There was data shared about golf scores that got me thinking about stats and populations.
Historically it’s been said about 10% of all golfers score less than 100. And that’s about right from my anecdote evidence. Now recently a stat was published for golfers officially recording their scores showing about 60% score less than 100. From my experience, that also seems correct.
So how can both be right? Well it’s based on the surveyed populations. The first stat includes all golfers while the second stat only counts golfers paying a membership fee to enter their official scores. So the populations are very different.
This is true for all statistical reporting. Look, with a surprisingly small sample of a large population, you can have very accurate results (recall all those stats that say "plus or minus 3%").
So when reviewing stats, be sure to always consider the population surveyed - what’s included and what’s excluded. And when you generate stats, clearly define your surveyed group for your readers.
That's my 8020 Framework. What's yours?
Комментарии