These tests are nothing special. They ask "are you a thief" and most people, if they are thieves, will say "why, yes, yes I am".
Or they'll ask "Have you ever stolen anything, even if it was very small, from work?"
Thieves will say "yes".
Or they will ask "When was the last time you did coke while at work"
never, 5+ years ago, 1-5 years ago, within the last six months, within the last six weeks.
A druggie will answer something other than "never".
It's like when the cop pulls you over and you were speeding. He asks "do you know how fast you were going"; someone that was going 20 will say "two or three over", because they want to be a little honest. Similarly, your dumb ass will say "5+ years ago" because you want to be a "little" honest about how you're a coke head.
They will also include a "Liar" or "faking" scale or socially desirable answer choices vs. functional useful answer choices. Even more useful, though, is just saying "we know if you are faking answers"; which they don't, but will change the behavior of 30% of respondents.
This is the 'theory' that drives most personality tests
Big 5
Essentialy:
You want to be:
Open to new experiences
Extroverted
Conscientious
Emotionally stable
Highly Agreeable
Those are the "right" answers; you could get away with "wrong" answers in some jobs (introverted accountant): but even then why risk it?
You also need a "I command my fate" perception of self and NEVER say you won't do something because it's "not in my job description".
BTW
Interviews are the most bias least-valid method of choosing someone. It's worse than using hand-writing analysis (which is total bullshit, but at-least it's chicken-bones bad not 'justification for my implicit biases' bad)
That said, to ace an interview: 1) Repeat back what the interviewer says, use similar mannerisms and gestures 2) Sell yourself, talk about your good points at all turns.
Doing these two things accounts for nearly 60% of the variance in interview-based selection.
