Estimates of fakeability on the Eysenck Personality Questionnaire

The fakeability of scales on the Eysenck Personality Questionnaire was estimated in a homogeneous sample of female undergraduates. Three instruction conditions were used: standard or honest, fake good, and fake bad. Various estimates of fakeability were used and the results supported Gordon and Gross’s (1978) position that at least 2 statistics be used so as to estimate both parameters, σ_D2, µ_D2, of the discrepancy distribution or of faking. The questionnaire appeared to be moderately susceptible to faking, at least in the fake good condition; however, the use of cutoffs on the lie scale should be considered to discard respondents.