To the Editor:
Self-reported measures are widely used in epidemiologic research; however, the validity of such reports may be questionable when studying sensitive topics. Self-administration of surveys removes interviewer effects and may reduce social desirability bias, but at the cost of potential subject confusion or data-entry errors. Audio computer-assisted self-interview (ACASI) systems provide for a high level of confidentiality, while removing literacy barriers.
The power structure of prisons combines with a history of abusive research involving prisoners1 to create an environment where pressure to misreport may be especially high. Rates of illiteracy among prisoners are high,2 limiting the utility of self-administered paper surveys. We examined the performance of a color-coded ACASI system for gathering sensitive data from male prisoners.
The participants completed a questionnaire regarding their adverse childhood experiences exposure as part of a survey of prisoner tobacco use.3 The Ohio State University Institutional Review Board and Ohio Department of Rehabilitation and Corrections approved all procedures, and all subjects gave informed consent to participate. The survey covered 7 categories, including physical, emotional, and sexual abuse.4 We collected data on a laptop computer-based, color-coded ACASI system, programmed using Visual Basic for Applications within Microsoft Access 2000 (Microsoft Corporation, Redmond, WA). The system displayed the question and responses onscreen while participants heard an audio recording of the materials over headphones. For example: “How often did a parent, step-parent, or adult living in your home swear at you, insult you, or put you down? If never, press red. If once or twice, press orange. If sometimes, press yellow. If often, press green. If very often, press blue. To skip this question, press black. To hear it again, press white.”
Participants entered their responses using a customized data entry device with 7 color-coded keys, built by modifying a number pad (Figure). The program had participants confirm each answer. For example, “You chose ‘never.’ If this is correct, press black. To change your answer, press white.”
The interviewer remained present during a sample question to adjust volume settings and answer any participant questions, but moved out of sight during survey administration. Following confirmation of the final response, the program instructed the participant to notify the interviewer.
The questionnaire was completed by 198 of 200 eligible participants (99%). Item response rates were high among those who received the questionnaire (range: 98%–100%; median = 100%). Complete responses were received from 191 men (96%); 2 men (1%) declined to answer 1 question but provided sufficient information to determine exposure to adverse childhood events for all categories, and 5 men (3%) declined to answer one or more questions in a manner that made it impossible to definitively determine their exposure.
Nearly all participants were willing and able to answer questions using the color-coded ACASI system, regardless of their prior experience with computers or comfort with reading. Intake data, collected by the Ohio Department of Rehabilitation and Correction from “self admissions, social and criminal history records,” identified physical and abuse in 8% of male inmates and sexual abuse in 5%.5 When using the ACASI system, participants in the current study reported about 3 times the prevalence of both physical (24%) and sexual (16%) abuse, possibly indicating a higher degree of comfort in sharing sensitive information when using the confidential system.
To promote dissemination of ACASI systems, the computer code used to create the system is available, for free, through the author's website for use or modification by other researchers (http://publichealthresearch.org/FreeCASI), as are instructions for building a color-coded keypad. Color-coded ACASI systems, such as the one described in the present study, offer an effective means of gathering sensitive data, even when participants have difficulty reading or using computers.
Development of the project was supported by the Behavioral Cooperative Oncology Group of the Mary Margaret Walther Program, Walther Cancer Institute, Indianapolis, IN.
Ross M. Kauffman
College of Public Health
The Ohio State University
Ryan D. Kauffman
Oakhill Medical Associates
West Liberty, OH
1. Kalmbach KC, Lyons PM. Ethical and legal standards for research in prisons. Behav Sci Law
2. Linacre JM. The prison literacy problem. Rasch Measurement Transactions
3. Kauffman RM, Ferketich AK, Murray DM, Bellair PE, Wewers ME. Measuring tobacco use in a prison population. Nicotine Tob Res
4. Felitti VJ, Anda RF, Nordenberg D, et al. Relationship of childhood abuse and household dysfunction to many of the leading causes of death in adults. The Adverse Childhood Experiences (ACE) Study. Am J Prev Med