The use of psychometric and other assessment centre measures in predicting performance on a naval command course

The Admiralty Interview Board (AIB) is the Royal Navy‘s assessment centre whose role is to select young people for officer training. The two aims of the study were (1) to investigate the relative value of psychometric versus other assessment centre selection measures and (2) the value of these and other approaches for selecting naval commanding officers for practitioners. The AIB selection data was used to investigate the long-term prediction of some of the selection measures, particularly the psychometric tests, in predicting the outcome for students attending the Submarine Command Course. Few pieces of research have looked at the long-term prediction of a real command situation. This research examines the prediction of a practical naval command situation where the student has to make rapid decisions under pressure and where failure to make the correct decision could be costly. A literature review showed that whilst cognitive tests, personality inventories and other assessment measures can predict job performance and training successes, the meta-analytical techniques used to pool research studies have produced inconsistent findings that could confuse practitioners. The students attended the command course, on average, thirteen years after the initial AIB selection process. Selection scores were available for 93 students, 57 of whom also completed a 'Big-Five' personality inventory and an Occupational Stress Indicator (OSI) at the start of the 24-week course. The average age of the students starting the course was 32. The students were assessed throughout the course and were graded as pass or fail. They were also given an A to F Course Grade. In addition, 88 students were graded on twenty aspects of performance covering eight tactical grades, three administrative grades and nine personality grades. The results showed that the means of the pass and fail groups on the AIB Non-verbal test were statistically significantly different with a moderate effect size. The correlation for this relationship (with the A to F Course Grade) was 0.20. The Non-verbal test score also correlated with the course instructor‘s grades on tactical performance at 0.30 and the Verbal test correlated 0.23 with the administration grades. None of the other AIB selection measures showed significant results. While this is a disappointing result, the students were a very homogeneous group and to obtain these findings for the Non-verbal and Verbal test after thirteen years shows the predictive power of these tests. Although the findings may be of theoretical interest the low correlations mean that not much variance in performance is explained. The tests would not be a useful screening device to reduce the failure rate on the course because there would be too much misclassification. None of the Big Five personality scales predicted success on the course or the other course grades but statistically significant differences were found for the means of two the OSI scales: these were for 'Ambition' and a Type A Behaviour measure. Ambition was the only scale which correlated significantly with the A to F Course Grade at 0.43. Ambition was also found to correlate with the total score for the twenty performance grades, the tactical grades and the personality grades and several individual performance grades including Practical Ability, Leadership and Command Presence with correlations approaching 0.4. Further research on this aspect of behaviour may be worthwhile. However, there are lessons to be learned. The literature review shows that practitioners need to scrutinize journal articles and book chapters on the validity of selection measures extremely carefully. It may be that measures which have been shown to predict the performance of junior staff are inappropriate for the selection of more senior staff with similar job experience.