Letter: Some of this is false

From Charles Nightingale

There are even more disturbing aspects of the contemporary research scene than John Ioannidis describes (16 February, p 44). As he says, many studies declare their results significant if the probability that they could have got the same results by chance is no more than 1 in 20; and there is bias against publishing negative results.

New but false ideas may start to be accepted on the basis of the randomly favourable results – of which there will be many in popular areas of research.

In areas where financial or ideological influences are at work, even on an unconscious level, desired results may eventually be obtained by chance.

It would be simple enough to remedy this by introducing a worldwide register of studies, so that those that produce positive results could be seen against those which were not published. Unregistered studies could be discounted.

Felixstowe, Suffolk, UK