At the risk of starting a stats nerdfest….
They didn’t actually sample 100 riders and record what kit they were using. They recorded which forks 100 riders were using, then recorded what shocks the next 100 riders were using, etc etc.
So the overall sample size was larger, but you can’t really draw comparisons between the groups…. ie: forks/shocks, shocks/frames or wheels/tyres, because the populations were different. For example 3 people had DVO Jade Coils in the shock categories. Those 3 people could have also had DVO forks – but they weren’t sampled for forks.
Also, depending on how quickly riders were “cycling” through uplifts – some of the populations might overlap….. ie: the riders sampled in each group were neither unique or identical. So in some categories there may be correlations between groups, but not all. For example: A bike (frame) might ship with an particular OEM component, but because your population is neither unique or identical, you have no way of knowing if that resulted in a corresponding number of riders using that component.
Not sure if that makes it statistically more representative or not…. we would need some stats nerds for that 😀