Abstract We tested the feasibility and performance of the Interactive Voice Response Technology (IVR) in the assessment of sexual behavior self-reports, rel- ative to self-administered questionnaire (SAQ) and Timeline Followback (TLFB) methods. The sample consisted of 44 sexually active Hispanic students recruited at the University of Texas at El Paso who reported daily about sexual behaviors and substance use. Thirty-three participants (75%, 18 women, 15 men) were retained for at least 80 days of the 91-day IVR. At follow-up, sexual behaviors and substance use were assessed by questionnaire (summary) reports and by TLFB, referring to the same 3-month interval. ANOVAs with normalized variables indicated less reporting in the TLFB and over-reporting of substance use in the questionnaire relative to the daily IVR self- reports. Gender moderated the effects of assessment mode, which were observed among women only. HLM analyses indicated a significant decrease in self-reports over time, suggesting reactivity of self-monitoring via IVR on behavior. Keywords Sexual risk behavior Substance use Gender TLFB IVR SAQ Introduction Most sexual risk behavior research is based on retrospective self-report, which can yield substantial measurement error due to inaccurate recall, motiva- tional bias towards social desirability, and rounding of behavioral counts (Catania, Gibson, Chitwood, & Coates, 1990; Schroder, Carey, & Vanable, 2003; Weinhardt, Forsyth, Carey, Jaworski, & Durant, 1998b). Much research has focused on assessment methods and conditions facilitating accurate self- report of socially sensitive behavior. Comparisons between different assessment modes, such as face- to-face interview, self-administered questionnaire (SAQ), and computerized assessment methods, have revealed significant intra-individual variations in sexual behavior self-reports (Catania et al., 1990; Schroder et al., 2003). Unfortunately, assessment mode comparisons are of limited value without a ‘‘gold standard of comparison’’ that could be used to determine the method that yields the smallest measurement error (Catania et al., 1990). Many researchers rely on daily self-reports as the best approximation to an objective standard of comparison, reasoning that diary methods can greatly reduce memory error as one major source of inaccuracy rela- tive to retrospective self-reports (Graham, Catania, Brand, Duong, & Canchola, 2003). What facilitates honest, unbiased reporting? Recent research and literature reviews indicate that anonym- ity, privacy, and self-administration reduce socially desirable responding and enhance self-reports of sex- ual risk behavior (Durant, Carey, & Schroder, 2002). For example, a common finding in the literature is the relative underreporting of socially sensitive behaviors K. E. E. Schroder (&) Department of Psychology, Utah State University, 2810 Old Main Hill, Logan, Utah 84322-2810, USA e-mail: kerstin.schroder@usu.edu C. J. Johnson J. S. Wiebe Department of Psychology, University of Texas at El Paso, El Paso, Texas, USA AIDS Behav (2007) 11:313–323 DOI 10.1007/s10461-006-9145-z 123 ORIGINAL PAPER Interactive Voice Response Technology Applied to Sexual Behavior Self-reports: A Comparison of Three Methods Kerstin E. E. Schroder Christopher J. Johnson John S. Wiebe Published online: 14 July 2006 Ó Springer Science+Business Media, Inc. 2006