Je noted a striking contrast between the
reliability and validity of personality questionnaires; their reliability has been consistently reported to be fairly high, but their validity has been generally found to be poor. We traced this divergence to the errors of self- estimate on which the questionnaire
responses ordinarily depend. A tendency to make flattering and fictitious responses was found to have been frequently suggested in previous investigations.It was presumed to account for the errors of selfestimate.
We noted that the tendency to fake responses
was not of necessity conscious and deliberate. It
also seemed to influence a self -rater sometimes without his knowledge. This suggested the operation of some persistent factors of distortion which affect the questionnaire responses in a uniform manner and thus account for their high self -consistency.
A survey of the allied literature suggested
two probable factors of distortion: (1) Lack of
insight or the capacity for self -knowledge; and (2)
The need for conforming to the social standards of
attitude and behaviour.
We failed to determine the nature of insight.
We could, however, find indications of the correlates of insight from. previous work. Those generally indicated ere abstract intelligence, liability to projection. and sense of humour.
We have reiterated many times that our aim in
this research was purely theoretical. The success of
our endeavour can be judged from the fact that we have been able to isolate two persistent tendencies which are at work in lowering tie validity of personality questionnaires. We have been able to describe some of their characteristics also and give them a rough operational definition. We have also devised some tentative tests for measuring these tendencies,some of which have proved quite suitable. There is no doubt that they need reformulation and a thorough refinement. Nevertheless,they may serve as a useful starting point.