Content uploaded by Christopher Hertzog
Author content
All content in this area was uploaded by Christopher Hertzog on May 18, 2014
Content may be subject to copyright.
A preview of the PDF is not available
Officer Career Development: Analytic Strategy Recommendations
Abstract and Figures
Strategies are recommended for analyzing information from the data bank developed by the Personnel Distribution and Career Development (PDCD) work unit for the purpose of establishing empirically-based decision guides to assist in the design and implementation of career policy and practice in the U.S. Navy. A set of analytic models is proposed wherein each model addresses an important issue concerning the development of empirically-based decision guides for career development. The statistical assumptions underlying each model are reviewed, as are methods that may be used to reasonably satisfy these assumptions. Estimation techniques and procedures for avoiding common errors in estimation also receive attention. Keywords: Analytic strategy, Latent variable, Time series, Cohort analysis, Moderator analysis.
Figures - uploaded by Christopher Hertzog
Author content
All figure content in this area was uploaded by Christopher Hertzog
Content may be subject to copyright.
ResearchGate has not been able to resolve any citations for this publication.
In this article, we provide guidance for substantive researchers on the use of structural equation modeling in practice for theory testing and development. We present a comprehensive, two-step modeling approach that employs a series of nested models and sequential chi-square difference tests. We discuss the comparative advantages of this approach over a one-step approach. Considerations in specification, assessment of fit, and respecification of measurement models using confirmatory factor analysis are reviewed. As background to the two-step approach, the distinction between exploratory and confirmatory analysis, the distinction between complementary approaches for theory testing versus predictive application, and some developments in estimation methods also are discussed.
Path analysis or structural equation modeling is a technique for testing the consequences of proposed causal relationships among a set of variables. The technique rests on specific procedures and important assumptions (uncorrelated residuals, 1-way causality, linearity, additivity, interval measures). These assumptions, along with the problem of multicollinearity, are discussed, and specific techniques are offered to deal with them. Path analysis studies from the industrial/organizational psychology literature are reviewed to illustrate the specific consequences of disregarding these assumptions and procedures. The relationship between cross-lagged correlation and path analysis is explored, particularly with regard to the differing but complementary purposes of each. Finally, the problems due to both shared and random measurement error and the application of path analysis to experiments are discussed. (37 ref) (PsycINFO Database Record (c) 2012 APA, all rights reserved)
Describes a general data-analysis strategy for analyzing relationships among multiple data sets. This strategy may be employed in an exploratory mode, in a hypothesis-testing mode, or in some combination of the two. Furthermore, one may focus on the original variables or summarize these variables by a small number of principal components. For the case of two data sets, the output includes (a) the complete multiple correlation analysis of each variable (component) in one set regressed on all variates of the other set, (b) a canonical correlation solution, and (c) a final transformation to enhance substantive interpretation of the data. Emphasis is placed on examining information available at each stage of the analysis. The final solution allows the data matrix to be partitioned into elements due to interset factors, battery-specific factors, and an error component. (28 ref) (PsycINFO Database Record (c) 2012 APA, all rights reserved)
To improve the efficiency of selection techniques, a method is proposed in which individuals whose criterion score predictability—as measured by the difference between their standard test scores and standard criterion scores—is high, are screened first on a measure (such as an occupational inventory) that correlates with this difference, and then those so screened are further screened on the basis of their test scores. An example based on data for two samples of taxicab drivers, using a tapping and dotting test for prediction of criterion performance, and an occupational level inventory measure to identify the "predictable" individuals, is presented. (PsycINFO Database Record (c) 2012 APA, all rights reserved)
Controlling for variables implies conceptual distinctness between the control and zero-order variables. However, there are different levels of distinctness, some more subtle than others. These levels are determined by the theoretical context of the research. Failure to specify the theoretical context creates ambiguity as to the level of distinctness, and leads to the partialling fallacy, in which one controls for variables that are distinct in terms of appropriate theory. Although this can occur in using any control procedure, it is especially likely to occur in multiple regression, where high-order partial regression coefficients are routinely obtained in order to determine the relative importance of variables. Four major ways in which these regression coefficients can be seriously misleading are discussed. Although warnings concerning multicollinearity are to be found in statistics texts, they are insufficiently informative to prevent the mistakes described here. This is because the problem is essential...