Table 6 - uploaded by Teresa Russell
Content may be subject to copyright.
Source publication
The Armstrong Laboratory and the Army Research Institute cosponsored a project to develop a Joint-Service classification research agenda, or Roadmap, for reducing redundancy of research across Services and improving inter-Service research planning. The Joint-Service Classification Research Roadmap is a research agenda designed to enhance the Servic...
Citations
... Classification systems can focus on one or more of a variety of goals (Bobko, 1992;Campbell, 1993;Wise, 1992). The design of the system and the way its usefulness is evaluated may be greatly influenced by the goals that are chosen. ...
The Joint-Service Classification Research Roadmap is a research agenda designed to enhance the Services' selection and classification research programs. It is composed of numerous research questions that are organized into seven broad activities. Ordered roughly from highest to lowest priority, they are: Building a Joint-Service policy and forecasting data base, capturing criterion policy, modeling classification decisions, developing new job analysis methodologies, investigating fairness issues, conducting criterion measurement research, and conducting predictor-related research. The first two activities, 'Building a Joint-Service policy and forecasting data base' and 'Capturing criterion policy,' will facilitate research planning. 'Modeling classification decisions' and 'Developing new job analysis methodologies' are activities wherein long-term research is needed. Classification is important because (a) changes in the ASVAB will result in revised composites, (b) recent innovations make classification research timely, and (c) downsizing makes classification more important. Job analysis research is needed to (a) facilitate innovations in predictor and criterion development and (b) facilitate management of selection and classification for future jobs. Fairness is important from a policy perspective. Criterion and predictor-related research are important, but the Services have researched them extensively. Extended research on experimental measures that have yielded promising results is recommended.
Using T. A. Cleary's (1968) model of test bias, relations between aptitude scores and training performance were evaluated for race and gender subgroups of 12,166 male and 1,292 female military officer candidates. Regression analyses on data from archival files revealed level bias with minority performance being overpredicted by a small and constant amount at all aptitude levels, suggesting that test usage results in higher selection rates for female and Black cadets. These results are consistent with the literature in education, industry, and prior studies conducted in the military. (PsycINFO Database Record (c) 2014 APA, all rights reserved)
The Armstrong Laboratory, the Army Research Institute for the Behavioral and Social Sciences, the Navy Personnel Research and Development Center, and the Center for Naval Analyses are committed to enhancing the overall efficiency of the Services' selection and classification research. This means reducing redundancy of research across Services and improving inter-service research planning, while ensuring that each Service's priority needs are served. The Roadmap project products describe across-service military classification research issues. This report serves two purposes. First, it is a reference document that selection and classification experts in the Services can use in making decisions about predictor measures. It provides information about operational and experimental predictors. Second, it refines and supplements predictor-related research objectives that emerged from the Services! current selection and classification practices and interviewing military selection and classification experts to identify selection and classification research objectives. Measurement, Individual differences, Roadmap.
The open problem of establishing an accurate adjustment formula for the Brogden (1959) table used in estimating assignment benefits is solved by taking a distribution-theoretic approach in the case of equally correlated jobs. This approach establishes a more sound adjustment formula which provides more accurate results than the formula provided by Alley & Darby (1994). Implications for the use of the adjusted table in the case of correlated jobs is discussed.
