Scientists Call For Abandoning The Concept Of "statistical Significance"

haidut

Member
Joined
Mar 18, 2013
Messages
17,440
Location
USA / Europe
When somebody emailed Peat with the latest study showing a highly statistically significant outcome that immunosuppression was a viable treatment of autoimmune conditions, Peat's response was that most studies with "statistically significant" results are actually quite insignificant and often based on heavily manipulated data. He also said that often the studies that do not show statistical significance are the ones to pay attention to. It looks like quite a few scientists in the medical field agree and, according to the article below, there is now a petition to NSF to abandon the term statistical significance the other binary categorizations schemes. And as the article aptly notes, often it is people without any statistical background that are capable of drawing viable conclusions based on the data while statisticians are paralyzed by their own biased analysis.

Scientists rise up against statistical significance
"...When was the last time you heard a seminar speaker claim there was ‘no difference’ between two groups because the difference was ‘statistically non-significant’? If your experience matches ours, there’s a good chance that this happened at the last talk you attended. We hope that at least someone in the audience was perplexed if, as frequently happens, a plot or table showed that there actually was a difference. How do statistics so often lead scientists to deny differences that those not educated in statistics can plainly see? For several generations, researchers have been warned that a statistically non-significant result does not ‘prove’ the null hypothesis (the hypothesis that there is no difference between groups or no effect of a treatment on some measured outcome)1. Nor do statistically significant results ‘prove’ some other hypothesis. Such misconceptions have famously warped the literature with overstated claims and, less famously, led to claims of conflicts between studies where none exists."

"...Let’s be clear about what must stop: we should never conclude there is ‘no difference’ or ‘no association’ just because a P value is larger than a threshold such as 0.05 or, equivalently, because a confidence interval includes zero. Neither should we conclude that two studies conflict because one had a statistically significant result and the other did not. These errors waste research efforts and misinform policy decisions."

"...These and similar errors are widespread. Surveys of hundreds of articles have found that statistically non-significant results are interpreted as indicating ‘no difference’ or ‘no effect’ in around half (see ‘Wrong interpretations’ and Supplementary Information). In 2016, the American Statistical Association released a statement in The American Statistician warning against the misuse of statistical significance and P values. The issue also included many commentaries on the subject. This month, a special issue in the same journal attempts to push these reforms further. It presents more than 40 papers on ‘Statistical inference in the 21st century: a world beyond P < 0.05’. The editors introduce the collection with the caution “don’t say ‘statistically significant’3. Another article4 with dozens of signatories also calls on authors and journal editors to disavow those terms. We agree, and call for the entire concept of statistical significance to be abandoned."

"...We must learn to embrace uncertainty. One practical way to do so is to rename confidence intervals as ‘compatibility intervals’ and interpret them in a way that avoids overconfidence. Specifically, we recommend that authors describe the practical implications of all values inside the interval, especially the observed effect (or point estimate) and the limits. In doing so, they should remember that all the values between the interval’s limits are reasonably compatible with the data, given the statistical assumptions used to compute the interval7,10. Therefore, singling out one particular value (such as the null value) in the interval as ‘shown’ makes no sense. We’re frankly sick of seeing such nonsensical ‘proofs of the null’ and claims of non-association in presentations, research articles, reviews and instructional materials. An interval that contains the null value will often also contain non-null values of high practical importance. That said, if you deem all of the values inside the interval to be practically unimportant, you might then be able to say something like ‘our results are most compatible with no important effect’."

"...Last, and most important of all, be humble: compatibility assessments hinge on the correctness of the statistical assumptions used to compute the interval. In practice, these assumptions are at best subject to considerable uncertainty7,8,10. Make these assumptions as clear as possible and test the ones you can, for example by plotting your data and by fitting alternative models, and then reporting all results."

