Bill Jefferys points to this news article by Denise Grady. Bill noticed the following bit, “In male rats, the studies linked tumors in the heart to high exposure to radiation from the phones. But that problem did not occur in female rats, or any mice,” and asked:
Forking paths, much?
My reply: The summary of the news article seems reasonable: “But two government studies released on Friday, one in rats and one in mice, suggest that if there is any risk, it is small, health officials said.”
But, yes, later on they get into the weeds: “Scientists do not know why only male rats develop the heart tumors, but Dr. Bucher said one possibility is simply that the males are bigger and absorb more of the radiation.” They didn’t mention the possibility that variations just happen at random, and the fact that a comparison happened to be statistically significant in the data is not necessarily strong evidence that it represents a corresponding pattern in the population.
Bill responded:
Yes, it was the weeds that bothered me.
Overall, then, yes, a good news article.
The size argument can be tested. Good science leads to testable predictions. The size of males and females can be measured and the increased risk should be proportional in some sensible way. So if the authors propose size, they should be obligated to test their hypothesis. That said, I suspect other explanations, including the stats issues, are in play, including sex-specific differences in DNA and cellular repair as well as other physiological mechanisms. Bottom line: make and test prediction.
Joe:
+1 on testable predictions. These are lab studies; it shouldn’t be so hard to get some more rats and do some experiments, right? Although it says in the news article that these studies cost $25 million, so I guess it’s not as simple as it sounds!
They could always use the data they have already collected to see if there is any correlation between size and radiation levels. That would at least tell you whether there is at least some support (however weak) for the hypothesis. Why shoot from the hip when you have data available?
I don’t think there is any data available on “radiation levels”. I mean, the exposure rate (in W/kg) is the thing the researches were controlling and not an outcome of the experiment.
They probably already trawled a lot of variables before they came up with the size explanation?
+1
Likely will have to wait for a jury trial to sort this out :-(
E.g. https://www.cnn.com/2018/08/10/health/monsanto-johnson-trial-verdict/index.html
Because jury trials are known for their accurate and subtle teasing out of scientific truth…. (sarcasm)
Roundup, eh? Well, we can look forward for lots more weeds to be getting into.
Of Scythes and Weeds: I suspect Andrew has some insight re: this: https://www.quantamagazine.org/universal-method-to-sort-complex-information-found-20180813/ … or so I hope. Or more likely, I’m just in the midst of the KNN chapter and over-thinking things.
Perhaps think this through first https://today.duke.edu/2017/07/opening-lid-criminal-sentencing-software via https://ww2.amstat.org/meetings/jsm/2018/onlineprogram/AbstractDetails.cfm?abstractid=333068
Having found the neighbors the prediction rules seem intuitive but my superstition is that with a good Bayesian model you will do better and understand more.
However, the authors of the above actually tried Bayes first – Scalable Bayesian Rule Lists https://arxiv.org/abs/1602.08610 – and then moved on to machine learning…
As a medical physicist, let me assure everyone that the difference in heart sizes leading to more of the radiation being absorbed is not a very plausible explanation. The wavelengths are so large compared to a rat heart, that the absorptivity won’t be affected. If organism size is a big factor, then maybe we should only use small people to clean up after radiation accidents–after all, they will absorb so much less radiation.