For a recent experiment, we divided respondents into two groups. To both groups, we showed a trailer for an upcoming movie. One group was given two different descriptions for the trailer (as if there were two different trailers) and asked to chose one or the other. We have found that the presence of this "choice" had a significant effect on recall of key facts from the trailer. We are doing some follow-up work now to clarify a few things, but what we found was pretty interesting.
My colleague Rob St.Louis and I are going to the ARF's Audience Measurement conference in New York in June to present a paper with the results. Come say hi if you are there. Our thing starts at 1.50pm on Monday June 11 in Majestic on the 6th floor of, I think, Marriott Marquis.
Oh, and we have also just published the results of another study that showed how fiddling with smartphones distracts people from TV and what could be done about it