NCJ Number
85995
Date Published
1981
Length
126 pages
Annotation
Psychometric analysis was applied to some drug knowledge and attitude data to demonstrate how this method can be used to transform test and questionnaire scores into measures of knowledge and attitude.
Abstract
The method of examining score matrixes focuses on the main line of inquiry, the intended variable, which led to collection of the data displayed in the matrix. It extracts estimates of where the persons and the items are located along this intended line. In doing this, person and item scores are transformed into person measures and item calibrations into 'logits.' These logits have advantages over raw scores. They facilitate commensurating person and item scores because they share a common unit on a common variable and thus are on one and the same line. This enables one to see where any individual person is located among the items and also to specify expected mastery or agreement levels for them. Furthermore, because these measures and calibrations are on an interval scale, it is possible to interpret equal differences in logits as implying equal differences in knowledge or attitude. Finally, the logits have an objectivity not possessed by scores, in that they need not reflect the difficulty of the items upon which persons are measured or the knowledge levels or attitude positions of the persons upon whom items are calibrated. This method provides a way of checking on the appropriateness of trying to measure a particular person or calibrate a particular item (i.e., a way of performing quality control by pointing out item or person 'misfits' in the overall pattern). Charts and tabular data are given, as is a 25-item bibliography. The appendix contains the study instruments.