ABSTRACT
Imagine a lecture theatre of 100 students. I give half of them tea and the other half coffee; the 50 who receive coffee all die. Which statistical test should I use to determine whether the coffee was poisoned? The answer is none, as the answer is obvious. But what if I decided to try and improve my teaching, as 10% of students fell asleep in my statistics lectures? I revise my notes, slides, jokes, etc. and find 6% fall asleep during my next lecture. Is my new lecture better? I don’t think the answer to this question is obviousthe variation (a reduction from 10% to 6%) may be due to chance or it may be caused by the new teaching methods. We are uncertain-I say it’s my brilliant new lecture, you say its chance-how can we decide? It is here that statistical tests come to the rescue.
