Prediction Differs from Description: General Versus Specific Intelligence Testing for the Selection of Ab Initio Air Traffic Controllers

Meta-analyses and recent large-scale primary studies concerning the importance of intelligence indicate that for almost all jobs general mental ability (g) alone predicts performance well. However, there is a controversy concerning the question whether specific abilities (s) are needed to predict job or training performance. In the present study performance test data of 5223 applicants from the DLR program for selection of ab initio air traffic controllers at DFS Deutsche Flugsicherung GmbH were analyzed. Results of different approaches based on exploratory factor analysis and structural equation modeling to estimate g- and s-intelligence components were compared. In addition to this, the criterion-related validities of different g- and s-measures were tested using training performance criteria from 282 DFS trainees. It is argued that the preference for an intelligence model depends in part on the utilized theoretical approach and the objectives of the diagnostician (description vs. prediction)