Debunking Junk Science

In her article, Junk Science in the April 2014 issue of Popular Mechanics, Sarah Fecht offers six warning signs for junk science.  Although one might find Popular Mechanics to be an unusual publication for such work, science editor Ms Fecht has published widely in The New York Times, Scientific American, Forbes, Popular Science, National Geographic, and other places.She has a good grasp of what qualifies as junk science.  She lists:

• Was the paper published in a [credible] peer-reviewed journal?  She admits peer-review may be flawed, but says "It stands between us and really poor science."
• What is the Journal's Impact Factor? In other words, how often are its articles referenced in other works?
• Do the researchers mostly cite their own work?  This is a red flag that this work is outside scientific consensus.  That does not disqualify it, but suggests one tread cautiously in citing this author.
• How Many Tests Where Used? Most tests improve reliability.  We add "Has the author made the raw data public?" Transparency improves credibility.
• Does It Rely on Correlation? Just because two events have a similar history does not imply one is caused by another or that they are even related.  The infamous attempt by Al Gore to correlate proxy global temperature anomalies with carbon dioxide concentration in the atmosphere is a remarkably disingenuous demonstration.
• Have the Results Been Reproduced? No science result is credible until its has been reproduced by independent researchers.  In 1989 Stanley Pons and Martin Fleischmann  claimed to have produced cold fusion. Their work was never reproduced and by the end of 1989 was totally discredited.