"...What will retiring statistical significance look like? We hope that methods sections and data tabulation will be more detailed and nuanced. Authors will emphasize their estimates and the uncertainty in them — for example, by explicitly discussing the lower and upper limits of their intervals. They will not rely on significance tests. When P values are reported, they will be given with sensible precision (for example, P = 0.021 or P = 0.13) — without adornments such as stars or letters to denote statistical significance and not as binary inequalities (P  < 0.05 or P > 0.05). Decisions to interpret or to publish results will not be based on statistical thresholds. People will spend less time with statistical software, and more time thinking."
 

Malris

Member
Joined
Mar 12, 2017
Messages
30
Interesting. One of the things that really bothers me about science is that they make little effort to acknowledge what state the person is in at the beginning. I mean, it's one thing to say "diabetics who eat XXX have YYY result." It's an entirely different thing to say "and because of that result everyone should eat/not eat XXX."

It's partly the media's fault (see How the "chocolate diet" hoax fooled millions - CBS News, where a scientist made a study hoping for a specific result, and then manipulated the media to get a specific interpretation of that result promulgated) but it's also the scientists' fault for jumping to conclusions, just like in the article that you listed, Haidut.

I think, rather than getting rid of confidence intervals, it would be better to force everyone to state how specific study results relate to a mental model of the body, similar to what Peat does. Tell us what your supposed mechanism of action is, and why that would work, and include the body's first level response to that. Kind of like the (likely incorrect) theory "sugar causes diabetes because of fatigued beta cells in the pancreas" type thing. That would then allow others to correct broken mental models, and propose ones that they think are more accurate, like "low blood sugar, combined with a damaged small intestine, cause diabetes. This happens through release of stress hormones cortisol and adrenaline, which cause the liver to perform gluconeogenesis. The hormones are like a blunt hammer, and cause too much blood sugar to be released. Therefore, ensuring that the blood sugar does not fall is the best way to fight diabetes."

With the right tools, a more accurate mental model of how the body works would be built up and corrected over time.
 

managing

Member
Joined
Jun 19, 2014
Messages
2,262
As somebody with an extensive science background and graduate statistics courses, this is a bit mind blowing.

I've always treated "intelligent interpretation" of science as a focus on methods, assumptions, and the critical thinking involved (or not) in the conclusions drawn from the data. But I've never questioned statistical significance. P is king, as one of my statistics professors used to say.

The malcontent contrarian in me wants to embrace this. The well-trained statistician is frankly disgusted by it!

If anybody has a good example of how confidence intervals (compatibility intervals) are used in discussing/assessing study data, I'd be intrigued to see it. I think at the end of the day, statistical significance is disposable if the replacement is superior. I just can't quite wrap my head around how that looks yet.

Thanks @haidut , this may just be the most disruptive thing you've ever posted.
 
