I rarely take stock in these things. Usually its hard to get a hold of the actual study without paying some academic journal for it, which means the methodology and original conclusions are going to get filtered through whoever reported on it, usually amounting to little more than "multi-tasking good" or "multi-tasking bad".