Scientists seem to have finally had enough of activists who promote every study in mice as if it were relevant to humans. A Twitter account, @justsaysinmice, calls out such bad behavior and quickly grew to 57,000 followers by linking to studies produced each day and noting they are just in mice. Clearly there was a lot of frustration pent up about hyperbolic claims in media using animals.
There is nothing wrong with animal models, of course, we would be doomed without them. But they are exploratory. A mouse study can disqualify a benefit or harm for humans, but it can never show it. That is why it is necessary to go to human trials for new drugs. Despite that limitation animal models get an irrational amount of media attention. Cancer has been cured 5,000 times and we can live to age 200 on a starvation diet if we believe mice are little people.
They aren't. One thing that can help add strength to findings using animal models (in cases where human experiments are not ethical, like smoking) is epidemiology. Smoking was a big win for statistical correlation, and epidemiology also showed that smog (PM10) and alcohol were carcinogens. But those victories caused a rush into the field, with everyone wanting to get into The Atlantic for finding the next literal smoking gun, and now the public are abused by statistical correlation claims just like they are animal testing.
In ethical hands, epidemiology is powerful, but it has an exploit. If you put a bunch of diseases in columns and a bunch of foods or chemicals in rows and use a large enough sample, you are sure to get a p .05 result linking some product and some disease. We can even use this common epidemiological technique (and it is common, Chensheng Lu of Harvard TS Chan School of Public Health and others there use it so much they could carry its name) to show coin flips are not random with p .05 authority. It's not valid for real science but it can considered valid when it's time to get a paper published in a journal.(1)
A recent paper is a mashup of both, with a smattering of test results thrown in. Paul Mills, Ph.D, is a contractor (basically adjunct) at U.C. San Diego School of Medicine, which allows him to use a professor title to gain authority his work does not have. He doesn't have a medical degree and even his Ph.D. is from Maharishi Institute, better known for claims they can help you levitate than for any science.(2) Though his work is often bizarre the UC San Diego patina of authority is still enough to get him published in a journal when the compound he is attacking is glyphosate. And he attacks it often.
Glyphosate is an herbicide, a weedkiller that is very popular among farmers and gardeners even after 40 years in use. In California, a state friendly to anti-corporate and anti-science beliefs, a San Francisco jury decided last year that it might have caused a lawnkeeper's cancer and so another another lawsuit followed, and then 13,000 more, because that is what trial lawyers do.
I don't need to go into how it's impossible for the product to cause human cancer, other people have done that better. What's been interesting is how the organic food industry and lawyers suing over the competitor to organic food have been able to get Mills, his collaborator John Fagan, Fred vom Saal, and a small group of others to write papers just at the time their trial lawyer friends need help.
Mills is always at the ready to pitch in and a new article has gone a step farther than his usual claims, which involve suggesting that the presence of a compound at any level has an affect (basically, homeopathy, except causing cancer): In his latest press release on glyphosate, he claims his next paper will feature those who only ate organic food and will show it can prevent liver disease.
Wait, what? Kroger is causing liver disease by selling food without copper sulfate?
Conventional food causes liver disease but organic food does not? If you are reading here, you will already have tripped him up with a simple four word skeptical question; where are the bodies? There are none. Now, I've never gotten a penny from Monsato or Bayer, nor has Science 2.0, so let's dispatch the "Science Is A Corporate Conspiracy" insinuation that organic industry trade groups engage in whenever critical thinkers ask awkward legitimate questions about their beliefs. Without any conflict (unlike Mills, who engages in fraud by refusing to note his conflicts of interest) I can tell you with 100% confidence +/-0% margin of error that if America switched from the 4% organic that rich people buy now to 100%, 96% of liver cancers would not disappear due to a diet he and his friends prefer.(3)
The method: A few hand-picked test cases here, an animal model there, and epidemiology over there
(1) An actual statistics journal, however, will not. They know better. Surveys, which were somehow claimed to be science after getting the Presidential election results right in 2008 and 2012 (though in the latter no more accurately than bookies in Europe did) collapsed in credibility in 2016.
This can show you why if a statistical result claims 95% confidence interval it is meaningless. Looking at the chart below for 95% confidence with +/-5% accuracy you can see that can be achieved with 250 people. It sounds authoritative but it isn't, which is why poll averaging with secret sauce for weighting (Nate Silver and others) isn't very accurate, unless the result is easy to know, like in the 2008 and 2012 elections.
(2) The man thinks he can talk to ghosts, and that something called a singing bowl has magical recuperative powers. His claims about science are just as ridiculous. but he is a favorite of journalists who want to twerk the mainstream community using glyphosate, because he will create any link you can imagine in his war on science. Maybe this paper he wrote with Deepak Chopra on herbal therapy can show how to cure the liver disease a weedkiller somehow caused.
His alma mater is now called Maharishi University of Management in Iowa and if that still sounds like an odd place to get a Ph.D. in neuroscience, your instincts are correct. It was even less credible in the 1980s but even today is not accredited by the American Psychological Association or the Accreditation Council for Graduate Medical Education. It hasn't stopped the group founded by the former Beatles guru from creating a $5 billion alternative to science empire.
(3) In a JAMA paper with his ally John Fagan, Mills chose not to disclose that he created a crowdfunding site to raise money to buy glyphosate test kits from HRI, a boutique testing service Fagan created to attack glyphosate. Rather than retract it, the magazine simply added an edit after we exposed it.
