In the world of advertising, research is no different from creative work -- some of it is excellent and powerful, and some of it is worthless and dangerous.
My experience has been that market researchers are good at counting, and very little else. So if you want to know "how many", you can hire a research firm and be pretty confident you'll get an accurate result. But if you want to know, "why" or "how" or "which one", look out. This is particularly true of copy testing.
There is an advertising blog called Copyranter which I like to read. It is foul, libelous, largely incomprehensible, and often hilarious. Copyranter responded to a Business Week columnist who was confused about the meaning of a study about advertising effectiveness in the following way:
"Just like every other study that's been done to figure which ads "work," this one means absolutely nothing."
SUBSCRIBE TO THE AD CONTRARIAN