The Journal of Socio-Economics 33 (2004) 615–630
Statistical reform in medicine, psychology
and ecology
Fiona Fidler
a,b,c,*
, Cumming Geoff
c
,
Burgman Mark
a
, Thomason Neil
b
a
Environmental Science, School of Botany, University of Melbourne, Melbourne 3010 Vic., Australia
b
History and Philosophy of Science, University of Melbourne, Melbourne 3010 Vic., Australia
c
School of Psychological Science, La Trobe University, 3086 Vic., Australia
Abstract
Over-reliance on Null Hypothesis Significance Testing (NHST) is a serious problem in a number of
disciplines, including psychology and ecology. It has the potential to damage not only the progress of
these sciences but also the objects of their study. In the mid 1980s, medicine underwent a (relatively)
major statistical reform. Strict editorial policy saw the number of p values in journals drop dramatically,
and the rate of confidence interval reporting rise concomitantly. In psychology, a parallel change is
yet to be achieved, despite half a century of debate, several editorial inventions, and even an American
Psychological Association Task Force on Statistical Inference. Ecology also lags substantially behind.
The nature of the editorial policies and the degree of collaboration amongst editors are important
factors in explaining the varying levels of reforms in these disciplines. But without efforts to also
re-write textbooks, improve software and research understanding of alternative methods, it seems
unlikely that editorial initiatives will achieve substantial statistical reform.
© 2004 Elsevier Inc. All rights reserved.
Keywords: Statistical reform; Confidence intervals; Medicine; Psychology; Ecology
Over 30 years ago, Morrison and Henkel (1970) commented on the “parallel but quite
independent scrutiny” (p. 182) Null Hypothesis Significance Testing (NHST) had undergone
*
Corresponding author. Tel.: +61 3 8344 4405; fax: +61 3 9347 5460.
E-mail address: fidlerfm@unimelb.edu.au (F. Fidler).
1053-5357/$ – see front matter © 2004 Elsevier Inc. All rights reserved.
doi:10.1016/j.socec.2004.09.035