Monday, 26 May 2014

Never be afraid to ask "are you sure?"

The star economist of the year, Professor Thomas Piketty, who has topped the best seller lists with his 577 page 'Capital in the  Twenty-first Century', has been exposed by the Financial Times as having made transcription errors, used incorrect formulas and cherry-picked data.  It makes little difference whether, as he claims, his core argument stands up or indeed whether he has been lauded for his statistical work - the method was wrong and that undermines its credibility.

You may also recall the recent discussion on research on statins. Research published in the British Medical Journal “overestimated the side effects of statins by more than 20 times”. There are fears that those who stopped taking statins as a result of the original claims may have been unnecessarily exposed to heart attacks. It may be, as they say, 'academic', to some that the data was unreliable.  

In both these cases it would have taken a lot to have challenged the assertions of the authors but it turns out they were built on foundations of sand.

Of course this isn't the first time I have discussed these problems. You may recall the case of the correct numbers being inputed the wrong way round. I could go on but that may act as a distraction.

I would like you to recall these examples the next time you find yourself being pressurised in a tender evaluation by an expert - even the great and the good can make mistakes. But then again, so can you and I. Perhaps it's wise to ask "am I sure this is robust?"

No comments:

Post a Comment