Integral Review

A Transdisciplinary and Transcultural Journal For New Thought, Research, and Praxis

Posts Tagged ‘psychometrics’

Investigating the Validity of the Ogive Aggregation Method, Including the use of Rasch Analysis for the Sentence Completion Test and the STAGES Model

Tom Murray

This project assesses psychometric aspects of the STAGES sentence completion test (SCT) using data from 740 scored surveys, and some of the analysis applies to all variations of the SCT for ego development (meaning making maturity). The goals of this research project include: (1) to apply item response theory (IRT) and Rasch analysis to determine item-level psychometric properties of the SCT that were previously unaddressed in SCT research; and (2) to further investigate suspected problems with the ogive cutoff method for aggregating item scores in the SCT and propose alternatives. The psychometric analysis includes: within-test item normality, item standard deviations, test length analysis, factor analysis, characteristics of and correlations among each item, overall test strength, and construct levels discrimination. A range of issues with the standard ogive cutoff method are described, and a new item aggregation method is then proposed.

Tags: , , , , ,

Models, Metrics, and Measurement in Developmental Psychology

Zachary Stein and Katie Heikkinen

Abstract: Developmental psychology is currently used to measure psychological phenomena and by some, to re-design communities. While we generally support these uses, we are concerned about quality control standards guiding the production of usable knowledge in the discipline. In order to address these issues precisely, we provide an overview of the discipline’s various facets. We distinguish between developmental models and developmental metrics and relate each to different types of quality-control devices. In our view, models are either explanatory or descriptive, and their quality is evaluated in terms of specific types of disciplinary discourse. Metrics are either calibrated measures or soft measures, and their quality is evaluated in terms of specific psychometric parameters. Following a discussion on how developmentalists make metrics, and on a variety of metrics that have been made, we discuss the two key psychometric quality-control parameters, validity and reliability. This sets the stage for a limited and exploratory literature review concerning the quality of a set of existing metrics. We reveal a conspicuous lack of psychometric rigor on the part of some of the most popular developmental approaches and invite remedies for this situation.

Tags: , , , , , , , , ,

Current Issue

Recent Issues