NCJ Number
65432
Journal
International Journal of the Addictions Volume: 13 Issue: 3 Dated: (1978) Pages: 359-367
Date Published
1978
Length
9 pages
Annotation
FOR DRUG-USE EPIDEMIOLOGISTS, THIS PAPER DESCRIBES AN EXPERIMENTAL STUDY WHICH EXAMINED THE ADVANTAGES OF EMPLOYING THE RANDOMIZED-RESPONSE TECHNIQUE TO OBTAIN ESTIMATES OF DRUG USE.
Abstract
THIS TECHNIQUE TAKES ADVANTAGE OF GREATER ANONYMITY IN ENCOURAGING RESPONDENTS TO ANSWER MORE HONESTLY. THE PROCEDURE PRESENTS A PERSON WITH TWO ALTERNATIVE QUESTIONS (ONE CONCERNING SENSITIVE BEHAVIOR, THE OTHER AN INNOCUOUS QUESTION). ESSENTIAL TO THE TECHNIQUE IS PRIOR KNOWLEDGE OF THE PROBABILITY OF SELECTING EITHER QUESTION FOR RESPONSE. IN THIS STUDY, DRUG-USE INFORMATION WAS OBTAINED FROM ABOUT 800 HIGH SCHOOL STUDENTS (AGED 15 TO 19) WHO RECEIVED RANDOMLY ASSIGNED SETS OF QUESTIONNAIRES. FOUR ALTERNATIVE METHODS OF INQUIRING ABOUT DRUG USE WERE USED: THE SIX-ITEM CONDITION, ONE-ITEM CONDITION, AND TWO SAMPLES OF THE RANDOMIZED-RESPONSE TECHNIQUE. THE DATA ANALYSIS INCORPORATED IN THIS STUDY IS EXPLAINED, NOTING THE EQUATIONS USED TO DETERMINE MEAN DRUG USE AND MEAN ENTERTAINMENT ACTIVITY, AS WELL AS DETERMINATION OF STANDARD ERRORS. FINDINGS REVEAL THAT A SIGNIFICANTLY LARGER PROPORTION OF PERSONS FAILED TO COMPLETE THE DRUG-USE QUESTION WHEN ASKED IN A TRADITIONAL, DIRECT FASHION (I.E., IN THE ONE-ITEM AND SIX-ITEM CONDITIONS) THAN WHEN THE RANDOMIZED RESPONSE PROCEDURE WAS USED. WITH ONE EXCEPTION (HALLUCINOGEN USE), MEAN REPORTED DRUG USE WAS SIGNIFICANTLY GREATER WHEN THE RANDOMIZED RESPONSE PROCEDURES WERE EMPLOYED THAN WHEN DIRECT QUESTIONING WAS USED. REFERENCES AND SOME TABULAR DATA ARE INCLUDED. (WJR)