Thursday, 28 August 2014

Blowing my own trumpet and casting scorn on others... again

Today in The Guardian, the Daily Telegraph, Newsweek and elsewhere, there is a report of a scientific experiment at a US university to see if applying magnets to the brain improves memory.  The results are published in Science magazine.  
So it must be true then.
Well, not so fast.
Firstly, there were 16 treatment subjects.
Yes, 16.  A miserably small number.
The improvements were "20 to 25%" over the control, but only lasted 24 hours.
We don't know how many other papers on the same topic were not published due to failure.

We did something with memory in India.  We took 100,000 women.  That's a sample size of 6000 times bigger.
We made most of them literate.  They learnt 50 letters of the Hindi alphabet and how to string them into sentences.  The effect lasted a lot longer than 24 hours.   (We don't think literacy fades much if they keep practising.)  Our control group - the other hundreds of millions that we didn't put on our course - didn't get literate.  So the improvement was from zero words a minute reading ability to say an average of 10 words a minute.  That's 10 divided by 0 is.... an infinite improvement!

No comments:

Post a Comment