Punishment in human choice: direct or competitive suppression?
Punishment cuts the punished response directly; it does not make other responses extra valuable.
01Research in Context
What this study did
College students played a computer game. They could press either of two buttons. Pressing earned small amounts of money on variable-interval schedules. Losing money acted as punishment. The team wanted to know which math model fit the data best.
One model says punishment simply cuts the rate of the punished response. The other model says punishment also boosts the value of any other activity. The study pitted the two ideas against each other.
What they found
Most students’ choices lined up with the first, simpler model. The direct-suppression story won. Money loss mainly reduced the punished button presses. It did not make the other button extra attractive.
How this fits with other research
McGonigle et al. (1982) saw a different picture. When they punished or restricted one response, only the most likely alternative rose. That result leans toward the two-factor, competitive view. The clash is only on the surface. J used response restriction and watched sequential links, while S used money loss and watched steady choice. Different methods, different windows.
Smith et al. (2010) used the same lab set-up but with reinforcement instead of punishment. People still picked the low-effort option. The choice mechanics look similar whether the consequence is gain or loss.
Emmelkamp et al. (1986) also ran head-to-head trials in the same population. They compared two correction styles for learning patterns. Like S, they found one clear winner, showing that single-case lab comparisons can cleanly separate procedures.
Why it matters
When you add a punisher, expect direct suppression of the target response. Do not count on a big jump in other behaviors. If you need a replacement skill to grow, program extra reinforcement for it. Keep your data simple: one line for the target rate, one for the main alternative. The single-factor model gives you a quick visual check that usually fits.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →Graph the target behavior and one top alternative; if punishment suppresses the target but the alternative stays flat, you are seeing direct suppression—no need to hunt for hidden gains.
02At a glance
03Original abstract
This investigation compared the predictions of two models describing the integration of reinforcement and punishment effects in operant choice. Deluty's (1976) competitive-suppression model (conceptually related to two-factor punishment theories) and de Villiers' (1980) direct-suppression model (conceptually related to one-factor punishment theories) have been tested previously in nonhumans but not at the individual level in humans. Mouse clicking by college students was maintained in a two-alternative concurrent schedule of variable-interval money reinforcement. Punishment consisted of variable-interval money losses. Experiment 1 verified that money loss was an effective punisher in this context. Experiment 2 consisted of qualitative model comparisons similar to those used in previous studies involving nonhumans. Following a no-punishment baseline, punishment was superimposed upon both response alternatives. Under schedule values for which the direct-suppression model, but not the competitive-suppression model, predicted distinct shifts from baseline performance, or vice versa, 12 of 14 individual-subject functions, generated by 7 subjects, supported the direct-suppression model. When the punishment models were converted to the form of the generalized matching law, least-squares linear regression fits for a direct-suppression model were superior to those of a competitive-suppression model for 6 of 7 subjects. In Experiment 3, a more thorough quantitative test of the modified models, fits for a direct-suppression model were superior in 11 of 13 cases. These results correspond well to those of investigations conducted with nonhumans and provide the first individual-subject evidence that a direct-suppression model, evaluated both qualitatively and quantitatively, describes human punishment better than a competitive-suppression model. We discuss implications for developing better punishment models and future investigations of punishment in human choice.
Journal of the experimental analysis of behavior, 2003 · doi:10.1901/jeab.2003.80-1