Premium
REANALYSIS OF THE EQUATION FOR SIMPLE ACTION
Author(s) -
WarrenBoulton Frederick R.,
Silberberg Alan,
Gray Michael,
Ollom Randolph
Publication year - 1985
Publication title -
journal of the experimental analysis of behavior
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.75
H-Index - 61
eISSN - 1938-3711
pISSN - 0022-5002
DOI - 10.1901/jeab.1985.43-265
Subject(s) - reinforcement , variance (accounting) , action (physics) , constant (computer programming) , matching law , structural equation modeling , simple (philosophy) , function (biology) , statistics , matching (statistics) , scaling , value (mathematics) , accommodation , mathematics , psychology , computer science , social psychology , geometry , physics , philosophy , accounting , epistemology , quantum mechanics , evolutionary biology , business , biology , programming language , neuroscience
De Villiers and Herrnstein (1976) have shown that the equation for simple action, derived from the matching law, predicts change in behavioral output for some 40 experiments in which the value of a single source of reinforcement has been varied. Using only the positive‐reinforcement studies they cite that used five or more different reinforcement values, we found the high percentage of variance they report accommodated by this equation (94%) is predicated on instances of averaging rates of behavioral output before making a least‐squares fit of the equation. In our reanalysis, which minimizes rate averaging, only 78% of the data variance is accommodated. This diminished data‐variance accommodation can be improved by adding parameters that permit the equation's scaling constant to change as a function of reinforcement. Although these parameters permit acceptable levels of accommodation of data variance, they correspond to no obvious psychological processes. These findings support the view that the equation for simple action is an inadequate model for behavioral output.