CONSTRUCT-VALIDITY OF THE SWEDISH ENLISTMENT BATTERY

The construct validity of the Swedish Enlistment Battery (SEB) used fo r classification of conscripts according to intellectual capacity was examined. The covariance matrices of three samples of conscripts (N = 501, 1058 and 1057) were analysed in Study 1, testing the hypothesis o f one general factor influencing all the four subtests and one residua l factor General visualization (Gv') influencing two of the subsets. T he fit of such a model was good. In Study 2 seven reference tests were added to accomplish better definitions of the factors and also to be able to identify a crystallized intelligence (Gc') factor. Subjects we re 113 national servicemen. A general factor was identified, accountin g for the dominating part of the total variance. Two factors orthogona l to the general factor were identified; one crystallized-verbal (Gc') and one general visualization (Gv') factor, both accounting for a sig nificant but low part of the total variance. More tests must be develo ped and brought into use in the Enlistment Battery in order to add to the variances in the Gv' and Gc' factors. This is a prerequisite for h igh determinacy and high differential predictive validities of the fac tors.