>If they own a dSLR, they're more likely to be good at composing photographs.
That's a very valid point, selection bias is the most-missed bias in studies of every stripe. But they're not beginners:
>And we also found similar numbers looking only at people who uploaded all three types of photos.
They didn't show a graph of that data, because this is an article rather than a formal study. Getting similar results in an intersection study is an excellent way to show that selection bias was at least not overwhelming - depending on the sample size, it might not even be significant.
The conclusion they made was pretty well supported, and seems like something you would agree with - the details they give are quick guidelines, but it boils down to "photographic skill matters".
Comments
>If they own a dSLR, they're more likely to be good at composing photographs.
That's a very valid point, selection bias is the most-missed bias in studies of every stripe. But they're not beginners:
>And we also found similar numbers looking only at people who uploaded all three types of photos.
They didn't show a graph of that data, because this is an article rather than a formal study. Getting similar results in an intersection study is an excellent way to show that selection bias was at least not overwhelming - depending on the sample size, it might not even be significant.
The conclusion they made was pretty well supported, and seems like something you would agree with - the details they give are quick guidelines, but it boils down to "photographic skill matters".
> similar numbers looking only at people who uploaded all three types of photos
Um.. we’re talking about the skill of the photographer here, not the skill of the model.