Nothing new with replication

In the November issue, Simon Robinson asked to be credited with the idea of a British Journal of Replication Psychology. Sorry to disappoint, but he's 37 years too late.

Following on from the Cyril Burt debate at the British Psychological Society's Annual Conference in Aberdeen in 1981, I published 'Avoiding Scandal' in The BPS Bulletin (Vol 34 p.127-28) which detailed 11 ideas to prevent future fake/faulty research being published. Amongst the suggestions I made was for a Journal of Non-Significant Results and Replications.

Other ideas included limited tenure for journal editors, lodging all data with an independent source, a taught PhD in Constructive Criticism (with journals making space for reviews of previous articles), review of any complicated statistics by a competent/independent statistician, and a voluntary ban on publication in journals edited by a colleague (to prevent departmental nepotism, conscious or unconscious).

I also took a swipe at the curse of 5 per cent and 1 per cent probabilities rather than meaningful differences, and the assumption of standard bell shaped distribution. I hasten to add that all the ideas/suggestions arose out of discussions at the Conference.

Ewen Rennie
Retired Senior Educational Psychologist
Aberdeen City Council