stdbrouw

70 ms·

571.
by stdbrouw 11y ago
The anomaly itself does not seem to be disputed. It's true that when looking for arbitrary clusters and unusual patterns you can pretty much always find something -- though it's more of a Texas Sharpshooter fallacy [1] than the ki
572.
by stdbrouw 11y ago
But (1) statistics is really not the kind of math that is being talked about in the article, (2) even if you study statistics in a graduate program, you'll be inundated with hypothesis tests and alpha=0.05 and any deeper insight just k
573.
by stdbrouw 11y ago
I'm sure you just meant to highlight that reading tutorials is often a better experience than just messing around and hoping to get somewhere (which is why I didn't downvote you), but on first reading it comes across as "if y
574.
by stdbrouw 11y ago
You have to prove a drug saves or improves lives to get it approved in the first place. With the help of survival analysis and clinical trials, we most certainly can quantify it.
575.
by stdbrouw 11y ago
It's great work. Constructive criticism can make it even better.
576.
by stdbrouw 11y ago
You're creating a false equivalence. Yes, different interpretations of the same data are possible, but that doesn't make all interpretations equally valid. Occam's razor would prefer the simple explanation that a question ask
577.
by stdbrouw 11y ago
On the one hand I love that people are thinking about how to make algebra more approachable and more usable in everyday life, but on the other hand I fear that it might mess with the learning curve. Once you want to do something that's
578.
by stdbrouw 11y ago
Feynman is great, and I'd add Meehl's paradox: http://www.fisme.science.uu.nl/staff/christianb/downloads/me... A lot of the research that relies on null hypothesis significance testing really is lou
579.
by stdbrouw 11y ago
Why? What would be more efficient or effective?
580.
by stdbrouw 11y ago
Think of a prior as what's plausible or not plausible given the current state of our knowledge, rather than as an objective state of the world. Mind you, that doesn't mean priors are necessarily subjective. In medicine, we often h
581.
by stdbrouw 11y ago
You do see the irony in criticizing a study for its imperfection when you yourself rely on nothing more than your personal guesstimate of what "most people" think and a just-so story about the possible influence of an obscure bibl
582.
by stdbrouw 11y ago
That seems a little far-fetched. You ask someone about distributing cash between them and their neighbor and of course the participant will assume that what you're talking about is their wealth relative to the sum total wealth of the
583.
by stdbrouw 11y ago
I can see why my comment could come across as arrogant and in hindsight shouldn't have written it with that tone. However, I wrote it immediately after reading https://www.optimizely.com/statistics/ which makes a
584.
by stdbrouw 11y ago
That would be best practice if you insist on using null hypothesis significance testing and only wanted to use classical frequentist statistics. We really can do much better these days with multi-armed bandits, and by focusing on effect siz
585.
by stdbrouw 11y ago
After reading the blog post and reading through the comments, it looks like people are drawing the wrong conclusion from this. The problem is not that AB-testing is overrated, doesn't work, is bullshit etc. but that Optimizely used to
586.
by stdbrouw 11y ago
Their Stats Engine does resolve the issues, but I have to laugh at their marketing materials calling it 21st century statistics, because even the new approach is comically behind the times. Looks like the poor schmucks accidentally outsourc
587.
by stdbrouw 11y ago
Nitpick, but I don't think you could still talk about an improvised explosive device at that point.
588.
by stdbrouw 11y ago
Submitting a package is the least important part of distributing it. For CRAN you can use the submit form you link to, for Python it's `python setup.py sdist upload` and you're done. CRAN in fact has much stricter standards for pa
589.
by stdbrouw 11y ago
Is it still possible to be an informed citizen? Most laws are the result of work on congressional committees and then voted on according to the party line because even politicians have had to admit that the issues are just too complex to de
590.
by stdbrouw 11y ago
Guesstimate is napkin math. It makes no sense to spend an inordinate amount of time fine-tuning the distribution when both the distribution and its parameters are just best guesses. The important part is the propagation of uncertainty acros
591.
by stdbrouw 11y ago
Most of the world's data analysis still happens on data sets that are at most a couple of thousand observations. It'd be neat if ggplot2 was faster for large datasets, but I can imagine that's not a priority.
592.
by stdbrouw 11y ago
It refers to research, heavily publicized by Freakonomics, which claims that the reduction in crime rates in the US is not because of police interventions etc. but because more people are having abortions and abortions are legal in more sta
593.
by stdbrouw 11y ago
Well, if you really want to learn how to think probabilistically in everyday life, I'd recommend Douglas Hubbard's "How to Measure Anything" which contains detailed advice for how to calibrate your estimates (so you don&
594.
by stdbrouw 11y ago
I wonder if the people upvoting the article are actually reading it. It essentially just says "be careful" without any kind of practical advice for how to think like a statistician. Two much better articles: * https://s
595.
by stdbrouw 11y ago
I think I'd prefer it over office politics, though: it doesn't happen behind your back. (Not that that's a choice anyone has to make.)
596.
by stdbrouw 11y ago
It probably depends on what you study, but I feel most professors realize this. It's why open book exams and take home exams exist, and why you're graded not just on exams but also on assignments. Nothing wrong with just using Ank
597.
by stdbrouw 11y ago
Just a nitpick, but you can't go around and say "successful people tend to read a lot of books, but anyway, that's probably just a correlation, not causal" and then proceed to make the argument that people should read bo
598.
by stdbrouw 11y ago
Another interesting opportunity is data journalism and news application development: every project is on a 1-4 week deadline and because every news app / data visualization / ... is a standalone thing with no maintenance burden, y
599.
by stdbrouw 11y ago
We have identified various causes and we know that wearing glasses actually makes shortsightedness worse, but what can you do? It's like saying "I'm struggling to understand why people die in car accidents, when we all know
600.
by stdbrouw 11y ago
Actually, doh, you're right, my bad, turns out sample quantiles are only asymptotically unbiased, and a better estimator would take some sort of weighted average of P99 and P99.5 with the weights depending on the sample size. (You stil

More ›