Estimation of indifference points with an adjusting-delay procedure.
Keep step size at one second or smaller in adjusting-delay tasks to avoid noisy, inflated indifference points.
01Research in Context
What this study did
The author tested how step size changes the results of an adjusting-delay choice task.
In this task one reward stays fixed while the delay to the other reward shifts until the subject picks each side half the time.
Different step sizes were tried to see which gave the cleanest indifference point.
What they found
Steps larger than one second made the indifference point bounce around.
Big steps sometimes gave inflated estimates of how long the subject was willing to wait.
Other tweaks to the setup had no clear effect.
How this fits with other research
Torres et al. (2011) later used the same 1-second rule and added fancy math to show that choices wobble at first then settle down.
Cox et al. (2015) warn that adjusting-delay tasks can look less reliable than fixed-delay tests, but they used bigger steps, so the noise Mazur (1988) predicted may explain their weaker results.
Fraley (1998) took the tiny-step method and applied it to procrastination, proving the tool can reach beyond simple delay discounting.
Why it matters
When you run any adjusting-progress procedure, lock your step size at one second or less. This keeps data clean and prevents false estimates that could mislead your treatment decisions. It is a quick setting that saves hours of noisy data.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →Check your assessment software and lower any adjusting step to 1 s before the next session.
02At a glance
03Original abstract
In a series of conditions, pigeons chose between 1.5 s and 3 s of access to grain, each preceded by some delay. The delay that preceded the small reinforcer was constant throughout a condition. The delay that preceded the large reinforcer was increased or decreased a number of times each session in order to estimate an "indifference point," a delay at which the subject chose each alternative about equally often. The experiment was designed to determine whether variations in any of four features of this adjusting-delay procedure would systematically alter the estimated indifference points. The four features were the total trial duration, the number of center-key responses necessary to begin a trial, the number of choice trials that preceded each change in the adjusting delay, and step size--the size of each increment and decrement in the delay. Manipulation of the first three features had no systematic effects on the indifference points. As step size was increased from 0.5 s to 6 s, within-session variability of the adjusting delay steadily increased, and the 6-s step size produced larger indifference-point estimates for some subjects. The results suggest that, within certain limits, these procedural features can be altered without affecting the indifference-point estimates, but that the use of a large step size can distort the estimates. Some theoretical implications of the relative constancy of indifference points across these procedural variations are discussed.
Journal of the experimental analysis of behavior, 1988 · doi:10.1901/jeab.1988.49-37