nonbel

94 ms·

811.
by nonbel 8y ago
Yep, that's the "switcheroo" that reverses the long successful logic of science. I'd suggest starting with Meehl 1967: >"In the physical sciences, the usual result of an improvement in experimental design, instru
812.
by nonbel 8y ago
> "A more stringent significance standard doesn't just make it more expensive, since if what you're seeing is a statistical fluctuating, it will most probably disappear upon collecting more data, not get stronger." I
813.
by nonbel 8y ago
> "At 5σ, you lose a lot of the downsides of standard p < 0.05 hypothesis testing because the p value is so low." I don't see how this addresses any of the issues, it just makes it more expensive (ceteris paribus). As
814.
by nonbel 8y ago
>"we pick the null hypothesis so that we can when we collect data we can calculate probabilities of that data assuming the null hypothesis is true." That is an awful reason to choose a null hypothesis. You should choose a null
815.
by nonbel 8y ago
>"If that were the case, we would have made a lot more discoveries by now." In physics you still have quantitative predictions done in addition to checking the null hypothesis (which I why I said that was superfluous), if NHST
816.
by nonbel 8y ago
A more stringent significance level just means it costs more money or the data is much cheaper to collect. You can look up "NHST controversy", "NHST pseudoscience", etc to find many people discussing the problems for ove
817.
by nonbel 8y ago
> "In that specific case, no, a number of existing experiments appeared to confirm the "w/o sterile neutrinos" model, so it was assumed that was correct. Just one older experiment fit the "w/ sterile neutrin
818.
by nonbel 8y ago
What are you saying is nice?
819.
by nonbel 8y ago
>"The method described is the extremely standard Hypothesis Testing method....The only difference is that in particle physics the standard isn't p < 0.05 but a 5σ significance." Yes, I know. This isn't at all a pos
820.
by nonbel 8y ago
In that case report: "these results are consistent with model A and model B, since model A has been studied much more so we will stick with that for now". There is no reason to include these additional (usually elaborate) steps. A
821.
by nonbel 8y ago
> "For instance, in your example, if they had observed 1.8x, but the uncertainty on that were 0.8x, no discovery would be claimed despite the relativistic model explaining the data better than the non-relativistic model." The
822.
by nonbel 8y ago
>"Question number one would be: Did I or did I not establish a discovery? Question number two would be: How well does my alternate model describe this discovery?" https://cds.cern.ch/record/1099994/fil
823.
by nonbel 8y ago
>"observations wrought by Newtonian physics" I'm not sure what this means. Is it different from just making an observation? Eg, "this light in the sky is of this color at that location", or "this object drop
824.
by nonbel 8y ago
>"Are you saying that it doesn't make any predictions (and is therefore unfalsifiable)?" As I understand, its the opposite. String theory "predicts" any and every possible observation, then they cherry pick whate
825.
by nonbel 8y ago
I don't know how useful this tool will be: >"This trial is ongoing. It must report results 82 years, 8 months from now." http://fdaaa.trialstracker.net/trial/NCT00068692/ Yep, they scraped the co
826.
by nonbel 8y ago
I'm not sure how they argue that ice core CO2 is representative for the earth, but on the face of it I wouldn't compare Mauna Loa measurements (from near the equator and at thousands of feet above sea level on the largest active v
827.
by nonbel 8y ago
>"Every day, I grow more and more convinced that we are just utterly screwed." Have you ever experienced "doom porn", fomo, etc? I suspect not, because it sounds like that is what you are going through. Once you have
828.
by nonbel 8y ago
Just try a problem like this: https://stats.stackexchange.com/questions/302891/hyper-param... The space of possible configurations can be simply huge, but still you can usually find a reasonable (not optimal) on
829.
by nonbel 8y ago
This sounds like standard biomed to me: - Refusal to share data - Refusal to share data analysis methods - Not keeping the basics of a double-blinded test - No reproduction, and not even the possibility of independent reproducibil
830.
by nonbel 8y ago
>"8.7*10^349236 possible configurations" Are you thinking there is only ~1/8.7e349236 viable solutions? There is no reason to expect you need to exhaust even a tiny fraction of the space to get something reasonable.
831.
by nonbel 8y ago
There can be an HIV vaccine, but not an AIDS vaccine... Its like saying there can be a "runny nose and cough" vaccine or a "rash" vaccine.
832.
by nonbel 8y ago
>"And the article hardly explains how reason has been defeated. It barely mentions the propaganda machine (Fox, Limbaugh, etc.) that played the biggest role in impugning the idea of truth." I saw the is the other day, authored
833.
by nonbel 8y ago
>"Cryptocurrency mining has to use the cheapest available power, or else it loses money because the mining rewards float at just above the cost of the world's cheapest power." How much energy is being wasted each year due
834.
by nonbel 8y ago
It sounds like you just need to use a different methodology, not that methodology is unimportant for whatever ml 101 examples that became popular like MNIST, etc.
835.
by nonbel 8y ago
>"If you use financial performance to measure the success of a model, methodology is critical." When wouldn't it be? And why does this sound exactly like the other post I just read?
836.
by nonbel 8y ago
>"When applying CS/ML to finance, methodology is absolutely crucial." When wouldnt it be?
837.
by nonbel 8y ago
And here is the pubmed link (only abstract available): https://www.ncbi.nlm.nih.gov/pubmed/29808008
838.
by nonbel 8y ago
Sure, go to Nature Medicine: https://www.nature.com/nm/ Click the first research article they show. It has various public funding sources: "The collection of the ELC and genetic data for the American samples was
839.
by nonbel 8y ago
>"6-months embargo paralyzes progress." As long as they put the code on github or a blog or whatever then who really cares? Do people really get that much out of these academic style papers in this field? They are more like a p
840.
by nonbel 8y ago
> "The article discusses the methodology in-depth including the meat of the specific points you raise." The meat of my methodological issues is that the survey has not been validated in any way as a measure of mortality. Where

More ›