Unsound autism research claims

Post Reply New Topic

Over the years I have been on WP, there have been a lot of "research shows" posts, some valid, many not nearly as robust as the press releases claim, and sensationalization is not only done by media - some researchers overclaim, and some do it knowing that the lay community will never scrutinise the actual research (if they can even access it) and even if they can, most will be stumped by the statistical stuff. Over-promoting results is done for many reasons, sometimes because of unconscious bias, or quite conscious reasons such as eg: the hope of improving the chances of getting future grants, to accelerate a career climb, and maybe in some cases, vanity. Also, corruption in science exists, and it is not beyond the bounds of credibility that large 'research fees' from wealthy organisations might buy them the result they want to quote as unqualified fact ("research shows.."). Researchers are human too..


Here's a recent example of faulty overclaiming: you may have read the headline "Premature Babies More Likely to be Autistic". Much was made of it.
Here is an informed appraisal of the actual study, which is markedly different from what the media reported so widely:
http://www.thecogitoblog.com/blog/no-yo ... /#more-143

Compare the above with how the research was reported in headlines (in bold)

Autism Risk Increases the Earlier a Preemie is Born | BabyMed.com
By Sandy Hemphill, Contributing Writer, BabyMed.com
View on www.babymed.com

Extremely Premature Babies at Greater Risk for Autism: MedlinePlus
View on www.nlm.nih.gov
Preview by Yahoo

Preterm babies more prone to autism
View on www.medicalnewstoda...
Preview by Yahoo

Even the websites you might expect to be more "science based", and to practise critical evaluation in their reporting failed to do so.

Feel welcome to post other examples of this phenomenon..

Remarkably (sarcasm there), there is a study proving that corn syrup causes autism, published in a journal and again, reported on as if the claims were factual; read a critique here:

http://doubtfulnews.com/2012/04/autism- ... tion-mill/

The kind of journal a study is published in isn't an automatic guarantee of robust scientific findings either (The Lancet published the Andrew Wakefield faked research as if it was factual and credible). There have been so many peer review scams in the last couple of years that even journals like Nature and other similarly high profile, evidence-based journals have been taken for a ride.

This is a very good point, and even the best-meaning, most ethical scientists can design a study with poor methodology, design flaws, biased sampling, study groups that are too small for statistically significant results, unintentionally biased results, and poorly applied statistical tests. Good peer review is essential, and I really wonder how the Wakefield study managed to get by their editors.

I always wince when the mainstream press starts making a big fuss over a scientific result before it's gone through peer review and follow-up studies. There have been a lot of disappointing follow-ups to grand or scary announcements. Anybody remember ALH84001, the Martian meteorite that supposedly had fossilized bacteria? The jury is still out on that one, but further studies cast doubts on early hypotheses, scientists were able to duplicate the shapes with non-biological means, and even initially scientists were trying to get everyone to not get their hopes up.

Remember cold fusion? In 1989, two scientists claimed to have been able to create a fusion reaction at room temperature, as opposed to "hot fusion", like what you have in a hydrogen bomb or the heart of a star. Everyone was pretty excited -- cold fusion could be controlled and harnessed for alternative energy, unlike the destructive power of a hydrogen bomb. Lots of other scientists were unable to replicate the experiment, and it was finally agreed that the original scientists made a mistake in data gathering.

I agree GodzillaWoman - given its glaring faults, how on Earth did Andrew Wakefield get his paper past the editorial board of the Lancet in the first place?

For those unfamiliar with this history, a brief overview:


by Richard Feldman: Vaccines and autism: Numerous studies indicate no connection

British researcher Dr. Andrew Wakefield authored completely bogus research in 1998 that linked the measles-mumps-rubella vaccine to autism. His fraudulent research was finally exposed; he was completely discredited and lost his British medical license. In 2011, the British Medical Journal published an investigative piece by Sunday Times reporter Brian Deer, debunking Dr. Andrew Wakefield’s vaccine/autism study as “an elaborate fraud.”

Only 12 children were studied. Doubts were raised about the manner in which they were recruited and the science with which the study was conducted.

As well, it was discovered Wakefield was on the payroll of a group that had launched a lawsuit against manufacturers of the MMR vaccine — and their claim would be based on his evidence.
...........................

The other really shocking aspect of this disastrous fraud is that reputable scientists had 13 years to expose the shoddy stuff that Wakefield wrote and The Lancet published. It was widely known in the serious research community that attempts by genuine scientists to replicate the fraudulent findings claimed by Wakefield were dismal failures throughout those years. It appears that no-one from that community was willing to give voice to what they knew was going on. Then after all those years - and how ever many deaths of children whose parents refused to vaccinate them - it took a journalist working for the Sunday Times to expose the scandal. The history of the silence is itself scandalous.