Traditionally, cross platform titles sold better because of the larger base -- especially on the PS3, but this phenomenon doesn't come across strongly in this study. It could mean that the 1000 - 2000 gamers they surveyed are not representative. It's hard to get the right proportion for Wii gamers vs 360/PS3 gamers though (How do they divide the population ? by visits to the stores ? Which stores ?).
EDIT: Actually, if the analysts don't double check their result against known trend and at least explain the discrepancy, then we should probably give them more time to refine their techniques.
Actually if you're speaking about the difference about Nielsen institute and gameplan it will be difficult to discuss as the news I read gives no references for the Nielsen study (outside that it's a really serious institute). And When I say no references it's no references, we don't even know if they consider Us. north america or the whole world. Gameplan is obviously US/North america by looking at madden figures. I may search to find out from where the Nielsen figures have been leaked, may be we will have better informations.
I don't agree in regard to discrepancy, if you look at gameplan figures moder warfare 2 is actually in second position (432 on 360 / 245 on ps3). I don't bother checking other titles but GH5 is hight in the chart if you add Wii, ps3 and 360 .
Out side of this basically having a representative is the basic of a marketing study/survey, it's not easy I think it's even the toughest part of the job. But I would not say that it's impossible or that video games are tougher than some others sectors. I don't get why they would focus their analysis on "gamer" the panel could be made as any panel which is a blend of various family (different ages for parents/kids/numbers of kids/revenues), young adults. Then there is how they ponder the results from the people from their panel. They coud even use a pretty standard panel representative of north america society and the mystic could be in how they ponder their results.
My guess is that 2000 is not enough for a country as US or North america thus they must use a custom panel and then ponder their results. Anyway if they think that their estimations have any value they won't public state how they do their in-house sauce.
But clearly while making the panel I can't see them focus on gamers, the point is to have something representative of the society thus is come down to calling a lot of people (more than 2000) and ask them if they want to be part of the "poll/survey" then ask random questions about family situation, revenues, interests, etc. not focus on gamers (which may not be representative their are a part of the market) if I were to do it I would not even focus on video game at first.
Then you sort the result, do your magic pondering and you come with a panel of 2000 persons.
Between 2000 persons can be missleading as a family of four can't be count as four as depending of the kiddies ages (it's mom or dad who have the buying power) or if teenagers have some buying power.
Basically then you call/e-mail back, you accepted a while ago, blablba, a survey, doing a study about video games I've a list blabla, please send me back the result.
Basically once kids/teen have answered with daddy, teens with money, young adults/adults have answered too you must ave something like what they call "awareness index"
lol: can help it but I always think of Jean Claude Vandamne when I read aware... such a freak...), I understand it as the raw result of the survey.
It looks like they have info about preorders (may be not taken from the survey) I guess is that it's use to fine tune/ponder the raw result.
Then the final result comes form the result of the first contact with the people (where you learn about revenues, numbers of kids, how money in gifts, etc. etc.)
Then you ponder the awareness about the product with the various economical factors you managed to gathers about the people.
It's basically marketing 1o1, but it's a lot of work to make the panel properly (most of the work)