We are in the midst of a revolution. With so much "bad science" being put out there, I think it's important for our community to educate and reinforce what statistics can and cannot prove.. .what "statistically significant" and "insignificant" truly mean.
All too often, I see article after article making wild claims and then you go into the methodology and see glaring errors. What this article talks about (the infamous "p" value) is a more insidious error in that the math is correct but the interpretation isn't, an interpretation often born more closely out of dogma and begging the question than any valid experimentation.
I think this is a very important article for anyone interested in science and knowledge in our society...
Being published is currency in the scientific world and like money, it has been used for the greater good and to enrich the bad. As a published author (just 2 small studies), I realized, that like in the real world of the internet (irony), click bait titles attract attention. That said, I have come to heed the constant line added to many scientific articles: ...more studies are required to corroborate these finding.
My advice to reading scientific (based) articles is:
Do not give credence to a press release or main stream media article of an actual scientific article. Go to the source.
The reputation of the journal is important (Nature, Cell, see: [scimagojr.com]. These articles have a strong/rigid peer review process.
Repeatability i.e. demonstrating a growing body of evidence from many articles
Study design: ex. double Blind placebo controlled clinical trials, sample size: bigger is better
Actual study: cell culture and animal studies does not equate to applicable to humans but it is a first step in a body of evidence
Lastly: controversial articles, meaning going against a large body of evidence as for anti-vaccination or climate change for example, often cherry pick the data (big red flag!)
Very, very, very often, almost always, serious authors admit to their biases by declaring any conflict of interest (who paid for the study) and how their study can be discounted as having weak points in their conclusions. Be skeptical if none are discussed. Great discoveries and eureka moments are very rare so the more an article elludes to their findings as being unique is also a red flag.
Science, like life, is always in motion; trends are more important than outliers.
@CrazyQuilter interesting? yes, but insignificant without corroboration or repeatability.
Most recent example I've seen: frequent use and/or strong strains of marijuana linked to psychosis. The statistics were insignificant, the control group was small, the study didn't span a very long time, but the story made NPR and the local news. Point being, if you're looking to prove something, you can find or create statistics to back yourself up.
I'm beginning to think there's nothing left worth believing in. Clearly not religion, not politicians, and now, sadly, "science."