selectionbias

Also approximate Bayesian computation, principal stratification, and the Bayesian Bootstrap.

I used to live in Boston and visited recently for the first time in a while. The red line was so slow between Central Square and Harvard that I thought there must have been some kind of mechanical failure, but when I mentioned it to friends afterwards they said that it had been like that for months.

(Ignore my previous reply I found it myself). To be fair to the authors, it is not their primary specification, that was a linear probability model. The logit model is just a robustness check to make sure the linearity assumption isn't driving the results.

Which paper was this?

Citing a paper needn't be a celebration of it, you can cite a paper to say "these guys find X but..."

Sorry to be pedantic, but although Monte Carlo simulations are based on pseudo-randomness, I still think it is good practice that they have deterministic results (i.e., use a given seed) so that the exact results can be replicated. If the precise numbers can be reproduced then a) it helps me as a reviewer see that everything is kosher with their code and b) it means that if I tweak the code to try something out my results will be fully compatible with theirs.

If you haven't seen it already, the Farm Hall transcripts provide a fascinating insight into the failure of the Nazi nuclear effort: https://www.nybooks.com/articles/1992/08/13/the-farm-hall-tr...

FYI, the word is 'tenet', a tenant is a person who lives in a particular residence.

I sympathize with the author's disdain for brutalist and some deconstructivist work. But to suggest that all contemporary architecture is a movement towards creating deliberately alienating and impractical buildings, seems to me very wrong. I think Zaha Hadid's work (which the author critiques) is aesthetically pleasing to most people, not alienating, for example https://assets.newatlas.com/dims4/default/e09670a/2147483647... . The author expresses a uniform disdain for skyscrapers, but many skyscrapers are popular with the general public, for example, the Gherkin in London https://i.pinimg.com/474x/b5/78/c0/b578c0732b532b91b5e8455de... or say, The Empire State Building. Personally I do not find the shiny glass and sleek curves of many of these buildings unsettling or alienating nor, I think, do most people. In short, if we accept a basic premise of the author's: that what is good architecture is what is pleasing to most of the people who view and interact with it, then the author's critique is (I believe) too broad because many of the architects and buildings the author implicitly or explicitly criticizes are in fact popular.

Furthermore, something that the author does not address is the movement in contemporary architecture to carefully consider the practical effect of building design on the people within it. For example, how the flow of people is directed by the building, how the layout can help its occupants interact with each other, how interior walls can support privacy or erode it, and how to cater the response to these concerns to the function of the building. This is the opposite of the approach in Eisenman's house design mentioned in the article.

Quasi-experimental variation refers to situations in which assignment of treatments is 'as good as random' but treatments weren't assigned by the researchers in an RCT. A classic example is Angrist and Lavy https://www.google.com/url?sa=t&source=web&rct=j&url=https:/...

Well, at the very least, the argument over hypothesis formation is not specific to economics. The philosopher of science Karl Popper argued that science should aim to empirically falsify hypotheses, and that the way scientists develop those hypotheses, while an interesting psychological question, is irrelevant for the scientific method. He argued that hypothesis formation always contains an element of irrationality and instinct and he quotes Einstein who expressed views to that effect.

I admit I haven't read or even heard of Kate Raeworth, but I am an academic economist. Most modern economic research is empirical, a paper poses a policy-relevant empirical question 'did policy X reduce unemployment' and then empirical evidence is presented (perhaps using data from a randomly controlled trial or using quasi-experimental variation). The statistics are calculated and the assumptions required for the validity of the statistical analysis discussed critically and at great length. Undergraduate econ classes unfortunately leave students with the impression that academic econ is mostly unrealistic theoretical models of behavior, but those models are just handy tools for hypothesis formation, and it is the empirical testing of hypotheses that makes a discipline a science, not the manner by which those hypotheses were formed.

Look at the latest edition of QJE. You will see mostly studies addressing a particular policy question, e.g., 'what was the effect of this policy change on unemployment' which they answer using randomly controlled trials or quasi-experimental methods. Which aspect of this falls apart when you prod it?

My understanding is that medieval Rabbis (notably the Maimonides) explicitly discuss cannabis consumption and its psychological effects. The plant also gets some discussion in the Talmud but in the context of its use in fabric and as candle wick. It is speculated that certain plants mentioned in the Torah refer to cannabis but these strike me as a little tenuous. At the very least the use of cannabis as an intoxicant does not seem to be explicitly mentioned in the Torah. So I wonder, if it was commonly used for that purpose in the ancient middle east, why little to no mention of this? I mean, there is a mountain of detail about all the other minutiae of ancient custom.

I think this misses a possibly very important effect: agglomeration. Programmers in the Bay Area were not all born there, many (most?) chose to move there, some from elsewhere in the US, some from abroad. If those programmers who are more skilled tend to move to this area then the resulting greater productivity could explain higher pay. So why might more skilled programmers move to the Bay Area? Perhaps because other skilled programmers live there, and programming skills are complementary. A highly skilled developer may be worth more to a company with other skilled developers who can work together to create advanced products. Note that local PISA scores are irrelevant if people were not educated in the place that they work.

I like Schiele's work becuase I find his grimy, corpse-like figures to be very aesthetically pleasing. I do not find it 'beautiful' in the sense that an old Venetian painting of a cherub is beautiful, I simply find the deliberate ugliness pleasing to the eye. I do not find his work shocking and I do not see why I should care about the level of technical skill that went into it.

If you have a finite set of symbols, then the set of finite sequences of those symbols is countable. The key here is 'finite sequences', if you were to allow for infinite sequences then the set if uncountable.

