Point of View
Michael Bracken is a professor of epidemiology at Yale University and co-author of a study published in the British Medical Journal that looks at the usefulness of animal studies.
An article in the Summer 2007 issue of HealthBeat discusses the study and one that came to similar conclusions.
By Michael Bracken
Yale University
There is an urgent need for health and science journalists to have a better understanding of the scientific method. The impression is that many health and science journalists have never themselves had training in science, they have a simplistic view of science, and they do not understand the inherent limitations of the scientific method.
Scientists view their knowledge as building by accretion and replication. They know that most initial reports will not be replicated and that even replicated findings may still reflect ultimately spurious associations.
Ninety-nine percent of animal studies are of no relevance or interest to the public because they represent a single observation, they are likely wrong (in no small part because of poor methodology, documented in the British Medical Journal piece), and even if correct are likely to have little human application without a lot more work. The same applies to many in-vitro studies. Scientists know this, journalists appear not to.
Even 95 percent, say, of human research is likely to be eventually shown to be incorrect. Over-eager journalistic reporting of human studies has an even more negative effect. Journalists and their readers are likely to think that because the factors studied in human research (coffee drinking or eating carrots and cancer) are more readily understood, they have understood the reported study. The fact that these are highly likely to be spurious observations is not appreciated and journalists are all too quick to draw inferences about what the findings mean to human behavior (don't drink coffee, do eat carrots). This leads to the public being irritated (or bored) with what to them seem to be continually contradictory results being reported in the media, and to their not being able to sort out the truly scientifically based evidence from the non-reproducible results.
None of this is helped by having a science-illiterate population, but this simply makes the work of accurately reporting science more necessary.
Journalists need to present new findings in the context of existing knowledge, which requires a lot of research by them into the topic, not just calling a resident "expert" for perspective. Almost every journalist that calls me has an immediate deadline. It is very obvious they do not have time to research "context" and so we get the quick, and woefully simplistic, story.
