Ben Goldacre helps us to evaluate science and shows publication bias

Often on this blog we have exposed that which has been deemed “scientific” and revealed it to better categorized as science fiction. So, the question becomes how do we distinguish between true science based  upon sound research and transparent from pseudo-science or science fiction that has somehow entered into the courtroom as both are termed “scientific”?

It has been long recognized that there is a publication bias in science in general. I submit that no where is this more so featured than in the endeavor that is called “forensic science.” Rarely, if ever, do we find in the Journal of Forensic Science an article that is not enormously pro-prosecution. As a generality with notable exceptions, the publications are over-expansive as to its conclusions based upon a limited data set. Again, with notable exceptions, to a degree the ones that are published are policy papers that are wrapped around the idea of science. It is not as if articles from the other side of the equation do not exist. They do. It is just they are not published. This is a problem. Good and solid science should always be published in the true scientific community regardless of its political point-of-view of the editors or the organization. Data rules.

The above clip helps us to create a hierarchy of scientific studies that I suggest makes sense.

Ranked order (Least scientific to most scientific/empirical):

  • Authority (Just because I have letters after my name that you do not, I am right and you are wrong)
  • Observational Study (based upon testing in a limited environment and limited population)
  • Trial (split population and then compare data-open, blind, and double blinded)
  • Trials against placebo
  • Trials against the current best available method/technique

This whole distinction is more about the abuse of science by salesmen than bad science per se. Scientists are human and are not immune from all forms of cognitive bias.

Letters after a name do count in so far as they establish a degree of reliability and trust, but they just do not excuse the person’s work from proper scrutiny.

Leave a Reply

Your email address will not be published. Required fields are marked *