A Feast of Feedback

by Brian Dunning
August 19, 2025

Sponsor an episode!

If you're going to do a show like this, you're going to get feedback — everything ranging from praise to death threats, and from "constructive criticism" to rambling novel-length screeds. You're also going to get proper feedback: the pointing out of actual mistakes; thoughts from people who were actually central characters in the stories discussed, or descendants of theirs; and people knowledgeable enough to offer relevant feedback or updates on the subject. Those are the best. I've heard from many people over the years who had some connection to the actual urban legend discussed. I've also heard from fellow academics who have researched the same story and had learned either more or less than I had. And frequently, this feedback includes something relevant enough for me to convey here to you, my listener, and to update my original transcript accordingly. So today we have an episode consisting of feedback and followups. No error corrections, as those will go into a different show; but today's enhancements are equally cool and amazing.

So without further ado, we'll get started with two interesting bits of feedback on the recent episode about an infamous story from the Australian outback:

Did the Dingo Take the Baby?

Episode #977 was about the origin of the phrase "a dingo ate my baby" or some variation on it, and it turns out that its origin is exactly what it sounds like: a horrifying case of an Australian wild dog killing a baby, way back in 1980. Why then, we asked, is the quote normally used comedically, when it's from such an unspeakable tragedy? Listener Stephen thought it had to do with the movie version of the story, which starred Meryl Streep as the mother, Lindy Chamberlain:

For Australians a particular issue with "A dingo took my baby" was that the pronunciation of this by Meryl Streep grated with Australian ears. So there was an element of "taking the mickey" by Australians initially in building the popularity of the phrase.

Now the funny thing is that I heard this from several Australian listeners, that Meryl Streep's accent was so bad it was hilarious. However, I also heard the opposite from at least two Australians who noted that the Chamberlains were originally from New Zealand, and given that, they thought Meryl Streep's accent was spot on. So I have no idea what to think there; but it is one good theory for why it's considered funny.

Another email came from listener Chris, about one piece of evidence in the lengthy murder trial Lindy Chamberlain had to endure:

If the Chamberlains are truly innocent (which it sounds like was proved beyond a reasonable doubt), why was the blood found in the car? Of course this only matters if it was the baby's blood. If not, it's just noise.

In the episode, I said forensics experts had found "traces of blood on the seats of the car," and this was a key piece of evidence for the prosecution. In fact, it wasn't blood at all; it was Dufix, a sound-deadening compound sprayed on various parts of that model of car during manufacture. The defense showed that other examples of this same car also had overspray of Dufix in the same places.

The forensic examiner who misidentified it as blood had made all kinds of blunders, several of which she later admitted; and the case has become known as one of the most significant forensic failures in Australian legal history. Flawed scientific evidence and bad testing protocols can and do lead to wrongful convictions — as we discussed in Skeptoid #821 on forensic pseudoscience.

Shaking Up the Earthquake Predictors

Episode #873 was about the prediction of earthquakes, particularly about crackpots on the Internet who claim to have developed some method for accurately predicting specific earthquakes — something which even today remains impossible. However, about a year after my episode came out, an article appeared in Nature Communications by Girona and Drymoni who claimed to have developed an artificial intelligence model that was able to accurately predict major earthquakes months in advance, all by looking at patterns in low-magnitude seismicity preceding them.

The world headlines ate it up with a spoon, happily declaring that earthquake prediction had finally been cracked. So with such headlines out there, it's not surprising that quite a few Skeptoid listeners emailed to tell me I should either correct or retract my episode. Well, as I've mentioned before, my son happens to be a seismologist, and so I went straight to him for the real scoop. He hooked me right up.

Girona and Drymoni's article had become something of a laughing stock in the world of seismology, because it was so horrible. It did not remotely support their claims. They made their software and data freely available, so geologists Kyle Bradley and Judith Hubbard re-ran their programs. They had made 20 different machine learning models, all based on actual historic seismic data preceding two major quakes: a 2018 one in Alaska, and a 2019 one in California. Nineteen of the models failed to predict either quake. Model number 20 worked on both — but with some important problems. The Alaska prediction was based on partial data only. Bradley and Hubbard filled in the missing data, and the prediction vanished. The California prediction was based on only a small amount of local data; when Bradley and Hubbard provided the complete regional data, that prediction vanished as well. In other words, everything Girona and Drymoni tried failed 100% of the time. There was not even a false positive as we would expect from random chance.

