The question is now whether the flawed earlier study resulted from the usual problems with broad epidemiological studies or if the study was designed to create the desired result. I'd go with a combination, providing the "scienticians" doing the thing with an alibi.
This shows an interesting way to get govt grants for research without having to prostitute yourself to the Bush Administration. When they let it be known that you'll only get the grant if you find that abstinence is the best solution, give them what they want. Then turn around and give your data to someone who'll put out an honest study.