Observing behavior in a computer game.
Good news on a screen keeps kids clicking; bad news does not—so program your apps to cheer, not boo.
01Research in Context
What this study did
The team built a simple computer game. Players could click to see either good news or bad news about their score.
The researchers watched which clicks kept happening. They wanted to know if good news works like candy for the eyes.
What they found
Good-news clicks kept going. Bad-news clicks faded out.
The value of the information, not the screen flash, kept the behavior alive.
How this fits with other research
Cullinan et al. (2001) later used the same computer-task trick to spot which reinforcer features a student cares about.
Root et al. (2017) stepped back and said, "Nothing is a reinforcer until the moment and the person say so." The 1990 data already showed that: the same pixel color worked only when it carried good news.
Dembo et al. (2023) moved the idea online. They changed one sentence from "Check the box to join" to "Uncheck to skip." Sign-ups doubled, proving that tiny verbal frames still steer choice, just like good-news pixels did.
Dudley et al. (2008) added self-monitoring to computer feedback and fixed adults’ posture. The core rule—immediate screen consequences shape behavior—holds across decades and settings.
Why it matters
You already use screens in sessions. This paper tells you to load them with good news. Flash scores, points, or praise right after the right response. Skip the red-X buzzers. The data say they are weak reinforcers. Next time you open a tablet program, check: does it celebrate the learner? If not, swap it for one that does.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →Turn on the ‘celebrate’ sound in your software and mute the error beep for one student.
02At a glance
03Original abstract
Contingencies studied in lever-pressing procedures were incorporated into a popular computer game, "Star Trek," played by college students. One putative reinforcer, the opportunity to destroy Klingon invaders, was scheduled independently of responding according to a variable-time schedule that alternated unpredictably with equal periods of Klingon unavailability (mixed variable time, extinction schedule of reinforcement). Two commands ("observing responses") each produced stimuli that were either correlated or uncorrelated with the two components. In several variations of the basic game, an S-, or bad news, was not as reinforcing as an S+, or good news. In addition, in other conditions for the same subjects observing responses were not maintained better by bad news than by an uninformative stimulus. In both choices, more observing tended to be maintained by an S- for response-independent Klingons when its information could be (and was) used to advantage with respect to other types of reinforcement in the situation (Parts 1 and 2) than when the information could not be so used (Part 3). The findings favor the conditioned reinforcement hypothesis of observing behavior over the uncertainty-reduction hypothesis. This extends research to a more natural setting and to multialternative concurrent schedules of events of seemingly intrinsic value.
Journal of the experimental analysis of behavior, 1990 · doi:10.1901/jeab.1990.54-185