So, no: artificial intelligence has not cracked earthquake prediction.

Stop Giving Oxygen to Hyperbaric Oxygen Therapy

Episode #702 was about hyperbaric oxygen therapy (HBOT), where you put a person into a pressurized chamber of pure oxygen. It is approved as treatment for thirteen different conditions, but — unfortunately — it's also one of those therapies that the quacks have gotten hold of, and sell it for just about anything and everything you can imagine, most of which are backed by zero science. It's just one more way that charlatans take advantage of people who are suffering.

Listener Olivier wrote in about a tragic news item:

I was looking into this because of an interesting case happening in my city, a young boy was found in the bottom of the pool at the house rented by his parents. The hospital considered the boy brain dead and suggested the parents consider organ donation. The parents ended up moving their boy down to Louisiana for HBOT treatment and they are adamant the boy is improving.

Evidence that HBOT can help restore dead brain tissue is sorely lacking, and it's a treatment that is only sold by clinics who also claim it can cure autism, ADHD, cancer, and all kinds of things it most certainly doesn't. For each of these, there are a few small published studies supporting it, usually of poor quality and always in opposition to larger, high quality studies. This is the case with brain injuries as well.

The brain has a certain amount of neuroplasticity and can, in some cases, regain lost function by retasking undamaged parts of the brain. The claim that HBOT can facilitate this is not entirely implausible, but it remains unsupported by any compelling evidence. Whatever Louisiana clinic these parents took their child to is almost certainly just ripping them off — taking advantage of the grieving.

A Big Update on the Loveland Frog

Episode #473, guest hosted by Ryan Haupt, was about the Loveland Frog, sometimes called the Loveland Frogman, said to have been a group of froglike humanoid creatures just over a meter tall that were seen in Loveland, OH in 1955. Then in 1972, a police officer, Mark Matthews, claimed to have shot such a creature. Soon thereafter, however, Matthews admitted he'd made that up, and that what he actually shot was simply a large (but otherwise ordinary) iguana that was missing its tail.

This story came up in the news again in 2016, just about a year after Ryan's episode. Some people playing Pokemon Go claimed to see it, and even produced a video of something standing in a lake exhibiting bright eyeshine. Ryan wrote me that this refresh of the story prompted the police officer, Mark Matthews, to come forward with more information:

In my episode I referenced Matthews recanted his story back in the day, but didn't have a definitive source. He says he had already told the full true story to the author of a book on urban legends, but the part where he clarifies that it was an iguana was left out.

I had discounted the idea that an iguana could have survived the Ohio winters, and concluded that a mangy coyote was more likely, but with the officer's more clear admission that what he shot was a tailless iguana and the idea that a factory may have been discharging warm water into the river I would say that the possibility that a sickly escaped or released pet that was barely hanging on by the river increases.

So there we have it. Once again, we keep having to repeat the true explanation behind an urban legend; nobody seems to want to hear the true version.

Latest News on Fracking Waste Water

The process of fracking has come up in a number of Skeptoid episodes. This is pumping pressurized water into natural gas bearing rock in order to carry sand down there to prop open fissures and release more gas faster. Oil companies love it; environmentalists hate it.

One of its aspects is the final disposition of this water. Today, it's generally too expensive to treat it so that it can be used by farmers, etc., and so it is typically disposed of by pumping it into superdeep injection wells — a process which has now definitively been linked with clusters of small earthquakes.

In 2020, the EPA published a report following years of feedback from all the various stakeholders about what we really should be doing with these waste waters. I wish I could report that there is a clearer direction, or that any one particular discharge or reuse option is preferred, but the best way to summarize this report is that very little has changed:

• Some states, native tribes, and industry stakeholders favor expanded options, mainly because water is a crucial resource and water shortages continue to be a bigger and bigger problem;
  
  
• Many NGOs and some native tribes favor the continued superdeep injection of the contaminated wastewater due to concerns about environmental and human health impacts.

Often people ask if fracking should be allowed or not, with the anti-fracking crowd often exaggerating or even making up terrifying environmental impacts like earthquakes or poisonous groundwater. It absolutely does need to be stopped because it's the best way to extract methane, and methane is currently the greenhouse gas of greatest concern. But more than fracking, the entire fossil fuel industry needs to be given the heave-ho as climate change is already giving us the worst environmental disaster in history. That's the bigger issue than fracking; but if banning fracking can at least slow things down, then we need to do that.

By Brian Dunning
Follow @BrianDunning