The selection of officers to the Swedish defense forces was changed in the middle of the 1990-ies. A new computer-aided test battery was constructed for use in the selection process, as was a new personality inventory. This inventory, called Commander Trait Inventory (CTI), is at present subject to a series of validation studies. One of these studies is reported here.  Sweden has a long tradition of participating in UN peace-keeping operations. In the conflict in former Yugoslavia, Sweden has participated with a batallion in Bosnia since 1993. At present, changes in the selection of personnel for batallions in international operations are being prepared. The aim of the present study is to examine the correlations between CTI measures and measures of performance after completed international service.

A full description of CTI is to be found in its manual (Carlstedt & Widén, 1998). The reliability, expressed as Alpha coefficients, and the most characteristic statement for each of its eleven scales are presented in Table 1. The sample used was 1176 applicants for officer training in 1997.

All scales have a good homogeneity. Six of the eleven scales forming the Commander Trait Inventory are intended to assess cognitive style, having Jung´s theory of psychological types as a source of inspiration, but it was decided not to name them according to his theory. Instead, they have been named according to a comprehensive description of the content of the final scale.

The remaining five scales assess personality aspects presumed essential to the officer profession. Empathy and Leadership Motivation should be regarded as necessary but not sufficient qualities in leadership, while Egocentrism, Impulsivity and Ethnocentrism should be sufficient signs of unsuitability for the officer profession, mainly for ethical and moral reasons.

Confirmatory factor analysis (Jöreskog & Sörbom, 1993) was used in order to define and test an hierarchical measurement model, consisting of one "ideal employee factor" and four residual factors.

The model, with one general factor and residual factors is orthogonal, i.e. the correlations between the factors are close to zero, see table 3 below. Personality instruments yielding uncorrelated factors in selection situations are rare to find.

The following analysis was performed on data from applicants for the officer profession collected in 1997. The confirmatory factor analysis, accounted for in Table 3 below, shows an acceptable fit to data (GFI = .96, AGFI = .91, RMSEA = .09, chi² = 313.7, df = 31, p = .00).

All displayed factor loadings and factors are significant, with the exception of those in parentheses.

It is possible to find models with slightly better goodness-of-fit than the one presented above. The present solution, however, was chosen because of its simplicity and because it was judged to have a higher psychological credibility than the alternative solutions available. Validation studies against job-related criteria will show if the choice made can be justified.

High factor scores: For leadership optimal combination of different personality traits. High leadership motivation and high confidence in one´s fast intuitive decision-making. Empathic ability and ideological value orientation. A high score may indicate a socially agreeable response set and an obviously beautyfying self-picture.

Low factor scores: Cautious, bad self-confidence, not competitive, low interest for leading positions.

High factor scores: Structures life through rational and systematic thinking processes. Sorts experiences and facts in concrete categories or in terms of models, theories and value systems. This leads to an inflexible and "square" way of handling life.

Low factor scores: Attaches little importance to concrete facts or to systematizing and evaluating experiences.

High factor scores: Impulsive and little reflective. Strives constantly towards new sensory impressions and new experiences.

Low factor scores: Cautious and orderly.

High factor scores: Egoistic superficial docility and adaptability in limited social environments.

Low factor scores: Stubborn, uncompromising attitude towards the social environment.

High factor scores: Egocentric, ethnocentric and impulsive. Lack of empathy.

Low factor scores: Empathic and altruistic.

The correlations between factor scores on the five factors are shown in Table 4 below.

The correlations between the factors are lower than those between the scales. The factor solution can be said to be orthogonal. The correlations are so low that an individual´s profile can be interpreted directly to reveal his "true" values in the different factors.

The reliability of the factors cannot be calculated with traditional methods. However, Huang (1997) has presented a method that admits calculation of factor reliability, resulting in a measure called determinacy, which is shown in Table 5.