Building: Pinnacle
Room: 2F-Harbourside Ballroom
Date: 2016-07-02 11:00 AM – 12:30 PM
Last modified: 2016-05-22
Abstract
Introduction
General Mental Ability (GMA) is according to meta-analytical research the single best predictor of work performance. The question arises how speed and accuracy in answering questions across various ability areas underpin test performance and validity.
Objective
The present study investigates the psychometric of sub-scores of a short measure of GMA.
Method/Design
Psychometric studies were undertaken involving Logiks General which measures GMA through 24 Verbal, 16 Numerical and 10 Abstract problems in three separately timed sections of 4 minutes.
Results
Internal Consistency reliability on N=177 was satisfactory for the Total score (.76) and slightly below the .70 level required for stand-alone interpretation for Verbal (.64), Numerical (.62) and Abstract (.57) sub-tests. Same day completions (N=138) resulted in an adequate correlation of .766 for Total, .801 for Speed and .757 for Verbal sub-scores. Accuracy (.681), Numerical (.660) and Abstract (.599) fell slightly short. Test-Retest correlations (N=423) were adequate for the Total score (.741) whereas values for Speed (.548), Accuracy (.591), Verbal (.656), Numerical (.608) and Abstract (.489) fell short of the benchmark for stand-alone use.
Several construct validation studies were carried out. On N=762 Business School students the Logiks General Total score correlated highly with longer RfB Verbal (.492) and Numerical (.582) unsupervised tests. All sub-scores also correlated significantly. N=185 participants completed Logiks General and the RfB Business Support Numerical assessment in French or Dutch. Total score correlated .715 and all sub-scores at least .446.
In a concurrent validation study (N=123) Total, Speed, Accuracy, Verbal and Numerical correlated significantly with reviewer ratings of ‘Analysing and Solving Problems’, ‘Thinking Creatively’ and  ‘Conveying Credibility’ competencies. For the Abstract section Accuracy reached statistical significance.
Conclusion
The results suggest that all sub-scores contribute towards the validity of the overall measure and can e reported and fed back in the context of the total score interpretation.