Effects of mastery criterion on the emergence of derived equivalence relations

Fienup and team asked: how picky should we be when kids learn matching pictures to words? They tested 24 college students with no known disabilities. Each person sat at a computer and learned A-B and B-C matches until they hit one of three mastery rules: 12 correct in a row (block), 12 correct within every 12 trials (rolling), or only 9 correct in a row (looser).

The researchers then checked if the students could suddenly pick the never-taught A-C pairs. This untaught skill is called derived equivalence. The whole study took one short lab visit.

Both strict rules (12-block and 12-rolling) produced the same strong result: almost every student showed the new A-C matches right away. The easier 9-in-a-row rule failed about half the time. In plain words, asking for 12 correct answers, either all at once or spread out, is the safety zone for equivalence to pop out.

Oliveira et al. (2021) later showed that you can also save time by using a one-to-many (OTM) teaching order instead of a long line of pairs. Their college students learned faster, but they still used Fienup’s tight 12-trial mastery rule. The two studies team up: keep the high criterion and switch to OTM structure for fewer total trials.

O’Connor et al. (2020) took the same logic to three children with autism. Two kids derived new emotion-to-person links without extra teaching, just like Fienup’s students. The third child needed a quick correction program. The pattern shows the 12-trial mastery bar works for both neurotypical adults and kids with autism, but clinicians should stay ready to give a booster if emergence stalls.

Vladescu et al. (2021) looked like they disagreed at first glance. They found bigger stimulus sets (12 items) slowed down kids with autism during tact training. Fienup used 12 trials, not 12 stimuli, so the studies actually tweak different knobs. One warns against piling too many pictures in a set; the other warns against lowering the accuracy bar.

If you run equivalence lessons for name-to-picture or emotion-to-face drills, set the mastery bar at 12 correct responses before you test for emergent relations. You can check the 12 consecutively or within the last 12 trials—both work. Drop to 9 and you risk half your learners needing re-teaching. Pair this rule with an OTM layout to save time, and keep a quick correction plan handy for learners who don’t derive the new links right away.