Continuous Measurement
Frequency/Count: Number of times the behavior occurs. Use for discrete behaviors with a clear onset and offset. Rate: Frequency divided by time. Use when observation periods vary in length. Duration: Total time the behavior occurs. Use when the length of the behavior is important (e.g., tantrums, on-task). Latency: Time from stimulus to response onset. Use when response speed matters (e.g., compliance latency). Interresponse Time (IRT): Time between consecutive responses. Use when the spacing of behavior matters.
Discontinuous Measurement
Partial Interval Recording: Record if the behavior occurred at any point during the interval. Overestimates occurrence. Best for behaviors you want to decrease. Whole Interval Recording: Record only if the behavior occurred throughout the entire interval. Underestimates occurrence. Best for behaviors you want to increase. Momentary Time Sampling (MTS): Record if the behavior is occurring at the exact moment the interval ends. Provides an unbiased estimate with large samples. Best when you cannot observe continuously.
Decision Guide
If the behavior is discrete with a clear start and stop: use frequency or rate. If the length of the behavior matters: use duration. If how quickly someone responds matters: use latency. If you cannot observe continuously: use momentary time sampling for the best estimate. If you need to demonstrate a behavior is problematic: partial interval overestimates and may support intervention funding. If you need to show a behavior is improving: whole interval is conservative and credible.