Psychological Review

Logan, Frank A.

doi: 10.1037/0033-295X.86.6.507pmid: N/A

Hybrid theory of classical conditioning is extended to encompass FI schedules of reinforcement in operant conditioning. The basic assumption, familiar in other incentive-type theories, is that the expectation of reward becomes conditioned to response-produced feedback stimuli from the operant response. The resulting excitatory potential changes dynamically in relation to the stimulus-memory trace of the preceding reinforcer, the trace being conceptualized in terms of molecular elements with different lifetimes. Finally, excitatory potential is transformed into momentary probability of response emission on the basis of a randomly variable threshold, the mean of which also changes dynamically with the dynamogenic properties of response-produced feedback stimulus-memory traces. The theory implies a cyclical transition between 2 basic but related states, one corresponding to the postreinforcement pause in responding and the other corresponding to the running response rate. This theoretical analysis fits well with the empirical data presented. (70 ref)

Tversky, Amos; Sattath, Shmuel

doi: 10.1037/0033-295X.86.6.542pmid: N/A

Presents a probabilistic choice model called preference tree or pretree that describes choice as a covert hierarchical elimination process. Each alternative is represented as a measurable collection of aspects, and the entire ensemble of aspects is assumed to have a tree structure. At each stage, one selects an aspect (i.e., a branch of the tree) with a probability that is proportional to its measure and eliminates all alternatives not belonging to the selected branch. The process continues until a single alternative remains. Pretree is considerably more parsimonious than the more general model of elimination by aspects, and at the same time it is much less restrictive than models that assume that the strength of preference of one alternative over another is independent of other alternatives. An additional interpretation of pretree as an aggregate model is developed, and necessary and sufficient conditions for the existence of tree representations for binary choice probabilities are established. Pretree fits several sets of individual and aggregate choice probabilities and explains the prevalent violations of the constant ratio rule. (32 ref)

Roth, Alvin E.; Malouf, Michael W.

doi: 10.1037/0033-295X.86.6.574pmid: N/A

Examines the behavioral implications and empirical testability of the game-theoretic models of bargaining that follow in the tradition begun by J. F. Nash (1950, 1953). The classical game-theoretic models are reviewed with particular attention to the assumptions they make concerning the kinds of information shared by the bargainers. The experimental literature concerned with testing these models is also reviewed, with attention to the extent to which these experiments have conformed to the assumptions of the models they are designed to test. Some of the principal differences between the theoretical predictions and the results reported in the experiments seem to be due to questions concerning the information shared by the bargainers. A new game-theoretic model is presented, based on assumptions about information that more closely correspond to the conditions under which many of the experiments have been conducted. An experiment involving a bargaining game with undergraduate Ss manipulated the appropriate kind of information and supported the hypothesis that Nash's bargaining model has predictive value in situations that conform to its assumptions about information, but results also suggest that only a relatively narrow range of situations may conform completely to these assumptions. (47 ref)