When asked a control question, the examiner wants you to lie so they can measure the "response", and then compare it to your response to relevant questions.  If you react more to relevant questions than controls, you fail.  So you want to lie on control questions (as the examiner expects) but artificially RAISE your reaction.

This pertains only to CQT formated tests

Check out "the lie behind the lie dectector" under countermeasures, for more info.

TC

P.S.  If you DO use countermeasures, just realize that you are a butt twitching traitor, and a huge threat to western democracy!   

Thats clear. Thank you so much for your quick respoce.

Here's another tip.  If there is one question that you can't classify as either a control or a relevant, ask for clarification on that question, get him to rephrase it, et cetera.  Once you're squared away on that question, before running the test the polygrapher will bring up another question.  If that question is a control, then the question you weren't sure about and brought up is a relevant; if the question the polygrapher brings up is a relevant, then the question you weren't sure about and brought up is a control.

Why does this work?  Simple.  By you calling attention to a question, you will have a heightened reaction to that question, whether or not you're lying when you answer it.  They can't prevent this heightened reaction, but they can counter it by heightening your reaction to another question.  If you bring up a control question (and therefore heighten your reaction to it) they will need to counter that by heightening your reaction to a relevant question.  And vice versa.

Of course, if there are two questions you can't figure out and you bring up one and then the polygrapher brings up the other you may not be any better off, but you do know that they're not both relevants or both controls, which may be of assistance.

This technique should also be added to the next edition of TLBTLD.  In my humble opinion, anyway.