Junk Science or Sound? Five Questions to Ask

Junk Science or Sound? Five Questions to Ask
April 1, 1999



In her February 17 address to the Independent Women’s Forum conference, “Scared Sick? A Conference to Examine Unfounded Fear and its Effect on Health and Science Policies,” author Marcia Angell urged the audience to greet newly released studies with a healthy skepticism. To begin the process of sorting through the science, she suggested, ask the following five questions:

  • What’s the source of the report? A published report in a peer-reviewed journal is more reliable than a report at a scientific conference, and that’s far more reliable than a press release.
  • What data are reported? Anecdotes and investigations that focused on a single factor and ignored confounding variables are of no or little value.
  • Does the report fit with general knowledge? Implausibility suggests error--not a breakthrough.
  • How big is the effect? Reports frequently say something like changing diet or other behavior will reduce risk of a serious disease from 2 percent to 1 percent. Before deciding that making the change is worthwhile, it’s useful to consider that the chance of avoiding the disease is 98 percent if no change is made and 99 percent if it is. Is that increased likelihood of escaping the disease worth the inconvenience, discomfort, or cost of the change?
  • Is the report important to the individual or to public health? Taking a tiny risk to one person and multiplying it by the nation’s population produces a large number, and such large numbers increase the political pressure to make changes.