Fixed-interval performance: The dynamics of behavior and the interval length.
Under FI schedules, pause lengths barely predict the next pause, so change contingencies if you want different timing.
01Research in Context
What this study did
The team watched pauses after each reinforcer on fixed-interval (FI) schedules.
They varied the interval length and the size of the reinforcer.
The goal was to see if a long pause now meant a long pause next time.
What they found
Pause lengths only weakly predicted the next pause.
Longer FI times or bigger rewards did not change this link.
Once the pattern set in, it stayed stable across sessions.
How this fits with other research
Pisacreta (1982) saw humans lock into steady pause or no-pause styles that survived schedule swaps. H et al. now say those styles still show weak pause-to-pause links, so the lock is more about form than about strong memory.
Timberlake et al. (1987) later added competition and got clear break-and-run timing. That shows you can shift FI pauses, but you need extra contingencies; interval length alone will not do it.
English et al. (1995) gave rats a VR history before FI. Rates rose, yet pause autocorrelation stayed mild, matching the weak effect seen here.
Why it matters
If you run FI schedules in token boards, fluency drills, or DRL chains, do not expect the last pause to tell you much about the next one.
Instead of waiting for pause chains to break naturally, add contingencies such as brief competition or response cost if you need faster, steadier responding.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →Track post-reinforcement pause for five FI cycles; if the pattern hurts learning, layer a brief competitive timer or bonus response requirement to reshape timing.
02At a glance
03Original abstract
Postreinforcement pauses from successive intervals under various fixed-interval schedules (ranging from 15 seconds to 480 seconds in length) were subjected to lag-1 autocorrelation analysis. Results from both rats and pigeons suggested that there was a consistent tendency for pause values in successive intervals to be weakly positively related. This tendency did not appear to change systematically with interval length and was exhibited both when the reinforcer magnitude was constant and when it was variable at different interval values. The findings do not support suggestions that the dynamic properties of performance under fixed-interval schedules vary systematically with interval length, and are in the opposite direction from some previous findings suggesting that measures of behavior (such as post-reinforcement pause length or number of responses) in successive intervals are inversely related.
Journal of the experimental analysis of behavior, 1983 · doi:10.1901/jeab.1983.39-323