Control of human vigilance by concurrent schedules.
Human attention splits across tasks the same way animal responding does—relative reinforcement rate drives the share.
01Research in Context
What this study did
Researchers asked adults to watch two lights. One light blinked on a fixed-time schedule. The other blinked on a mixed schedule.
If the participant pressed a button when the correct light blinked, they earned money. The team added feedback and spoken hints partway through.
The goal was to see if human watching behavior would match the patterns seen in animal labs.
What they found
People quickly split their button presses to match the pay-off rates. They pressed more on the side that paid more often.
When feedback and hints were added, the split stayed steady for the rest of the session.
The result showed that human vigilance is controlled by the same schedule rules that control rat or pigeon key pecks.
How this fits with other research
Thompson (1975) ran a near-copy of this setup and got the same matching split, giving early confidence that the effect is real.
Pierce et al. (1983) later pulled every human concurrent study into one review. They counted Schmidt et al. (1969) as key proof that the matching law works with people.
Savastano et al. (1994) pushed further by using ratio plus interval schedules. Their adults did not chase every coin; they landed between matching and maximizing. This looks like a clash, but the schedules were harder. The 1969 result still holds for simple vigilance tasks.
Why it matters
If you want a client to share attention between two tasks, set clear, steady pay-offs for each. The matching law will do the splitting for you. Use praise or tokens on fixed or variable timers and watch the child drift toward the richer side without any extra prompts.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →Set two brief vigilance games side-by-side, pay one on a fixed 30 s timer and the other on a variable 15 s timer, then count responses to see the matching split emerge.
02At a glance
03Original abstract
Twenty four subjects were studied for ten 1-hr sessions to determine whether the human observer's visual monitoring of individual meters in a complex display can be differentially controlled by concurrent scheduling of signals. Subjects were divided into two main groups of 12 each. One group was given fixed-interval, variable-interval, and differential-reinforcement-of-low-rates schedules. The second group was given fixed-interval, fixed-ratio, and differential-reinforcement-of-low-rates schedules. Test subjects were instructed only to detect as many signals as possible. Results indicated that observing responses to the individual meters corresponded to the temporal patterns known to be associated with the schedules for the group given fixed-ratio instead of variable-interval as a component schedule. The group given the variable-interval schedule in the three-schedule combination tended to exhibit the same pattern of viewing across each of the three meters during any given session. However, subsequent testing was performed on two more subjects over 64 sessions, by adding initial feedback of signal detection results, and instructions concerning schedule construction. These results indicated that with knowledge of schedule construction and initial feedback of detection data, differentiated responding can be maintained efficiently over long periods of time by the combination including fixed-interval, variable-interval, and differential-reinforcement-of-low-rates schedules.
Journal of the experimental analysis of behavior, 1969 · doi:10.1901/jeab.1969.12-591