A year in statistics—the view from the trenches

Abstract

Last year we launched the EBM Methods Verdicts as part of an attempt to keep on top of developments in statistics that will impact on the way we do research.1 2 We focused on statistical methods and in particular on the output from five different journals: Journal of Clinical Epidemiology , Statistics in Medicine , Statistical Methods in Medical Research , BMC Medical Research Methodology and Biostatistics . The idea is simple: identify papers that we believe are likely to change the way we analyse or report evidence syntheses, randomised trials or studies related to diagnosis, prognosis or monitoring. Our editorial team typically focuses on one of these per month and aims to write a short commentary with a ‘verdict’ explaining how this will change how we do things. We are also interested in people submitting proposals for similar short pieces. Bayesian neural networks were the rage in the early 90s. Work in this area seemed to go quiet for most of the 00s but has exploded back into our consciousness in the last few years. This is clear by the number of papers published on the topic (figure 1) with nearly half of the total appearing in the last 2 …

Ähnliche Arbeiten