Lies, damned lies and……….

The well-known quotation “There are lies, damned lies and statistics” suggests that statistics doesn’t have the greatest of reputations. Why is this? Certainly, statistics can be used to mislead, but logically “lies, damned lies and statistics” reflects high levels of statistical misunderstanding rather than indicating that statistics are inherently ‘bad’. In Bad Science, Ben Goldacre highlights the incorrect presentation of statistical facts and figures provided in newspapers and television and attributes this to their authorship by humanities students who lack a science background. This is not surprising given that science students and clinicians also find statistics difficult to understand! The emergence of evidence-based medicine (and evidence-based optometry) means that clinicians need at least a basic understanding of statistics to be able to reasonably determine the validity of published research and apply it to patient care. Unfortunately, studies have reported a low level of knowledge of basic principles of research methods and statistical analysis among doctors. Even researchers (and reviewers and editors) can show a lack of understanding of statistical issues and a growing number of published papers are highlighting errors in statistical design and/or analysis of research studies that make it through the review process and are published in the biomedical and ophthalmic research literature. Of course, some of these may represent differences of opinion between statisticians, and others may include the (mis)use of statistical analyses that best help publication (for example,) given the academic pressure to “publish or perish”, but these seem relatively minor factors. The developments in evidence-based practice highlight the need to develop statistical literacy amongst clinicians and undergraduate curricula need to change if appropriate modules are not already present. In addition, it seems important for postgraduate programmes to include modules in principles of research design and statistics. Finally, given that statistical understanding continues to advance, researchers need to keep abreast of current knowledge and this is perhaps where research journals can help.