direct Interview and Anonymous Questionnaire have their own limitation in such epidemiological measurements. The main challenge is to overcome the twin problems of non-response as well as false response. The effort of the researcher has to be strengthened so as to enthuse the respondent with as much encouragement for active participation and as much built-in checks to ensure absolute anonymity that will give full confidence and avoid false response. The internal validity of such studies should thus be strengthened both quantitatively (by ensuring a high response rate) and qualitatively (by ensuring true responses). The Randomized Response Technique (RRT) offers such a strengthening in the measurement of culturally sensitive variables by using simple probability equations. This study was done to validate this innovative RRT vis-a-vis the earlier mentioned conventional methods in the measurement of pre/extra marital sex in a given community.

Material and Methods:

The three methods used in this study were (1) the direct interview method - wherein the respondent was asked frankly if he or she has ever had pre or extra marital sex (2) the anonymous questionnaire method - wherein the respondents were handed over a written questionnaire and requested to return their responses without mentioning their names and (3) the randomized response technique - this technique, by way of an open interview with the respondent, utilizes the simultaneous administering of two unrelated questions - written in two separate cards, card A and card B - shown to the respondent. These questions (in this study) were - Card A : Were you born in an even year? and Card B: Have you ever had pre/extra marital sex? The questions were formatted in such a way that they addressed discrete variables so that the responses were 'either Yes or No' to either of the questions. The respondent tossed a coin and without disclosing to the interviewer the outcome of the toss, he/she answered Card A if the toss showed 'heads' and Card B if the toss showed 'tails'. Hence only the response, namely, 'Yes or No' was known to the interviewer without the knowledge of which question was being answered.

Methods of analysis:

The analysis is based on simple probability equations in two steps.

Step l: An equation is obtained that relates the total (that is, the overall) probability of a 'Yes' answer to probability of a 'Yes answer if Card B was answered'.

Pr(Yes) = Pr(Yes/A).Pr(A) + Pr(Yes/B).Pr(B)
= 0.5 x 0.5 + Pr(YesB)(0.5)
= 0.25 + 0.5 Pr(YesB )

Step 2: The total response in percentage is applied to the equation derived in Step 1 to arrive at the percentage (prevalence) of pre/extra marital sex in the community studied.

Pr(Yes) = (X)

Pr (Yes/B) = [Pr(X) - (0.25) ] / 0.5

Results:

In this pilot study, a total of only 6 interviews could be done by the direct interview method (because many people refused to take part in the study by this method), 45 by the anonymous questionnaire method, and 62 by the RRT method.

Of the six respondents interviewed directly, none of them wanted even to respond and expressed frankly their inability/unwillingness to participate. On further request, they responded "no" to the question if they ever had pre/extra marital sex. The prevalence was, therefore, "0" by this method. Out of the 45 respondents to whom the anonymous questionnaire was administered (the same question), only 16 responded (response rate 35.5%). In this study, out of these 16 forms, one contained an affirmative response. Thus the prevalence in the study community of pre/extra marital sex was 6% (1/16) as per the anonymous questionnaire method.

Of the 62 respondents approached for the RRT method, one declined to participate, giving a response rate of 98.4%. Of the remaining 61 respondents, 18(30%) gave the 'Yes' response.

Using the probability equation in step 2 mentioned earlier,

Pr(Yes) = (0.30)

Pr (Yes/B) = [Pr(0.30) - (0.25) ] /0.5 = 10%

It is seen that the prevalence of 10% for pre/extra marital sex as measured by the RRT proves both the validity and, therefore, the reliability of the randomized response technique over the conventional methods in the measurement of culturally sensitive variables. Both the higher response (98.4%) and the better results (10%) are highlighted. Thus the internal validity of the RRT is strengthened.

Discussion:

Direct face-to-face interviews pertaining to information on highly sensitive variables are very embarrassing to the respondents. The tendency, therefore, is to decline to take part in the study leading to a high non-response rate. The disadvantage is too obvious. The second method, namely the anonymous questionnaire is, of course, a better modification of the direct interview; however, this method also suffers from the usual and well-known high percentage of non-response (64.5% in this study). In addition, there is always a lurking doubt in the respondent that somehow the identity might be traced and hence makes him/her insecure. Therefore, the reliability of the information is poor.

Comparatively the randomized response technique assures a very high percentage of participation (98.4% in this study). It is also a very useful method to measure highly sensitive cultural variables ensuring full anonymity to the respondent and fool-proof security of information. In addition, the participation is made interesting for the respondent because of the element of curiosity of the toss results.

Theoretically, though some amount of a deliberate false response is possible in the third method - either regarding a prank in answering Card A or Card B (since the result of the toss is not known to the interviewer) or by way of a false answer 'Yes/No' to whichever card is answered, these are definitely overcome to a large extent by (1) the built-in curiosity in the methodology which ensures participation and (2) the security afforded by the method in the non-identifiability of the respondent in such culturally sensitive variables. In addition, we have to acknowledge the human nature that if at all some prank is sought to be attempted, it would be the same for all the three methods and hence the superiority of the third method (RRT) in surpassing the insecurity in responding and hesitation in participating.

Conclusions:

The innovative Randomized Response Technique has been demonstrated to be a reliable and valid methodology in the measurement of culturally sensitive variables as compared to the conventional methods of Direct Interview and Anonymous Questionnaire method.

Acknowledgements:

The first author, Dr. MB Soudarssanane, would like to place on record his hearty thanks to Dr. Mathew Knewman who taught him Biostatistics at the University of Western Australia, Perth, Western Australia. The authors also place their sincere thanks to all the participants in all the three methods of the study.

Access free medical resources from Wiley-Blackwell now!

About Indmedica - Conditions of Usage - Advertise On Indmedica - Contact Us

Copyright © 2005 Indmedica