Originally Posted by Midpack
Clever, but how do you distinguish an anecdote from one data point?
People present anecdotes as conclusive here all the time, giving 'you should retire early because "I" know someone who died young' as a reason to retire early is abused here often...
Maybe I misread those, the people giving retirement reasons usually give them in context why it motivated them. Whether that's statistical relevant in some way is not really in play. It is relevant for that particular individual.
To go from anecdotes to data (and draw broad conclusions) you'll need a decent random sample and double-blind trials, with filtering out effects. Anecdotes can give you a clue which hypothesis to test, but not the test itself.
Haven't read the paper, but a rather small 2 point difference means any individual observation gets lost in the noise. IQs can easily differ 30 points between healthy siblings. Makes me wonder though if it's genetics (second and later children have older mothers on average) or nurture.
It was more surprising to me that temperament shows no ranking order bias, another finding cited in articles covering that study.