Can job applicants fake personality tests? This topic is one of the hottest in assessment research these days (in fact, there's a whole book about it), and there's quite a bit of research indicating these tests can be "faked", meaning people can figure out how they should answer given the job. But most research compares applicants to incumbents or asks the same subjects to be honest then "fake good" (with some exceptions)--this may or may not be the best way to investigate "faking."
Now, one of the editors of the above-mentioned book, Richard L. Griffith, has co-authored a new study titled "Do applicants fake? An examination of the frequency of applicant faking behavior."
Why is this study different from previous ones? Because they used a within-subjects design, meaning they compared scores of people taking exams as applicants to scores of the same individuals at a later point in time.
What did they find? That 30-50% of the folks elevated their scores when applying. In addition, this elevation resulted in significant changes in rank-ordering, which impacted hiring decisions.
The $400 million question that arises out of this is: Does it matter? I've talked a little bit about this before, but the major research done in this area (such as this and this) indicate it doesn't make much difference. Personality tests can still be very effective in predicting job performance--sometimes impressively so. Nevertheless, many people are still highly skeptical of personality tests. This may be one of those situations where some decision-makers are comfortable with the risks inherent in using the tool but do so anyway because of its proven value, while others simply don't want to go there. In reality, the same could be said about practically any assessment method.
My advice? Look to the job analysis. If personality factors are indicated as a major influence on job performance, don't ignore personality tests as a possible tool. There's too much riding on each hiring decision. Still interested, but not comfortable? Try giving a personality test as a research instrument (i.e., don't consider it for hiring purposes) and gather post-hire performance data--you might be surprised at how accurate it predicted success.