Decision (Including Decision Theory)

Introduction

Most problems dealt with in the practice of psychology can be solved only by the application of a treatment, e.g. an intervention. In most cases, more than one treatment is available. The psychologist has to choose between several options. His or her choice of or decision for a certain treatment cannot be made in an arbitrary way but has to follow certain rules which are the subject of a field of study called decision theory (cf. Klein et al., 1993). In the practice of psychology, an important component of the decision rules are case-related assessment data, which are collected in the course of an assessment process. The role of assessment data in the decision-making process is the topic of this entry.

A Classification of Decisions

For every decision it is necessary that a minimum of two alternative treatments is given, e.g. acceptance versus rejection of an applicant. Decisions which occur during the assessment process can be classified within a system proposed by Cronbach and Gleser (1965), who distinguished between six kinds of assessment-related decisions.

The first feature of decisions refers to whether the gain of a decision is in favour of an institution or an individual. A decision is institutional, for example, when an organization tests all individuals using the same standardized procedure. In such a case, a decision rule is required that yields as much benefit to the institution as possible from multiple (homogeneous) decisions over all individuals. Individual interests may be taken into account, but only insofar as they affect the realization of the goals of the institution. A decision is individual, for example, when an individual asks an institution for help in a decision making process. In order to get information that can support the individual in the decision, the institution arranges a specific test programme. Individual decisions are often unique. The choice confronting the decision maker may rarely or never recur. In this case only the individual benefit is important.

The second feature of decisions distinguishes between fixed and variable rates of acceptance. A fixed rate of acceptance exists, for example, when job openings in a company are limited to a certain number. In such a case the decisions [Page 294]depend on each other. A non-fixed or variable acceptance rate exists, when independence of the decisions is given. This is the case, for example, when there is a job for every applicant who fulfils the respective requirements.

Cronbach and Gleser (1965) also differentiate between single-stage and multi-stage (sequential) tests. In the first case, the decision is made in only one step, in the second case in multiple steps on the basis of a sequential procedure.

The fourth feature of decisions refers to whether persons are selected (e.g. for a job, training, therapy) or placed to different treatments. When people are selected, only a certain number is accepted. However, when they are placed, nobody is excluded from the institution, but each person is assigned to the treatment with the best fit to his or her individual characteristics.

Assessment data are either restricted to only one dimension (univariate information) or composed of more dimensions (multivariate information). The use of multiple predictors increases the validity of a decision, because multiple facets of the criterion can be considered.

...