Verbal self-reports of emergent relations in a stimulus equivalence procedure.
College students can mostly talk about their new equivalence classes, but occasional mismatches remind us to trust behavior first, words second.
01Research in Context
What this study did
Researchers taught 12 college students to match nonsense shapes. After training, they asked, 'Which shapes go together?'
They wanted to see if people could verbally report the new relations that emerged without direct teaching.
What they found
Most students could name the trained pairs and the new pairs. But a few said 'A goes with C' when their choices showed A-B, B-C.
The words and actions did not always match, hinting that some derived relations stay quiet inside the head.
How this fits with other research
Anonymous (1993) used the same lab set-up but added a 'not-this' cue. Their adults still formed classes, showing the core effect is sturdy.
Alonso-Álvarez et al. (2018) dropped self-report and used color cues instead. They got clean equivalence without asking anyone to talk, so the act can stay non-verbal.
Turkkan (1994) found college kids often mis-count how many times they were right when success is rare. That warns us: spoken guesses are fragile data.
Why it matters
If you probe emergent relations by asking, check the child's actual selections too. A quiet 'yes' on the tablet may tell you more than a spoken answer.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →After equivalence probes, ask the client 'Which pictures belong?' then quickly run a matching trial to see if the words match the selections.
02At a glance
03Original abstract
During a preliminary training phase, college students were taught to categorize each of their responses accurately in a conditional discrimination task as either correct or incorrect. Next, in the absence of self-reports, subjects acquired conditional discriminations (involving novel stimuli) prerequisite to the formation of two four-member equivalence classes. The self-report procedure was reinstated during probe sessions that tested for untrained relations indicative of equivalence class formation. Interspersed trials involving trained relations provided a positive control, and trials with no class-consistent comparison provided a negative control. Eight of 10 subjects demonstrated equivalence class formation; all accurately reported their performance on trained relations and on trials with no class-consistent comparison. Subjects also reported their performance on most untrained (emergent) relations accurately, but in several instances self-reports indicated failure or uncertainty despite nearly perfect emergent-relations performance. These inconsistencies add to a growing body of literature that suggests there are differences between individual types of emergent relations. We suggest that the present procedure may be helpful in understanding these differences and other equivalence-related effects.
Journal of the experimental analysis of behavior, 1996 · doi:10.1901/jeab.1996.65-355