The Wall Street Journal (WSJ) has a "numbers" guy that just posted an article on how numbers can be very misleading.
He makes a good point, despite coming close to arguing himself out of a job :).
The good point is that the term 'statistical significance' has taken on such a aura in popular culture and business that anything claiming 'significance' from data must be true and important.
It's generally not - on either account. Large sets of data can produce all sorts of strange results and small sets all sorts of wrong conclusions.
Just something to keep in mind.
He makes a good point, despite coming close to arguing himself out of a job :).
The good point is that the term 'statistical significance' has taken on such a aura in popular culture and business that anything claiming 'significance' from data must be true and important.
It's generally not - on either account. Large sets of data can produce all sorts of strange results and small sets all sorts of wrong conclusions.
Just something to keep in mind.
No comments:
Post a Comment