How do reinforcers affect choice? Preference pulses after responses and reinforcers
Use visit analyses, not just residual preference-pulse fixes, to see true reinforcer effects on choice.
01Research in Context
What this study did
The team looked at how reinforcers steer choice in concurrent schedules.
They compared two ways to read the data: old-school residual preference-pulse fixes versus newer visit-level counts.
Pigeons pecked on two keys for food while computers logged every peck and hop.
What they found
Visit analyses showed clearer, cleaner pictures of reinforcer impact.
Residual pulse fixes often hid or twisted the real effect size.
In short, the new method told the truer story.
How this fits with other research
DeRoma et al. (2004) first used visit counts and saw rapid choice shifts after each reinforcer.
Gomes-Ng et al. (2017) now says everyone should copy that move instead of trusting the older pulse fix.
Busch et al. (2010) had already mapped how preference pulses grow across sessions, but their data can be re-run with visits for sharper insight.
Boutros et al. (2011) showed single reinforcers bias the next response; visit analyses would make that effect stand out even more.
Why it matters
If you graph concurrent-schedule data for a client, add visit-level counts. You will spot the real reinforcer effect faster and avoid the noise that residual fixes leave behind.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →Open your last concurrent-schedule file and add a column that counts visits to each alternative; compare the new graph to your old pulse-corrected one.
02At a glance
03Original abstract
In concurrent schedules, reinforcers are often followed by a brief period of heightened preference for the just-productive alternative. Such 'preference pulses' may reflect local effects of reinforcers on choice. However, similar pulses may occur after nonreinforced responses, suggesting that pulses after reinforcers are partly unrelated to reinforcer effects. McLean, Grace, Pitts, and Hughes (2014) recommended subtracting preference pulses after responses from preference pulses after reinforcers, to construct residual pulses that represent only reinforcer effects. Thus, a reanalysis of existing choice data is necessary to determine whether changes in choice after reinforcers in previous experiments were actually related to reinforcers. In the present paper, we reanalyzed data from choice experiments in which reinforcers served different functions. We compared local choice, mean visit length, and visit-length distributions after reinforcers and after nonreinforced responses. Our reanalysis demonstrated the utility of McLean et al.'s preference-pulse correction for determining the effects of reinforcers on choice. However, visit analyses revealed that residual pulses may not accurately represent reinforcer effects, and reinforcer effects were clearer in visit analyses than in local-choice analyses. The best way to determine the effects of reinforcers on choice may be to conduct visit analyses in addition to local-choice analyses.
Journal of the Experimental Analysis of Behavior, 2017 · doi:10.1002/jeab.260