The aim of our study was to investigate what are the real objectives of the validation scales.

The sample consisted of 80 subjects (40 non-psychotic patients and 40 graduate high school students) to whom the Amoral-15 scale measuring antisocial tendencies was administered in two different situations - with the standard instruction (E1) and with the instruction to fake the responses (E2). The difference between E1 and E2 was taken as a measure of the ability to fake well. We correlated differential score (E1-E2) with the validation scales (the mix of Marlow-Crown and Eysenck's L scale from the EPQ). NEO-PI-R was administered under standard instruction to measure basic personality dimensions.

Results indicated no significant correlation between the ability to fake good and the scores on validation scales. In some items we found negative correlation which means that subjects who better present themselves in social desirable way have lower scores on the validation scales. By Hierarchical regression analyses we found that about 40% of variance on validity scales could be explained by Emotional Stability, Agreeableness and Conscientiousness and not by ability to fake food.

Socially desirable answers are serious problem in personality assessment and cannot be solved by giving the validation scales because these scales are measuring the basic personality dimensions and not the ability to fake good.