To your first point, look to the other comments, which clarify this beyond what I was aware of. To your second point, yes you are correct; I didn't feel like getting into that was necessary, but it's apparent in the image I put on the blog post.
If you do go on, I'll happily link to it in the article. There's no way I could have done everything, but I could have done a better job nodding to what's missing.
That's fascinating. I'm the author of the piece, and the material trick and its ramifications is new to me, thank you! This entire subject is not my area of expertise - I'm an historian of 17th century astronomy - but it's been fun dipping my toes into this world.
Hey, not the OP but the original author here! I was wondering why my server was going crazy; should've known it was HN. If you have issues, the page is already scraped into the Internet Archive: http://web.archive.org/web/20190116163255/http://scottbot.ne...
Carnegie Mellon Univ | Pittsburgh | Full time | Onsite
Hiring a "digital humanities developer" to support humanities & social science research. Research background is a plus, and work is mostly interactive data / web stuff, with some analysis or backend work.
If you define science as anything that's falsifiable, then of course anything that isn't, isn't science. And hey, free sciencecountry, that's your prerogative.
That said, the philosophers of science who originally discussed falsifiability have gone on to say its inadequate, and a huge chunk of work that's published and funded as science these days isn't strictly falsifiable.
Science is a word like any other, that we all agree on to make meaning. Right now lots of practicing scientists work on a definition that includes but isn't limited to falsifiability, but of course the great thing about it is you're welcome to decide what criteria is most comfortable to you. I like to think that the work Einstein did on Special Relativity before 1915, that wasn't mathematically distinct from Lorentz or Poincare, was scientific even though it wasn't falsifiable.
Pron, as a historian who has come up with quantitative models and resisted the urge to call them predictive, yes, I think that sort of restraint is within humanit(y|ie)'s capable grasp. =]
This ain't actually how science works, it's just an idealized account. Special relativity wasn't technically falsifiable compared to its competing theory by Lorentz, because both produced mathematically equivalent predictions. If you pick up a random science paper these days, odds are you'll find all sorts of unfalsifiable claims. Does this mean modern scientists are doing it wrong? No, but even if you think it does, it still shows modern historians are on the same footing.
This is true, but not entirely relevant. Life-based complex systems often share a propensity for punctuated stability specifically due to their own nature, because of the same circularity inherent in evolution (those that can survive to replicate, do). In this case, systems whose parameters tend towards stabilization persist specifically because they tend towards stabilization. The least self-undermining regularities persist (attractors).
Societies formed and persisted because they were good at it, because they were a stable attractor in a larger system. We might not necessarily be able to formally describe the entire system, but our propensity towards stabilization (at the biological level, the human level, the societal level) means we can do clever things at the stability points, like develop medicines that work, design groceries which are more likely to sell certain products, and predict the outcomes of presidential elections.
Now, whether or not the explanation given for the systemic outcomes are "accurate" descriptions of the underlying mechanisms is in question here, and it's an important one, but it's not a lost cause. When Copernicus set the world in orbit, there was a big argument of whether he was providing an actual explanation of the way the world works, or just a convenient mathematical shorthand for making accurate predictions. It turned out that the most parsimonious shorthand was also (ahem) less wrong than earlier mechanistic theories. So too can historians find explanations that, if not accurate representations of underlying mechanisms, can still be explanations which fit better to systemic tendencies than earlier explanations.
Edit: Which is just saying that the blog author's point is still a useful one, whether or not we can ever achieve complete mechanical account of human activity.
My stance is much softer than that, but I should have made it clearer, because similar arguments are often made in anti-frequentist rants. I think that there is an appropriate place for most statistics used (including NHST under the right circumstances) - and often, the differences between the results are entirely negligible.
My goal is to make people aware of the various stats out there, their benefits and pitfalls, and let people choose whatever is the most appropriate for their needs. Those choices need to be informed, and given that most introductory stats starts with p-values and seems to teach them wrong, that's where this post is aimed.
Regarding your aside, the universe of possible observations in a given experiment assuming 100 trials may be very different than the universe assuming trials until we run out of money, which happened to fall on 100.
I agree, lying is bad regardless! My post isn't anti p-values, it's anti poorly-understood-or-performed-statistics. NHST just happens to be the subject of choice, because it's particularly misunderstood. Anyway, I'm less worried about lying, and more worried about accidental inaccuracies, e.g., someone collecting data until they get tired or run out of funding, but running the calculation as though the "intent" was to get exactly that number of observations.