An advantage of the one used in the present study is that both construct and predictive validity of the instrument are demonstrated (Arvonen 2002).Behaviour description questionnaires devised to study leadership have been questioned by several researchers since this tradition began. The main claim has been that they measure the attitudes of subordinates towards the leader and not real behaviour. The research group at Ohio met the critique by presenting
studies which showed significant correlations between
subordinates who described the same leader (Ekvall
1992). The CPE questionnaire has been tested in
relation to influence of attitude. The change and production
dimension showed no correlations with attitude.
The employee dimension had a medium size
(0.40) coefficient, which is logical, because being
accepted and liked is a psychological drive in relationoriented
behaviour (Ekvall & Arvonen 1994). The
dropout rate was normal for this type of study and the
dropout analysis showed no selection bias.