Computerized behavioral observation: Implications for improving accuracy and reducing bias

Charles Edward Sicotte, University of Rhode Island


Direct observation of behavior is a form of behavioral assessment that maintains a low level of inference when applied to psychoeducational decision-making. The recent proliferation of computer behavior-observation systems has provided observers with an efficient means for collecting direct behavior observational data (Kahng & Iwata, 2000; Shapiro & Kratochwill, 2000a, 2000b). Although computerized methods are less cumbersome than other recording methods (e.g., partial-interval form, narrative recording), it is not clear if computerized methods offer improvement by enhancing observational accuracy or reducing judgment bias that is generally initiated in applied settings (e.g. referral information; Arkes & Harkness, 1980; O'Reilly, Northcraft, & Sabers, 1989). Undergraduate participants (N = 243) were trained to perform a 10-minute observation using a computer observation program, partial interval form, narrative recording method, or a no-recording method. In addition, the participants were randomly given an internalizing, externalizing, or a non-specific psychopathology referral. The participants who used computerized and interval-recording methods consistently produced more accurate observations and were less susceptible to referral bias than the narrative and no-method groups. Implications for applied practice and suggestions for future research are discussed. ^

Subject Area

Psychology, Behavioral|Psychology, Psychometrics

Recommended Citation

Charles Edward Sicotte, "Computerized behavioral observation: Implications for improving accuracy and reducing bias" (2003). Dissertations and Master's Theses (Campus Access). Paper AAI3103725.