> The fallacy with using tests of the kids as a marker for the quality of the teachers is that you just can't do that and get reliable results.
Which is not what is being proposed by people arguing for teacher evaluations drawing on standardized testing (https://en.wikipedia.org/wiki/Value-added_modeling), as the very name 'value-added' implies.
VAM is a good idea but it's really hard to get right and the trend has been make it very high stakes for teachers. You didn't really address the examples which mschuster91 provided and that's important to understanding the problem:
1. Limited parental support (note: this does not imply bad parents – working 3 jobs to pay the bills leaves little time to help with homework)
2. Unstable living environment
3. Strong financial restrictions
4. Need to care for siblings[1]
5. Food insecurity
How do you construct a VAM model which recognizes that a teacher who got a class full of students suffering from one or more of those problems and managed to improve them by one grade level had a LOT more work, and more complicated work, than the teacher in a wealthy suburb who got a bunch of students with affluent, involved parents who are both pushing their kids hard to excel and providing tons of extra support outside of school?
This isn't just a philosophical debate, either, since school districts are tying large parts of compensation to test scores. Starting with a hard job which doesn't pay particularly well, how many years are you going to spend not getting bonuses for your hard work or even being arbitrarily punished before you give up and find an easier job?
One estimate has ~12% of NYC public school teachers being punished by the flawed VAM in use there:
That's a high number to begin with and downright shameful when you consider that those schools are already facing a hard time getting qualified teachers. If hiring is hard, you really need to make an effort to retain the people you do manage to find.
1. My wife has had students who felt pressure to skip after-school extra-curricular activities or even go to an inferior college so they could care for younger siblings while their parents worked. That's not wrong in the sense of everyone involved having a sympathetic motive but it's a huge burden which more affluent kids never even have to think about, which is why simple-sounding ideas like making college admission or scholarships merit-based ends up reinforcing the existing socioeconomic status quo rather than changing it.
> You didn't really address the examples which mschuster91 provided and that's important to understanding the problem:
On the contrary, I addressed it entirely. mschuster91 seems to be under the impression that the teacher evaluation schemes boil down to nothing but the simplest possible before-after comparison of grades of students, ignoring all issues of demographics, differing student quality, differing school circumstances, etc. Such a scheme is indeed absurd, as his counterexample proves, but it is not what has been proposed by pretty much everyone! The actual proposals are well aware of what he thinks is the fatal problem, and go to often elaborate lengths to model and adjust for these sorts of heterogeneities in order to quantify the value-added of a particular teacher. The problem is recognized, included, and mostly dealt with. Whether the solution works entirely or is worthwhile is unclear, but he's arguing against a strawman.
> One estimate has ~12% of NYC public school teachers being punished by the flawed VAM in use there:
So I've looked at http://mathbabe.org/2015/04/02/the-arbitrary-punishment-of-n... and I have zero idea what she is trying to show. She assumes independence and treats it as a coin flip. Ummm.... what? With that sort of logic, you could show no one could expect to score a 1600 on the SAT. When criticized she links to a real analysis†, which shows considerable non-independence which means her numbers are wrong and will overstate how many will be denied tenure based on the VAMs. By the way, why are you phrasing it as 'punished'? That sounds like you're assuming your conclusion. If VAM doesn't affect hiring decisions, there's no point to bothering with it in the first place is there, but if it does affect hiring decisions, that means teachers are being 'punished'...?
† not that I think too much of it either, since it relies mostly on an argument from incredulity and pointing angrily at some scatterplots, and tries to ignore the r=.35 correlation of ratings from two subjects; to put an r=.35 in perspective, the correlation between years of education and intelligence is only ~r=.55! Even the best IQ tests won't correlate with Gf more than r=.7 or so. r=.35 is pretty good for a single pair. I don't know why he thinks a .24 is 'minuscule' when that means you're predicting half of variance... (I wonder if this is a graphing problem? He doesn't seem to jitter the datapoints, which for a large amount of discrete data will hide a lot of the density; a plot of r=.35 of n=6k should look much more striking, like this: http://imgur.com/KcwmJJH ) For implications, look at the first graph and think about classification rates. Look at the datapoints at 100 along one axis, then look across to see how many correspond to <10 on the other; hardly any do, and the 100s are almost all mapped onto 80+ on the other axis. Or look at the 0s. In terms of identifying the bottom decile, it's doing a good job.
Test motivation differs from person to person, so if you don't encourage everyone to at least try every question and guess, you'll get differences in scores which don't reflect the child's difference in knowledge (which is what the test is trying to measure) so much as willingness to try or guess. This willingness can differ systematically so you might get drastically lower scores for poorer children than they should. (This is one of the reasons schools like psychologists to do IQ tests, because they can spot when a child isn't trying or is deliberately underperforming.) So that's one reason. Another reason is that it's rare to have no idea whatsoever; even if you feel entirely uncertain, you can still often guess at above chance rates, showing that you did know more than nothing. Forced-choice methodologies expose that knowledge and again make the tests more accurate, because more items means more effective at distinguishing between students.
(Imagine a test of 10 questions, each substantially harder; one student manages to answer correctly up to question 5 before starting to feel uncertain and refusing to answer any more, and a second gets up to question 6. How sure are you that #1 knows less than #2? Now imagine that they instead 'guessed' on the remaining 5 questions, and #1 got 3/5 right and #2 got 1/4. Now how sure are you? Haven't you learned something from this apparently 'useless' guessing?)
> But in the real world, there is no 25% credit for guessing.
You can no more refuse to guess in the real world than you can refuse to make choices, take actions, or let time pass.
