Document Type
Article
Publication Date
2005
City
Dayton
Abstract
The present paper deals with the problem of data integration in the context of aviation psychological assessment. In the first pilot study 99 pilot applicants completed a comprehensive test battery. The general judgment of the candidates’ performance in a flight simulator served as an external criterion. To examine the predictive validity of this test battery, both a discriminant analysis as well as an artificial neural network were calculated and compared with each other with regard to classification rate, stability, and their respective differentiability of suited and not suited applicants based on their success probabilities. The results of this first pilot study demonstrate that artificial neural networks outperform classical methods of statistical judgment formation with regard to classification rate and differentiability of suited and not suited applicants based on their success probabilities. In the second study 264 applicants for the position of a commissioned officer in the air force completed a smaller test battery, which also included measures of personality traits. The general judgment of the candidates’ performance in a flight simulator served as external criterion. To examine the predictive validity of this test battery, a discriminant analysis as well as a logistic regression analysis and artificial neural network were calculated and compared with each other with regard to classification rate, stability, and capacity to separate correct and incorrect classifications. The results of this second study replicate the finding of the first pilot study by demonstrating, that artificial neural networks result into higher classification rates and a better differentiability of suited and not suited applicants based on their success probabilities. Based on these results it is concluded, that artificial neural networks provide a valuable tool for the selection of pilots which increases the objectivity and precision of diagnostical judgments derived from standardized test batteries.
Repository Citation
Sommer, M.,
& Schuhfried, D.
(2005). Validation of a Modern Aviation Psychology Test Battery Using Aertificial Neural Networks: First Results of Two Pilot Studies. 2005 International Symposium on Aviation Psychology, 694-699.
https://corescholar.libraries.wright.edu/isap_2005/118
