An Integration of the Training Evaluation and Job Performance Modeling Literatures: Confirming BE KNOW DO with United States Army Special Forces Training Data

training;performance;job performance;training performance;direct determinants;skill;affective;cognitive;learning criteria;performance modeling;overfitting a model;SEM;confirmatory factor analysis;CFA;ARI;training proficiency;special forces

Surface, Eric Alan ; Dr. Mike G. Sanders, Committee Member,Dr. Don W. Drewes, Committee Member,Dr. Bert W. Westbrook, Committee Member,Dr. Don L. Martin, Committee Member,Dr. Mark A. Wilson, Committee Chair,Surface, Eric Alan ; Dr. Mike G. Sanders ; Committee Member ; Dr. Don W. Drewes ; Committee Member ; Dr. Bert W. Westbrook ; Committee Member ; Dr. Don L. Martin ; Committee Member ; Dr. Mark A. Wilson ; Committee Chair

University:North Carolina State University

关键词: training;    performance;    job performance;    training performance;    direct determinants;    skill;    affective;    cognitive;    learning criteria;    performance modeling;    overfitting a model;    SEM;    confirmatory factor analysis;    CFA;    ARI;    training proficiency;    special forces;   

Others  :  https://repository.lib.ncsu.edu/bitstream/handle/1840.16/5160/etd.pdf?sequence=1&isAllowed=y

美国|英语

来源: null

【 摘 要 】

Training data from 1441 graduates of the U.S. Army Special Forces Qualifications Course (SFQC) offered an opportunity to test a multidimensional model of training performance. A three-factor model based partially on the Kraiger, Ford, & Salas (1993) framework was operationalized with level-of-performance training criteria (Sackett & Mullen, 1993) and successfully confirmed using confirmatory factor analysis (CFA). The model utilized the BE KNOW DO terminology from the U.S. Army's leadership model to describe the cognitive, skill-based, and affective training outcome factors. Several alternative models were tested and found not to be identified, suggesting the Kraiger et al. (1993) version provided the best description of performance. Additionally, the BE KNOW DO model was successfully confirmed for two individual phases of the SFQC training separated in time.Therefore, structural equation modeling (SEM) analyses were conducted to determine the relationship of similar constructs over time. Each construct from the initial training phase was found to predict its counterpart in the later phase. The degree of relationship varied for the constructs, suggesting some were more influenced by time and situation. Two performance modeling issues—the specificity of performance constructs and the impact of overfitting a model to the idiosyncratic characteristics of the initial sample on cross-validation—were investigated as well. Results related to the specificity of modeling performance content were inconclusive. Both the one- and three-factor construct models failed to provide adequate fit. The over-modified model provided a worse fit upon cross-validation in 11 out of 12 cases, demonstrating the importance of cross-validating modified models. An integration of the training evaluation and job performance literatures is presented and serves as the rationale for proposing a general three-factor performance model. The idea that all performance can be described in terms of three factors regardless of the context, content, situation, measurement method, or performance level should be investigated. Future directions for practice and research are discussed.