I'm not saying there aren't any valid reasons for doing it. I'm just saying the there's a "meta-lesson" there that has to be corrected. I want all my kids to grow up knowing that there's no shame in saying "I don't know," if you honestly don't know. Life is not a sounding smart contest.
> I'm not saying there aren't any valid reasons for doing it.
> I'm just saying the there's a "meta-lesson" there that has to be corrected.
And I'm saying that your meta-lesson is not a good idea as it will tend to teach underconfidence. The real world does not always let you off with a "I don't know"; you may not know to some high degree of certainty whether a cancer treatment is a good idea, but nevertheless you must decide whether or not to do it.
Looks like one'd describe him as an insurance adjuster:
> Kafka was rapidly promoted and his duties included processing and investigating compensation claims, writing reports, and handling appeals from businessmen who thought their firms had been placed in too high a risk category, which cost them more in insurance premiums.[41] He would compile and compose the annual report on the insurance institute for the several years he worked there.
And of course, Wallace Stevens was an insurance lawyer.
> This was a favorite zinger that my evolutionary biology teach liked to spring on unsuspecting students that tried to argue that they could demonstrate that low iq among AAs was due to genetic differences.
And what happens when AAs migrate out of America, or when Africans migrate into America, hmm? The debate is not that simplistic and easily resolved, and your professor did you a disservice by pretending that it is and not discussing why his anecdote is not airtight (for example, immigrant samples are almost always contaminated by serious selection effects which are hard to measure and vary by group). By the way, how sure are you that your professor was even right in the first place (https://unsafeharbour.wordpress.com/2012/01/13/burakumin-and...)?
You can download your genotype SNP data as a backup. I don't know about the health reports (are they so voluminous that you cannot open up each in a tab and then save them all as a batch of HTML files?) but there's a partial replacement in the form of the Promethease service which will take a SNP export and try to summarize any interesting hits.
If I scraped it myself, there'd be over 400 pages. Looks like they don't do async fetches for the main pages until you use the risk indicator (e.g. to try different age ranges to see when you're most likely to have a health problem).
So, as long as they don't ban your account for scraping, you could probably write a script to get most of it. I see personalized data in the view-source, which is a good sign (it means they server-render the personalized bit).
> The desire to believe in equality is weird sometimes.
It's not that weird. If you look back in the HN archives, you can find one or two submissions from transsexual authors who argue, essentially, that after changing genders, they were discriminated against more, and that this proves that society/Silicon Valley/etc must be extremely sexist & discriminatory because nothing else about them changed; many of the HN commenters agreed with the claims.
"adopting each other" is meaningless and the citation is vague. Some googling of The Advocate's website suggests this is probably a reference to Baker and McConnell, where the regular adoption did succeed: http://www.nytimes.com/2015/05/17/us/the-same-sex-couple-who...
So? The point remains the article specifically cites a situation where the adoption process fails, which directly contradicts the previous statement, and lends credence to the theory that the commenter hadn't actually read the majority of the article.
> The point remains the article specifically cites a situation where the adoption process fails
No, it doesn't. The article vaguely alludes to a third-hand description of the failure of a legal tactic which taken literally is nonsensical; and as far as I can tell, when those two people tried in the sensible standard manner used by everyone else (the manner in which the article is about), did succeed.
> This is completely fine if you're tooling around in a research lab or industrial lab, but even a 1% loss is probably too much for a human brain to bear and remain the same as before.
You probably lose at least that many neurons over a lifetime; consider the shrinking volume of the brain with age. And losses definitely easily exceed 1% in early Alzheimer's or dementia, but while not fun, people with early Alzheimer's clearly have not died or ceased to exist!
(And anyway, the real question about the dead cells is whether they still have the information about their synaptic weights and other functional information. Being able to revive the cell is overkill; the focus on revival is as an _a fortiori_ argument, since any revival proves that, even in the absence of future scanning advances, it must be possible.)
It would seem like there should be a difference between an operational brain losing neurons and a suspended brain losing neurons?
I'd assume that if inter-neuron communication speed is anywhere on the same order as the time-between-individual-neuron-loss, then that would alleviate the effect somewhat, no? My understanding was there was a lot of redundancy up there, so it would make sense that the brain would resilver itself in certain scenarios.
> I'd assume that if inter-neuron communication speed is anywhere on the same order as the time-between-individual-neuron-loss, then that would alleviate the effect somewhat, no? My understanding was there was a lot of redundancy up there, so it would make sense that the brain would resilver itself in certain scenarios.
The brain is remarkably resilient and adaptive; people manage to recover at least partial function even after major strokes. And, quite frankly, I'd take "functioning like a stroke victim" over "dead" any day.
It also sounds like a motivating example for machine-checked proofs: if one could feed Mochizuki's proof into Coq or something and be assured that it was correct, even if only in a purely formal sense, I suspect there would be much more interest in grappling with the concepts to understand what the proof is doing, whether it's acceptable, and why the proof is correct. As it stands, there's the risk of a wild goose chase.
Well, supposedly he did write a human-checkable proof--it's just that the reviewer must be familiar with several of his novel ideas. Like any human-intended proof, there's an assumption of foreknowledge.
That the Russians have been doing it for years and it's been so slow to catch on suggests the opposite, for the same reason that the increasing spread of autonomous cars is not a sign of the decay & fall of the West but its continued technological vitality.
Which is not what is being proposed by people arguing for teacher evaluations drawing on standardized testing (https://en.wikipedia.org/wiki/Value-added_modeling), as the very name 'value-added' implies.