|Geoff Cumming (Australia)||G.Cumming@latrobe.edu.au|
|Null Hypothesis Significance Testing
(NHST) is still the dominant statistical technique in psychological research,
despite evidence of its damaging effects, and cogent and sustained advocacy
of alternative methods by some distinguished scholars. However, substantial
change is now likely, especially with official support from the American
Psychological Association (APA) in the form of recommendations in the forthcoming
edition of the influential APA Publication Manual. Reform priorities include
a swing from inference to estimation, and wider use of effect size measures
and meta-analysis. Achieving these highly desirable reforms will require
re-education and substantial attitude change on the part of current researchers,
and drastic revision of statistics education within psychology.
This paper reports selected parts of a research program intended to guide and support these changes. Technology is used in a number of ways in the research, and software tools are seen as central to the reformed education needed within psychology. Projects to be described briefly include:
A case study of an attempt by a journal editor in psychology to encourage reporting of confidence intervals (CIs) and their use for interpretation. Authors were generally reluctant to change and many difficulties emerged with presentation and understanding of CIs.
A case study of an attempt by a journal editor in psychology to require authors to report effect size measures, including results of an email survey of authors.
An internet-based study of researchers' ability to interpret graphical representations of CIs for simple experimental designs.
A primer, with computer tools, designed to explain to researchers and students how CIs can be calculated and used for simple standardised effect size measures.
Development and evaluation of the ESCI (Exploratory software for confidence intervals) tools designed to promote understanding, calculation and use of confidence intervals.
Proposed Rules of eye (ROIs) to guide interpretation of CIs and other error bars shown on graphs of means.
Psychology is at present in the strange situation of being about to adopt substantial changes to statistical practice, but with scant evidence to guide how the reforms should best proceed. For example wide use of CIs may be desirable, but they are poorly understood in psychology and so far there has been little empirical investigation of how they can best be represented, interpreted, and taught to current and future researchers.
|Download in Adobe Acrobat format (137 Kb).|
Back to top
ICOTS-6, The Sixth International Conference on Teaching Statistics - International Program Committee (IPC) Website.
Copyright © 2001 by the IASE. All rights reserved. This information is subject to change without notice. This page was last modified on July 2, 2002.
For questions or comments, contact the Webmaster, Dagan Ben-Zvi.