Audio Computer-Assisted Self-Interviewing (ACASI)

Audio computer-assisted self-interviewing (ACASI) is a methodology for collecting data that incorporates a recorded voice into a traditional computer-assisted self-interview (CASI). Respondents participating in an ACASI survey read questions on a computer screen and hear the text of the questions read to them through headphones. They then enter their answers directly into the computer either by using the keyboard or a touch screen, depending on the specific hardware used. While an interviewer is present during the interview, she or he does not know how the respondent answers the survey questions, or even which questions the respondent is being asked.

Typically the ACASI methodology is incorporated into a longer computer-assisted personal interview (CAPI). In these situations, an interviewer may begin the face-to-face interview by asking questions and recording the respondent's answers into the computer herself or himself. Then in preparation for the ACASI questions, the interviewer will show the respondent how to use the computer to enter his or her own answers. This training may consist solely of the interviewer providing verbal instructions and pointing to various features of the computer but could also include a set of practice questions that the respondent completes prior to beginning to answer the actual survey questions. Once the respondent is ready to begin answering the survey questions, the interviewer moves to a place where she or he can no longer see the computer screen but where she or he will still be able to answer questions or notice if the respondent appears to be having difficulties and to offer assistance as needed.

ACASI offers all the benefits of CASI, most notably: (a) the opportunity for a respondent to input her or his answers directly into a computer without having to speak them aloud to the interviewer (or risk having them overheard by someone else nearby); (b) the ability to present the questions in a standardized order across all respondents; (c) the ability to incorporate far more complex skip routing and question customization than is possible for a paper-based self-administered questionnaire; and (d) the opportunity to eliminate questions left blank, inconsistent responses, and out-of-range responses. In addition, the audio component allows semi-literate or fully illiterate respondents to participate in the interview with all of the same privacy protections afforded to literate respondents. This is significant, because historically, in self-administered surveys it was not uncommon for individuals who could not read to either be excluded from participation in the study altogether or to be included but interviewed in a traditional interviewer-administered manner, resulting in the potential for significant mode effects.

Evidence from several large-scale field experiments suggests the ACASI methodology reduces socially desirable responding compared to both interviewer-administered and solely text-based self-administration methods for sensitive topics, including use of illicit drugs, sexual behaviors, and abortion. ACASI also allows for increased standardization in the presentation of the survey questions because a pre-recorded voice is utilized to administer the survey questions. As a result, each respondent hears all introductory text, questions, and response categories read in exactly the same way. Thus, the natural variation caused by differences in interviewers' reading skills, pace, and/or vocal quality is eliminated.

...