https://web.archive.org/web/20200221175201/http://kill-or-cu...

Too bad it's down, but here's an archive of it.

Author implies that Alwadi et al. was not forthcoming with sharing the data underlying their studies. Because otherwise they should be able to exactly replicate their results. "Reasonably close" does not cut it.

In reading the criticism, I noticed the authors keep using the term "prove" and they also keep trying to come up with mechanisms ("refueling of the plant"). Even the argument about plant worker exposure compared to people living far away doesn't completely work, because the plant workers are taking all sorts of precautions to minimize exposure to radiation, but there are still mechanisms where something could go out the cooling stacks and deliver something harmful downwind.

The biophysics of cancer causation is entirely nontrivial and looking for the actual sources of the cancer risk is challenging, and watching physics people argue with epidemiologists gets old quickly (my field is biophysics, and I've had a few physics people insist that non-ionizing radiation couldn't possibly cause cancer, "because it doesn't damage DNA". Unfortunately, that argument isn't good, because it presupposes a mechanism (DNA damage due to radiation); we know now that non-ionizing radiation causes cellular heating, stress response, and more, which are all associated (based mostly on in vitro studies) with increased rate of cancer.

Note the authors (and the institute they work for) have vested interest: Dr. Adam Stein is the Director of the Nuclear Energy Innovation program at the Breakthrough Institute, where his work centers on the technology, regulation, economics, and risk governance of advanced nuclear energy.

Deric Tilson is a Senior Nuclear Energy Innovation Analyst at The Breakthrough Institute, where he focuses on advancing nuclear energy as a critical pathway to a carbon-free and energy-abundant future.

https://pubpeer.com/publications/AEA87585FB0365719098E2C0D06...

Over the last several months, we have replicated the results of these papers. The authors supplied us with eight lines of code and answered a couple of questions about the covariates, which did not replicate the results. Most of our replication was done through first principles combined with trial and error. Once we were reasonably close to the results of the first national study on cancer mortality, we took the methodology and applied it to numerous other landmarks.

Any details of the methodology? Is it possible some mistakes were made during this "trial and error"?

edit: Some aggregated data is available from https://pmc.ncbi.nlm.nih.gov/articles/PMC12929679/ , but far from enough to reproduce: https://media.springernature.com/original/springer-static/es...

It is quite odd that the authors say that they don’t know if the original method can yield negative results. Some of the results are expressed as relative risk, so there’s a numerator and denominator. That should be enough to begin to work it out, since all possible exclusive numerators sum to the denominator. It feels like if you don’t understand whether some areas will be lower than average, then perhaps you don’t understand the method sufficiently.

It sounds like they reverse engineered the original method, but that means they could be applying an overfit model and getting spurious results that the original wouldn’t give.

That’s all speculation, hence we need the technical write up.

Alchemists conjured thousands of weird new substances, and we just shoved them everywhere they seemed fit. Asbestos everywhere, lead in gasoline, expanded polypropylene wrapping food and holding warm coffee, plasticisers practically everywhere..

And we only really learned about the ones that stood out as the worst. The effect of others, slower acting will keep being uncovered for decades to come.

And it’s not particularly difficult [0] to construct data on a bounded region of the plane where the Pearson correlation between any translation of the 1/r kernel and the data is positive, so “literally everything causes cancer”. (You need to avoid exploding due to the singularity at the center, and I assume that the authors also somehow did something about the singularity, but it’s really unclear what they did.)

I have not come up with a compelling reason why the cancer data in the analysis would have this nasty property. But, of course, the mere existence of the property is a very compelling reminder that correlation should be used with extreme caution.

[0] For discrete data it’s just linear programming.

- a complex interaction between hundreds of genes, which is just called 'bad luck' due its unknown nature

- impartial immunosuppression due to stress, lack of sleep or the hundreds of other external factors. (Without the cancer fighting immune system we would get a cancer very fast, though slower than fatal infections)

The first is barely recognizable and therefore non modifiable. We cam measure and manage only the second.

https://m.youtube.com/watch?v=EWNjw3pg9sA

Pretty much anything can cause cellular stress, including the normal operation of worn cells in an old body. Cataloguing the strength of the stressors is still useful.

The kind of crap “research” being criticized here makes me almost irrationally angry, because these people are at Harvard, and getting tons of international press for this slop - which they can crank out, precisely because they skip the hard parts that slow real researchers down. Meanwhile, hundreds of thousands of good researchers languish in anonymity, because it’s incredibly rare to find a high profile result using good methodology, and worse, it takes time.

In a just world, the people who get caught doing this stuff would be drummed out of academia.

[1] uncontrolled observational research

Even if the original authors never say anything about causation, inevitably the language of correlation (related to, is linked to, associated with, correlated with, coincides with, etc.) will turn into the language of causation (leads to, increases/decreases, has an effect on, causes, etc.) somewhere in this pipeline.

Then suddenly everybody around you is talking about how going to bed with shoes on causes headaches, or goes playing tennis increases longevity. There is some atrocious science journalism out there

The problem with nuclear is cost, not safety, and thus far it appears intractable in the real world rather than the fantasy world that the likes of BTI inhabit, whereas the problems and limitations of renewable energy are actually being solved.

The attributable number of deaths from this methodology is probably zero, but the attributable number of bad papers is at least four.

Brutal.

Nothing more to add, but sadly bunk or not these papers will surely be quoted by a lot of entities for a long time no matter how blatantly broken they are.