The conclusions of an individual scientific study are of little value when taken independently. We are all familiar with the myriad biases that crop up in both the process and presentation of science that undermine or invalidate such conclusions, but our understanding of the extent to which this happens is still recent.
In one sense, the entire point of and justification for the scientific method is that scientific findings are reproducible. If something has not been reproduced or, worse, cannot be reproduced, then it can hardly be considered settled science. For this reason, I am unmoved by arguments that rest on individual, non-reproduced studies. I am happy to talk about the implications raised by new studies, or what it might mean if a recent scientific finding is reproduced, but this is the stuff of conjecture, not fact.
If you view science as our best method of uncovering the truth about physical reality, and if you view intellectual integrity as a laudable goal when engaging in discussions, then I suggest the following two steps when citing studies to support arguments:
- Check to see if a study has been reproduced. If not, don’t cite it.
- Check to see if a study or its authors are listed at Retraction Watch. If so, find out why.
Any study worth citing should be able to pass this bare minimum of scrutiny, while every study definitely not worth citing will not. Pseudoscience, as its name implies, tries to pass for science. It dresses itself in the trappings of science, but it cannot withstand the rigorous scrutiny of the scientific method.