RBT Measurement Guide

Why measurement runs the whole show

Everything in ABA starts with data. Before your supervisor can write a plan, someone has to measure the behavior. During the plan, someone has to keep measuring to see if it works. That someone is you. About 17% of the RBT exam lives in this domain, and the questions are refreshingly concrete: which measurement fits this behavior, what does this graph show, which recording method overestimates. Learn the definitions cold and this domain becomes free points.

Continuous measurement: catch everything

Continuous measurement records every single instance of the behavior during the observation. Use it when you need the full picture.

• Frequency (count): how many times the behavior happened. Best for behaviors with a clear beginning and end, like hand raises or instances of hitting. Weakness: it ignores time, so 10 hits in 10 minutes looks the same as 10 hits in 2 hours.
• Rate: frequency divided by time (for example, 6 responses per minute). Rate fixes frequency's weakness and lets you compare sessions of different lengths. If the behavior is countable and time matters, rate is usually the answer.
• Duration: how long the behavior lasts, from start to finish. Use it for tantrums, crying, or on-task behavior where the length matters more than the count.
• Latency: how long it takes the behavior to start after the instruction or cue. "How fast did they start?" is a latency question, and it is a favorite on the exam.
• Interresponse time (IRT): the time between the end of one instance and the start of the next. Shorter IRTs mean the behavior is happening in rapid bursts.
• Permanent product: measuring what the behavior leaves behind, like completed worksheets or a cleaned room. No direct observation needed, just count the product.

Discontinuous measurement: sample smart

Discontinuous measurement chops time into intervals and records whether the behavior happened in each one. Use it when continuous recording is impractical, like tracking a behavior across a whole school day.

• Partial interval: mark the interval if the behavior happened at any point during it, even for a second. It overestimates behavior. Great for behaviors you do not want to miss, like aggression.
• Whole interval: mark the interval only if the behavior lasted the entire interval. It underestimates behavior. Good for continuous behaviors you want to increase, like staying seated.
• Momentary time sampling: look up only at the exact end of each interval and record whether the behavior is happening right then. The most efficient method, and the least biased of the three, but it can miss brief behaviors entirely.

Graphing: the line graph is your report card

Behavior analysts graph data on line graphs: time goes on the horizontal x-axis, the behavior measure goes on the vertical y-axis. Each data point is one session or observation, connected by lines so you can see the path of progress. When the intervention changes, a vertical phase line marks the switch, with labels like Baseline and Intervention.

Reading a graph means checking three things: level (are the points high or low), trend (are they going up, down, or flat), and variability (are they bouncy or steady). A good intervention shows the level moving in the right direction with a clear trend. Your job as an RBT is not to redesign the graph, it is to collect data cleanly enough that the graph tells the truth.

Quick-fire review

• Countable behavior + time matters = rate
• How long it lasted = duration; how fast it started = latency
• Overestimates = partial interval; underestimates = whole interval
• Time on the x-axis, behavior on the y-axis, phase lines mark changes

Last reviewed: October 7, 2026 against the BACB RBT 3rd edition task list.