Out here the answer isn't a number — it's a decision, and you have to defend it with the data. Two skills finish the job: choosing the better option when the numbers are messy, and knowing what data you'd need when you don't have enough yet.
Picking the Better Option
Here's the trap everyone falls into: a higher average isn't automatically the better pick. Imagine choosing a goal shooter for a final. Both players average 8 goals — a tie on the headline. But one is rock steady (her scores barely move, a tiny range) while the other swings from 2 to 15. For a one-off final you want reliable, so you start the consistent one. If you were desperate and needed a slim chance of a massive haul, the wild one might be worth the gamble. The right pick depends on what you're actually deciding — and you justify it with both the centre and the spread.
Other times the call is cleaner. If one phone averages 9 hours of battery and the other 6, with the same consistency, the higher one simply wins. The skill is knowing which kind of situation you're in: is the gap in averages big enough to decide it, or is spread the thing that really matters?
Working It Backwards — What Data Would You Even Need?
The cleverest exam questions flip it: they give you a decision and ask what data would let you make it fairly. A shop wondering which drink to stock more of needs sales counts — and ideally over several months to spot a trend — not the average price or the manager's favourite flavour. To claim a revision method raises marks, you need a big enough group who used it and a comparison group who didn't, or you can't tell the method from luck. Naming the data that actually answers the question — and binning the data that doesn't — is the deepest form of this skill.
Why This Is the Finish Line
Rung 1 showed you that one number hides the story. Rung 2 gave you the centre-spread-shape recipe. Rung 3 taught you not to over-claim. Here it all pays off: you read real data, weigh it honestly, and either make the call or say exactly what you'd need to make it. That's what statistics is for — and it's the doorway to designing your own investigations, next.