We support the AllTrials campaign.
We’ve taken the decision to lend the support of evidencebased.education to the AllTrials campaign. The campaign “calls for all past and present clinical trials to be registered and their full methods and summary results reported”, an aim we find no difficulty supporting.
The so called ‘file drawer’ problem – the metaphor used for publication bias – has been well-documented for many decades (Rosenberg, 2005; Rosenthal, 1979; Scargle, 1999). Publication bias is a huge problem for those making clinical decisions; without a full picture of all trial results (those results finding positive, negative and no effects of clinical interventions), clinicians really cannot make fully-informed decisions for patients. The file drawer of every researcher should be empty, and all studies (particularly those funded with public money) should be made available.
Publication bias is not, however, limited to medicine; indeed concerns about the problem were raised in education and psychology research long before they were in medicine (Egger & Smith, 1998). Today, it is pleasing to note that the Education Endowment Foundation – the game-changing organisation funding much of the high-quality research (including a substantial number of randomised controlled trials) in England is committed to publishing all its education research results. It is only when organisations such as this take steps such as this that teachers and school leaders can have trust in research evidence; as Kevan Collins made clear at a conference[1] in January 2016, trust is absolutely paramount to developing a better, more functional relationship between researchers, teachers, school leaders and policy-makers. A major pillar of that trust is the publication of all trial results, no matter the conclusion.
We urge you to support AllTrials.
References
Egger, M., & Smith, G. D. (1998). Bias in location and selection of studies. BMJ: British Medical Journal, 316, 61.
Rosenberg, M. S. (2005). The file‐drawer problem revisited: a general weighted method for calculating fail‐safe numbers in meta‐analysis. Evolution, 59, 464-468.
Rosenthal, R. (1979). The file drawer problem and tolerance for null results. Psychological Bulletin, 86, 638.
Scargle, J. D. (1999). Publication Bias (The” File-Drawer Problem”) in Scientific Inference. arXiv preprint physics/9909033.
Endnotes
[1] Osiris Learning and Achievement Conference, 13/1/16, London
