There are way too many variables to call this study conclusive by any means.  Study participants walk into a TV studio lab.  When viewing television, did they simulate a real home environment with a family, kids, and a pet to serve as possible distractions during the ‘event’ of watching a TV ad?  Did they assume one channel only (re: no jumping from one program to another during a commercial break)?  As for magazines, did it assume that everyone flips page by page rather than skipping entire chunks of the magazine?  As for the Internet, were the ads shown on CBS and Conde-Nast sites?  What kinds of ads were shown?  How about the relative CPP across each medium?  Was the TV watched meant to be primetime?  Late night?  Were the magazines up-market, down-market?  And how about questionnaire?  What was the research design behind it?

I applaud any research for taking a look into ad effectiveness, but when it’s commissioned by two companies that make their mint on TV and magazines, is it any surprise as to what comes out on top?  For years, those of us in the Internet ad community (disclosure: the majority of our work is digital although we have serviced clients cross-media) have looked at smart targeting for addressability, something which print will lack at scale due to the cost base of producing a customized magazine and still a few years away until projects like Canoe and Simulmedia get to full speed.  The study seems to assume that the ads behind the web pages are untargeted and for untargeted advertising I’m not surprised by the study results.