Behavior Assessment and Measurement
Assessment begins with a behavioral definition that is objective, clear, and complete - a stranger should be able to score it reliably. Avoid labels and inferences ("aggressive" is too vague; "closed fist contact with another person's body" is measurable).
Choose a measurement system that matches the behavior's dimensions. Frequency/rate fits discrete, short responses; duration fits behaviors with a clear start and end; latency is time to onset; inter-response time (IRT) is time between responses. When behavior is high-rate or continuous, use time sampling: partial interval tends to overestimate, whole interval tends to underestimate, and momentary time sampling gives an estimate at the end of each interval and is generally the least biased for duration. These directional biases are frequent exam traps.
Use indirect methods (interviews, rating scales) for efficiency but treat them as subjective; use direct observation and ABC recording for descriptive data. Functional analysis experimentally manipulates conditions to identify function, whereas a functional assessment may be descriptive only.
Collect baseline data before intervention, graph it, and monitor interobserver agreement (IOA) (e.g., smaller/larger × 100 for frequency; agreement divided by agreements plus disagreements for interval data) and treatment integrity. Rule of thumb: if IOA falls below about 80%, retrain observers before trusting the data. Worked example: a 20-minute observation records 10 responses → rate = 0.5 responses per minute.