Similar threads
Thread starter Title Forum Replies Date
A Article - Did Chinese scientists find autism’s missing puzzle piece? Articles & Scientific Studies 2
E UK Column... Political catcher of scientists in secret document Breaking News 3
Pointless Scientists Find 'secret Molecule' That Allows Bacteria To Exhale Electricity Scientific Studies 2
Mito Secret Group Of Scientists Backed By Billionaires Are Working On A 'Manhattan Project' To Stop COVID Miscellaneous Health Discussions 33
David PS Scientists Try To Figure Out If Summer Will Slow The Spread Of COVID-19 Viruses, Parasites, Fungus, Vaccines 7
HealthisWealth Coronavirus Breakthrough As Scientists Discover A Drug Used To Treat HEAD LICE Can Kill COVID-19 Cel Viruses, Parasites, Fungus, Vaccines 48
Mito Scientists Map How Immune System Fights COVID-19 Articles & Scientific Studies 0
JanP Bulletin Of Atomic Scientists: Nuclear War Is More Likely Than It Ever Was In The History Of Mankind Health 7
H Scientists Find First Evidence Of Humans Cooking Starches Articles & Scientific Studies 8
W Corvelva Italian Scientists Blow The Whistle On Vaccines Composition Scientific Studies 2
sweetpeat Scientists Are Going To Starve Teenagers Rant or Rave 3
S 50 Years Ago, Sugar Industry Quietly Paid Scientists To Point Blame At Fat Diet, Recipes 2
E Scientists Believe They’re Getting Closer To An Anti-aging Pill Scientific Studies 36
L Scientists Develop Lab-Made Mineral That Will Suck CO2 From The Atmosphere Articles & Scientific Studies 47
T Scientists Unravel The Structure Of Common Bacterial Wall-building Protein Scientific Studies 0
T Scientists Unravel DNA Code Behind Rare Neurologic Disease Scientific Studies 0
T GAPING Holes In Evolutionary Theory Leave Scientists Baffled (and Darwin Rolling In His Grave) 2018 Articles & Scientific Studies 3
Mito Scientists Have Revealed Which Countries Have The Highest And Lowest Levels Of Cancer Cancer 22
Mito Immunotherapy-first Cancer Treatment A 'game-changer', Say Scientists Cancer 1
Mito Scientists Think They Found A New Human Organ—a Fluid-filled “shock Absorber” Miscellaneous Health Discussions 0
P Scientists Discover An Entirely New Biological System That Regulates Body Fat Articles & Scientific Studies 2
Soren Scientists REVERSE Brain Damage In Girl, 2, Who Almost Drowned In A Pool Using Oxygen Therapy Articles & Scientific Studies 4
bodacious Russian Scientists Reveal Secret Of Eternal Life Miscellaneous Health Discussions 0
B Scientists Find Skin Cells At The Root Of Balding, Gray Hair Hair & Nails 5
Mito Scientists Find Skin Cells At Root Of Balding And Gray Hair Articles & Scientific Studies 4
Let Go Eating Bogies Is Good For Teeth And Overall Health, Scientists Conclude Articles & Scientific Studies 11
M Scientists Are Getting Closer To The "obvious" Truth On Naked Mole Rats Supplements 3
R Most Scientists 'can't Replicate Studies By Their Peers' Science 2
N Scientists "find Out" Glioma Fueled By Fat. Contradicts Previous "findings" Cancer, Degenerative Diseases 1
S Looks Like These Scientists Are Looking To Jump Into The Anti Aging Realm Supplements 21
E Scientists Still Have It Wrong On Obesity? Miscellaneous Health Discussions 45
EndAllDisease Study: Scientists Suggests Boosting Metabolism For Weight Loss Scientific Studies 0
charlie Scientists Finally Admit Climate Models Are Failing To Predict Global Warming Political Talk, Alternative World Theories 52
L The Geometry Of Histamine Understood By Russican Scientists Scientific Studies 0
Philomath Scientists Have Discovered Orgasmic Mushroom That Gives Women An Earth-Shattering Experience Articles & Scientific Studies 3
haidut Scientists Now Think That Being Fat Can Protect Your Health Scientific Studies 51
burtlancast Scientists Reverse Cell Aging By Genetically Enhancing Glycine Production Articles & Scientific Studies 4
Philomath Scientists unboil an egg - Peat relevance Ray Peat Topics 1
charlie LSD cures depression? Scientists plead for cash to fund ‘exciting’ drug study Alcohol, Tobacco, Drugs 2
haidut Why Can't The World's Greatest Scientists Solve The Consciousness Riddle Mind, Sleep, Stress 15
C Call For Help: Bone Health And Strength Diet 2
ecstatichamster A Call Out For Bamboo Shoots Digestion, Gut Flora 23
haidut A Call For Medical Nihilism - To Do Nothing Is Also Good Medicine Scientific Studies 53
E I Dont Know What To Call This Thread But Its A MUST SEE Cholesterol 38
charlie Wireless Wake-up Call Radiation 0
S It Takes a Lobby to Call a Lobby Out Political Talk, Alternative World Theories 1
M Call me Mamaherrera Meet & Greet 17
charlie Better Ideas On What To Call This Area? Ask For Help or Advice 10
charlie What to call Peat eating... Polls 31
T Abandoning the Ray Peat Diet Polls 84

Similar threads

Top