The Great 8 as a Framework for Validation Research Helen Baron, Dave Bartram and Rainer Kurz 1 SHL Group plc, The Pavilion, 1 Atwell Place, Thames Ditton KT7 0NE Over the last 50 years the validity of selection measures has been the subject of constant research. There have been operational studies of individual selection procedures to help a particular employer make better selection decisions, generalisability studies at the individual instrument level to determine over what range of circumstances it may be effective and more theoretically oriented studies of the constructs underlying good performance prediction which these days are likely to take a meta-analytic approach. From early on, ability measures have been acknowledged as good predictors of job performance and even better predictors of training performance. The early meta-analyses (e.g. Hunter & Hunter, 1984) showed the generalisability of this finding. However at this stage there was a consensus, at least among academics, that personality factors were not good predictors of performance. This view was strongly influenced by Guion (1965) and supported by the close to zero findings of unsophisticated meta- analyses which looked in an undifferentiated manner at the correlations with job performance of any personality scale of whatever nature, or in whichever direction it was defined (e.g. Hunter & Hunter, 1984). This did not prevent many practitioners from continuing to work with personality measures and being convinced of their usefulness. More recently, the academic literature has begun to support this contention. Studies concentrating on the ‘big five’ measures, has shown that first Conscientiousness and then Emotional Stability can be seen to have broadly generalizable relationships with overall job performance (Barrick & Mount, 1991; Hough, 1992, Salgado, 1997). While these factors do predict overall performance well enough to be very useful in selection, and incrementally over cognitive tests, they account for a relatively small part of the overall variance. Efforts to find other predictors of overall job performance have floundered. Given the variation in job demands, from individual to group oriented work, from physical to largely mental activity, from repetitive to constantly changing, from competitive to cooperative environments, it is reasonable to wonder whether much of the remainder of the variance might be explained by more job or situationally specific variables. Indeed from an ethical perspective it is to be hoped that there are different factors predicting success in different jobs, since the corollary of standard predictors is that people who are low on these factors do not have the capacity to be 1 The work reported in this paper was carried out while all three authors were with SHL Group plc. Helen Baron is now an independent consultant and Rainer Kurz is now with Outstanding Achievements