Psychological Measurement and Models of Individual Streams of Behavior

A general strategy for single subject research, based on the collection, exploration, and modeling of time and frequency information in a "stream protocol" (a complete record of the time and type of all behavioral changes during an interval of observation), is proposed and evaluated. A special technique for modeling time and frequency information in an individual stream is developed, a technique grounded in the proposition that both the time spent in a type of behavior and the frequency of that behavior are related to the attractiveness of the behavior. In this way observable time and frequency aspects of the data are related to each other via attractiveness. Several model variations which use Luce's (1959) choice axiom to aid in the time-frequency mapping are presented and the model variations are applied to Timberlake's (1969) data on the free behavior of the laboratory rat. It is shown that the data are fit well by a model conjunction which is formed out of a semi-Markov model with transition probabilities and relative time spent values that conform to Luce's choice axiom. This model conjunction is related to two previously developed stream models--Birch's (1984) Activation Time Scheduling model and a Markov model--through a common prediction. The general strategy is shown to be more comprehensive than conventional single subject research techniques, to allow for the assessment of a new treatment effects hypothesis, to provide an easy-to-use framework for model building, and to encompass presently available models for the stream of behavior.

Name thesis:degree_name

Ph.D.

Level thesis:degree_level

Dissertation

Discipline thesis:degree_discipline

Psychology

Grantor

University of Illinois at Urbana-Champaign

Year dc:date

2014

Author dc:creator

• Holtgrave, David Robert

• Psychology, Psychometrics

Identifier

(UMI)AAI8815355

OAI identifier oai:identifier

oai:www.ideals.illinois.edu:2142/69711