Independence of terminal-link entry rate and immediacy in concurrent chains.
Entry rate and immediacy act like two volume knobs, not one blended dial.
01Research in Context
What this study did
Connell et al. (2004) worked with pigeons in a two-key chamber.
Birds first pecked on a center key to enter a terminal link.
Once inside, a side key delivered food after a short or long delay.
The team varied how often each terminal link opened and how soon food followed.
They counted pecks to see if entry rate and immediacy blended or stayed separate.
What they found
Peck data fit a model that adds the two factors instead of mixing them.
High entry rate and short delay each raised responding on their own.
Changing one did not alter the power of the other.
The birds treated how often and how soon as two separate pieces of information.
How this fits with other research
Thomas et al. (1968) saw the opposite pattern: when two reinforcers were signaled together, response rates landed in the middle, not on top.
The difference is procedure: R et al. presented cues at the same time, while E et al. kept entry and delay in different links, letting each factor act alone.
Duker et al. (1991) also split timing effects, showing that the gap between sample and choice hurts accuracy more than the gap between choice and food.
Both studies agree that timing and frequency can be pulled apart if the schedule allows it.
Why it matters
When you build a token board or choice system, treat how often a student can access the reinforcer and how soon it arrives as two separate levers.
You can speed up work by raising either frequency or immediacy without touching the other.
If a plan stalls, check which lever is weak and adjust just that one.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →Keep the token exchange delay the same but add two extra chances to trade in this week and watch response rate climb.
02At a glance
03Original abstract
In Phase 1, 4 pigeons were trained on a three-component multiple concurrent-chains procedure in which components differed only in terms of relative terminal-link entry rate. The terminal links were variable-interval schedules and were varied across four conditions to produce immediacy ratios of 4:1, 1:4, 2:1, and 1:2. Relative terminal-link entry rate and relative immediacy had additive and independent effects on initial-link response allocation, and the data were well-described by a generalized-matching model. Regression analyses showed that allowing sensitivity to immediacy to vary across components produced only trivial increases in variance accounted for. Phase 2 used a three-component concurrent-schedules procedure in which the schedules were the same as the initial links of Phase 1. Across two conditions, the relative reinforcer magnitude was varied. Sensitivity to relative reinforcer rate was independent of relative magnitude, confirming results of prior studies. Sensitivity to relative reinforcer rate in Phase 2 did not vary systematically across subjects compared to sensitivity to relative entry rate in Phase 1, and regression analyses confirmed again that only small increases in variance accounted for were obtained when sensitivities were estimated independently compared with a single estimate for both phases. Overall, the data suggest that conditioned and primary reinforcers have functionally equivalent effects on choice and support the independence of relative terminal-link entry rate and immediacy as determiners of response allocation. These results are consistent with current models for concurrent chains, including Grace's (1994) contextual choice model and Mazur's (2001) hyperbolic value-added model.
Journal of the experimental analysis of behavior, 2004 · doi:10.1901/jeab.2004.82-235