Latency measurement in functional analysis and treatment of behaviors targeted for reduction

Mattson and team read every paper that used latency in a functional analysis. They found 27 studies that timed how long a behavior took to start after the condition began.

The review looked at kids and adults whose problem behavior was rare or dangerous. Instead of counting how many times the behavior happened, these studies measured how many seconds passed until the first hit, bite, or scream.

Latency worked as well as the usual count data for showing why the behavior occurred. The first response came faster in the condition that matched the teacher’s guess about function.

The studies also showed real-life perks: sessions finished sooner and staff stayed safer because they did not have to let severe behavior pile up.

Hansen et al. (2019) ran latency-based FAs in real classrooms. Their results only partly matched the standard count-based FA, so they warned not to trust latency alone for every student.

Frank-Crawford et al. (2026) compared latency and rate ways of picking reinforcers for self-injury. When the two methods disagreed, the old rate method picked the better toy.

These two papers seem to clash with the review’s upbeat tone, but they don’t. The review simply says latency is “viable”; the newer trials add the warning that it can miss the mark for some kids or for very automatic self-injury.

If a student’s aggression happens once an hour, counting will eat your whole day. Start the session, start the stopwatch, and stop at the first hit. You will finish in five minutes and still know which condition wins. Just double-check with count data if the latency picture looks muddy or if another study like Frank-Crawford hints that automatic reinforcement is in play.