Cities skew democrat, particularly inner-cities. Those built-up areas are also more vulnerable to the spread of disease.

The person to whom you replied meant that it now means palace but used to refer to other dwellings.

Among the first things you learn in microeconomic theory are the expected utility axioms (developed by Von Neumann and Morganstern), and Afriat's theorem. These results give conditions under which an agent's behavior is indistiguishable from utility/expected utility maximization. This is the standard justification for the use of utility maxmization in economics: that it is a good mathematical model of decision making, not that it captures what actually goes on in people's heads. This is not a heterodox idea, it has been mainstream since at least as early as the 1950s. Of course, modern economists like to empirically verify whether these models of decision-making are accurate or whether other behavioral models are more consistent with the data, because economics is a science. Praxeology, on the other hand, is the opposite of science.

I don't see how micro-founded macro models equate utility with willingness to pay? Work-horse New Keynesian models begin with a representative agent so willingness to pay of different consumers doesn't even make sense in this context. Moreover, these models are (to my knowledge) seldom used for welfare analysis but rather to examine things like the effects of montetary policy on employment and growth. Models with heterogeneous agents certainly don't assume willingness to pay is the same as utility. I'm an econometrician not a macroeconomist though so perhaps I'm missing something.

As someone has already commented, the definition of an 'efficient market' in no way equates willingness to pay with utility, and in fact economists seldom make this assumption. Pareto efficiency is very deliberately not utilitarian. A Pareto efficient outcome needn't be a 'good' outcome, it is just an outcome such that no other outcome would make everyone better off. If an outcome isn't Pareto efficient then there is room for improvement. It's worth noting that while Pareto efficiency is central to some very neat foundational concepts taught in introductory econ, modern ecenomic research uses a range of welfare measures to quantitatively evaluate policies. This includes utilitarian welfare analysis. These maybe better capture the actual ethical goals we should have when making policy decisions, but they are usually a bit ad hoc and don't lead to such neat results.

My problem with the 'Bayes=rationality' type of argument is that it ignores context and isn't really a case for reporting Bayesian vs frequentist estimates. If I am a researcher publishing results then I have an audience who interpret my results. If my audience is Bayesian and accept my model then all I need to do is report sufficient statistics and they can make their own Bayesian inferences given their priors, or better yet, I can just post my whole dataset. The very reason we need to report things like credible sets or confidence intervals rather than just sufficient statistics is because audiences in the real world want summary stats that they can easily interpret and are transparent. The best approach to inference is one that is the most useful to audiences, and that depends on context and practicalities rather than on some underlying philosophy of subjective vs objective probabilities.

Why wouldn't you just ride on the tugboat?

Both of your assertions are inaccurate. First of all the economics prize, while not an original prize, is recognized by the Nobel foundation, and that 'Swedish bank' is the central bank of Sweden.

Secondly, the idea that modern academic economics is driven by ideology and linear regression is demonstrably false. Take a look at the latest edition of QJE, the highest impact factor economics journal (link: https://academic.oup.com/qje/issue/134/4). Most of those articles are use empirical evidence to answer questions with obvious policy importance (e.g., causes of food inequality, effectiveness of workplace wellness policies). And modern econometric methods do not amount to linear regression, they include things like regression discontinuity design, instrumental variables techniques, panel data methods, that can provide convincing evidence of causal effects when only quasi-random variation is available. The Nobel prize your commenting on was awarded to researchers carrying out RCTs to assess the impact of specific interventions, how is that mere ideology and linear regression?

Conway's Game of Life is Turing complete, so if it is possible to build an artificial intelligence in a regular computer then it is possible to build one within the Game of Life. Also, understanding the basic rules of the game of life is not the same as understanding how an artificial intelligence within it functions. The rules that determine how patterns of pixels change in the game can be written down on a single sheet of paper, a description of an artificial intelligence built within the game would probably be absurdly complex.

In the context of the incompleteness theorem a 'theory' consists of a formal language to describe theorems, some primitive axioms and some rules that can be used to prove theorems from the axioms. Godel's incompleteness theorem states (loosely speaking) that if a theory is rich enough to describe the arithmetic of the natural numbers, is consistent and is 'effectively axiomatized', then there are statements that can be expressed within the theory and that are true, but that cannot be proven using the rules of deduction in the theory. In short, this is a totally different meaning of the word 'theory' to the one you are thinking of.

Suppose we are able to formulate a very neat, parsimonious mathematical model and it happens to extremely accurately describe every physical phenomenon, so accurately that we cannot find even the tiniest violation. Now, that does not mean the model is 'correct', it might be that we have just not measured precisely enough to detect its failures. But there does seem to be a mysterious tendency for very neat mathematical models to very accurately describe physics, and so maybe it is not unreasonable to conclude that this model is at least probably true (i.e., that it is never violated). Would this not then tell us something pretty profound? Even if the way we understand the math (in terms of ideal shapes, in terms of symbols) is inherently human, we would still have possibly discovered a full description of the behavior of the universe, and even if that in itself doesn't tell us why the universe exists or what it is, it would surely help us to answer those questions.

Why is the implication that this is due primarily to sexism?

Well perhaps because of the persuasive evidence the article cites from peer-reviewed journals. For example, in the paragraph below from the article:

"In one experiment, reported by Corinne A. Moss-Racusin and her coauthors in PNAS, 127 US scientists were asked to hire an undergraduate lab assistant and decide on a salary based on fictional CVs of equally qualified men and women. The scientists were more likely to offer the position to men as well as more hours of mentorship, and gave a lower salary to women, about eighty-eight cents to the dollar."