Escapist News Now: Xbox One Beats PS4 In Black Friday Sales

[cthulhuspawn82]

I actually don't see how we can even have any good data about this contest. As far as I know, the consoles were selling out everywhere to this point, although they "may" be in stock now. How can you judge which console is "selling more" when both consoles sell every unit they produce.

 

[Vosgy]

Probably should tell people what a good sample size is.
For instance, you want at least 2% of all sales reported, and gathered in a non biased way.
For statistics this is a good base line to base your predictions off of (assuming the data set is very large 2% of 10 isn't going to work ).

 

[smithy_2045]

In other words, even if the sample was significant enough to mean anything (and at only 1500 console sales, it's not even close to significant), it immediately suffers from an inherent self-selection bias (because only those who willfully participate in the data collection are being included) that will distort the validity of the conclusion.

All other possible concerns about the data aside (I don't know or particularly care enough to check), but a sample of 1500 is more than sufficient a sample size to draw conclusions from. The only thing a larger sample size does is reduce the variation in the sample. The only question is at what level of significance does the difference become statistically significant.

Also, regarding the self-selection bias, so long as you specifically state that that any conclusions are made about people who would choose to participate, they remain valid. Dunno whether that happened here, but the point remains.