The quality of measurement systems used in almost all natural sciences other than behavior analysis is usually evaluated through calibration study rather than relying on interobserver agreement. We demonstrated some of the basic features of calibration using observer-measured rates of free-operant responding from 10 scripted 10-min calibration samples on video. Five novice and 5 experienced observers recorded (on laptop computers) response samples with a priori determined response rates ranging from 0 to 8 responses per minute. Observer records were then compared with these predetermined reference values using linear regression and related graphical depiction. Results indicated that all of the observers recorded rates that were accurate to within plus or minus 0.4 responses per minute and 5 were accurate to within plus or minus 0.1 responses per minute, indicating that continuous recording of responding on computers can be highly accurate and precise. Additional research is recommended to investigate conditions that affect the quality of direct observational measurement of behavior. (Contains 2 tables and 3 figures.)