Do Functional Analyses Probe Learning Histories, Or Create Them? An Exploratory Investigation.

The team asked a simple question. Do our functional analyses show what kids already learned, or do the test conditions teach new problems?

They ran two steps. First, a quick reinforcer test showed what items or attention each child worked for. Second, a full functional analysis checked if those same things made problem behavior spike.

Eight children with mixed diagnoses took part in an outpatient hospital clinic.

For six of the eight kids, the reinforcer test did not perfectly match the FA result.

That mismatch is normal. It means the FA was reading the child's history, not inventing a new function.

Older papers already warned that odd stimuli can twist FA outcomes. Tassé et al. (2013) list idiosyncratic antecedents clinicians add when standard FAs look flat. Jolliffe et al. (1999) showed poor test-retest reliability across weeks or months. These worries raised the same core issue: are we measuring real functions or lab artifacts?

Heald et al. (2020) answers directly. The imperfect match between a quick reinforcer test and a full FA is expected, because the FA is probing deeper learning history, not creating false positives.

Matson et al. (2013) add that function usually stays the same even when you switch rooms or therapists. Together, the studies tell a coherent story: FAs are mostly valid, but you still need to watch for setting or method quirks.

You can trust your FA even when it does not line up with a brief reinforcer check. Keep running standard conditions first. If the data look muddy, follow the leads from Tassé et al. (2013): add one idiosyncratic antecedent at a time, then retest. This saves hours of second-guessing your results and keeps treatment planning on track.