Self-reports of alcohol use have widespread application because of other satisfactory and practical measures are generally unavailable. Research indicates that validity is influenced by a number of factors. For example, socially desirable response pressures may cause respondent to underreport their actual level of alcohol use. The quantitativeness of alcohol use response options may affect validity, but the available research is inconclusive. Studies in other topic areas suggest that respondents may be more likely to endorse the first response option presented them, but whether this effect generalizes to self-reported alcohol use is unknown. It has been suggested that response bias can be minimized by facilitating the retrieval process, but the precise details and best methods of facilitation have not been established. We propose to develop an interactive, computerized assessment method for investigating self-reported alcohol use and employ this method to study and improve the validity of self-reports. The program will be administered by microcomputer to 1240 respondents in driving under the influence treatment programs to study factors affecting self-reported alcohol use. A pretest- posttest design will be used to manipulate four factors: (1) order of frequency of alcohol use items relative to alcohol diagnostic items; (2) quantitativeness of response options (low,k Medium, High) for frequency of alcohol use items; (3) order of presentation of close-ended response options, and (4) feedback given to respondents. The feedback intervention has four different levels. In all levels, respondents will respond at Time 1 and be given a second opportunity to provide answers to the alcohol use items at Time 2. Under the second and third levels, respondents will be shown their original responses before responding to Time 2. Under the fourth level, respondents will receive information about the inconsistencies of their answers from Time 1 before responding at Time 2. We will compare self-reports with collateral reports. Analyses of these data will contribute to an understanding of the effect on self-reported alcohol use of (1) manipulating whether or not specific questions about alcohol use precede diagnostic items, (2) varying the quantitativeness of response options, (3) varying the order of presenting response options, (4) showing respondents their original answers and giving them a second opportunity to respond, and (5) providing feedback about inconsistencies and giving respondents an opportunity to resolve them.

Agency
National Institute of Health (NIH)
Institute
National Institute on Alcohol Abuse and Alcoholism (NIAAA)
Type
Research Project (R01)
Project #
5R01AA007852-02
Application #
3111819
Study Section
Clinical and Treatment Subcommittee (ALCP)
Project Start
1990-07-01
Project End
1993-06-30
Budget Start
1991-07-01
Budget End
1992-06-30
Support Year
2
Fiscal Year
1991
Total Cost
Indirect Cost
Name
Rand Corporation
Department
Type
DUNS #
City
Santa Monica
State
CA
Country
United States
Zip Code
90401