ABA Fundamentals

Effects Of Varying Sample- And Choice-stimulus Disparity On Symbolic Matching-to-sample Performance.

Godfrey et al. (1998) · Journal of the experimental analysis of behavior 1998
★ The Verdict

Dimming the sample hurts cue detection; dimming the choices hurts choice discrimination—fix the right part and matching errors fall.

✓ Read this if BCBAs running matching-to-sample or stimulus equivalence lessons on tablets or interactive whiteboards.
✗ Skip if Practitioners working only with vocal tacts or listener responding without visual matching.

01Research in Context

01

What this study did

The team ran a symbolic matching-to-sample task with adult humans. They kept the same three black shapes on every trial. The trick was how bright each shape looked. They dimmed either the sample shape or the two choice shapes by up to 70 percent.

By changing only pixel brightness, they could test two ideas at once. Dimming the sample tests how hard it is to see the cue. Dimming the choices tests how hard it is to pick the right answer. One parameter change, two separate effects.

02

What they found

When the sample got dimmer, people took longer to start picking. Their errors rose in a clean curve. When the choices got dimmer, people still started fast but then hesitated before clicking. The curves fit a simple detection model with fixed parameters.

The model said: sample clarity controls stimulus-response discrimination. Choice clarity controls response-reinforcer discrimination. The numbers stayed the same no matter which part was dimmed, proving parameter invariance.

03

How this fits with other research

Baron et al. (1968) showed that forcing an extra look at the sample speeds up learning. Eisenmajer et al. (1998) now show why: a clear sample lowers the brain’s detection threshold. Together they tell us to first make the sample pop, then worry about the choices.

Sigurðardóttir et al. (2012) used the same three-choice layout to build Icelandic word classes. They got near-zero errors once classes formed. The 1998 paper says those low errors only happen when sample and choice clarity stay above the model’s threshold. The procedures line up like Lego bricks.

Rose et al. (2000) found that reinforcer size breaks matching-law parameter k. Eisenmajer et al. (1998) found that brightness does NOT break their detection parameter. One parameter bends, the other does not. The difference is sensory versus motivational variables.

04

Why it matters

You can now troubleshoot matching-to-sample errors with a flashlight test. If the learner hesitates before touching, dim the sample or raise contrast. If the learner touches then pulls back, dim the foil choices instead. No new software, no extra prompts—just tweak brightness until errors drop. Use this on tablet programs, flashcards, or SMART Boards tomorrow.

FREE CEUs

Get CEUs on This Topic — Free

The ABA Clubhouse has 60+ on-demand CEUs including ethics, supervision, and clinical topics like this one. Plus a new live CEU every Wednesday.

60+ on-demand CEUs (ethics, supervision, general)
New live CEU every Wednesday
Community of 500+ BCBAs
100% free to join
Join The ABA Clubhouse — Free →
→ Action — try this Monday

Open your matching app, lower the brightness on one foil choice by 30 percent, and record if errors drop.

02At a glance

Intervention
stimulus equivalence training
Design
single case other
Sample size
6
Population
other
Finding
not reported

03Original abstract

Six pigeons were trained on a typical two‐stimulus two‐response symbolic matching‐to‐sample task involving the discrimination of sample and choice stimuli that were signaled by different probabilities that pixels in an area were lit on a computer screen located behind the response keys. The disparities of the sample and choice stimuli were systematically varied across five experimental parts by manipulating the probabilities of pixel illumination. Across conditions within parts, the ratio of reinforcers obtainable for matching responses was varied over five levels. A recent model of detection based on the discriminability between the stimulus—response relations and between response—reinforcer relations provided an effective description of the data. Consistent with this model, changes in the disparity of the sample stimuli led to decreases in stimulus—response discriminability and left response—reinforcer discriminability unchanged. Equally, changes in choice‐stimulus disparity caused a decrease in estimates of response—reinforcer discriminability and not in stimulus—response discriminability. Parameter invariance was thus obtained, and the variables expected to affect these parameters (sample and choice stimulus disparity) were correctly identified. The reasons for the failure of two recently reported studies to support parameter invariance under this model are discussed.

Journal of the experimental analysis of behavior, 1998 · doi:10.1901/jeab.1998.69-311