ToxicJosh on 16 January 2015
I get OP's point about sample size, to a degree.
For example, you have 2 samples; one, non-representative, comprising 50% of the market. and the other, representative, but only 1%.
The first doubled, would be less accurate, than the second x100. As the first, despite being bigger, isn't an accurate representation of the whole market, whilst the second is.
However we've no reason to believe that not having Toys R Us makes NPD non-representative, they may well have other toy retailers that provide insight into that market segment. And we also have no reason to believe that VGChartz sample is more representative. It could well be less.
