Morning Ed: Education {2017.06.19.M}

[Ed1] Yasmin Nair is concerned that there are no renegade academics anymore. This touches on a point I’m coming to, which is for all the conservative hair-ripping about the leftward drift of academics, I think it’s more of a consolidation towards a sort of High Liberalism that is also giving the old school left the shaft. It’s pretty easy to learn new things, but plumbing can be very difficult so if you just moved into a new dorm or house and are having issues with the pipes, then visit Peoria Plumbing Company to get some assistance. If you’re in the California and need orange county plumbing, visit their page to see what they can do for you.

[Ed2] It’s not enough to merely be a minority. College recruiters also expect you to play the part.

[Ed3] Awwww.

[Ed4] Behold, the power of liberal arts degrees at tech companies. {More}

[Ed5] Freddie points to a study suggesting trade schools have their limits.

[Ed6] While many here seek to reform our education system, parents abroad want to import it.

[Ed7] One neat little trick to paying off student loans: Generous parents.

[Ed8] The world of student debt forgiveness for physicians is really quick tricky (we could have paid off the student loans years ago but it wasn’t a good idea to!), but sounds more simple than compared to that for teachers.

[Ed9] The correlation (or lack thereof) between income and education spending by state:


Managing Editor
Home Page Twitter Google+ Pinterest 

Will Truman is a former professional gearhead who is presently a stay-at-home father in the Mountain East. He has moved around frequently, having lived in six places since 2003, ranging from rural outposts to major metropolitan areas. He also writes fiction, when he finds the time. ...more →

Please do be so kind as to share this post.
TwitterFacebookRedditEmailPrintFriendlyMore options

129 thoughts on “Morning Ed: Education {2017.06.19.M}

  1. Ed1: Many conservatives see anything to their left as dangerously radical. There is simply no difference between a liberal Democratic voting professor and a bomb-throwing Anarcho-Marxist. Its all the same to them. Professors can have theories that are out there but many of them are either gripping for survival as adjuncts or living a typical middle to upper middle class American life if they have tenure.

    Ed2: This seems like a law suit waiting to happen.

    Ed3: Why did they need to start with a teenage girl when the story is about the aid dog of a male student? The girl is unnecessary to the story.

    Ed4: All hail the liberal arts.

    Ed5: This relates to Ed4 and is also my criticism about trade school. Trade schools teach jobs you how to do a particular job but its really impossible to determine whether that job is going to be useful in the future. You learn how to be a wielder and a few years after your graduation, somebody events a wielding robot and eliminates your job. A more academic education can lead to greater flexibility when it comes to an uncertain economy.

    Ed6: The grass is always greener on the other side of the Atlantic or Pacific Ocean.

    Ed7: If your born to the right set of really generous parents you get to avoid student debt altogether.

    Report

    • Ed3 because it was her tweet that went viral and was the source for how buzzfeed picked up this story.

      Hey, they’re no longer recycling #content without credit. I’ll call that progress.

      Report

    • Ed7: If your born to the right set of really generous parents you get to avoid student debt altogether.

      True. I’m sure my Dad could have afforded to pay for all of my college. He choose not to.

      Report

    • The problem is in todays world one needs to know how to train themselves from written (and today video) presentations, not being spoon fed in class sessions. This applies in particular in the trades as things change rapidly. Consider plumbing if you go into the repair and remodel part of the business you have to deal with systems ranging from lead pipes thru copper and galvanized to plastic in the tree model, to the newest buildings with plumbing manifolds, flexible plastic pipes and the like. (one the waste side cast iron as well as plastic as well as some lead and copper as well). Then there is whatever the next plumbing idea that is adopted. Let alone talk about someone who learned auto mechanics in the late 1960s and did not keep up with changes, about all they can do today is change oil and tires.
      Of course keeping up does take time and many folks appear not to have the time to keep up.

      Report

      • Not having time to keep up seems to be an issue in all trades and professions. One explanation I’ve heard for the drug pushing done on television is that doctors tend to stick with the drugs they learned about during the medical school and residency years rather than the new drugs. Pharmacy companies do not like that. Doctors and lawyers have classes to help them deal with changes in their field and lawyers can hit hard if they don’t keep up with the changes so that helps. Plumbers, auto mechanics, and others seem to have a more difficult time.

        Report

  2. Ed2: Best comment in that article: “Diversity is an elitist term used to give respectability to acts and policy that would otherwise be deemed as racism.” But seriously, what would you expect? Basically, this is “well, he’s minority X, but he’s acting “white”. Can’t have that!

    [Ed4] There seems decent anecdotal evidence for this. My ex was a literature major. She ended up working in IT, under a guy who had a frech lit major. Then later on, she became a director of Business Analytics at company formed by a group of doctors. But I don’t think this is so common that people should rely on it.

    [Ed7] I remember my college days. The key thing for me was finding a very good job. I started out with my parents paying for my classes. I was living at home since my dad could not afford on campus housing, at least for the first year. The second year I was on campus and that summer my dad found a job for me (it was find a job or i’ll find one for you. of course i didn’t) So I worked in a auto factory assembling cars. I did that in to my grad school days. However, the old man said “since you’re making good money, you can pay for half of your tuition going forward (last 2 years of undergrad). And I did. For two years I paid every other semester’s tuition while living on campus, all my books for the year, 6 months of car payments, all the insurance, gas, maint. on “my” car, etc. I also had a job on campus during school tutoring to earn “beer money”. Dad told me grad school was “on me”. Your parents don’t need to pay off your debt, but helping you find a job, oh yeah baby!

    Report

    • Ed7: Both my parents had part time factory work for part of their college days. They tended to be pretty contemptuous of a lot of the jobs that teenagers did to earn pocket money but said that they really wished Saul and I could have done some factory work for a summer because that is real work. The days of part time factory for UMC youth were long gone by the time we were teens though.

      Report

      • Well, I wasn’t “part time” in the sense that I was working less than 40 hour a week. I was “part time” in the sense that I was not working the full year. I worked 6 10 hour shifts a week, 4pm to 2AM, and every other Sunday. I had to join the UAW and pay union dues–my state is a closed shop state. On the positive side, at least from my dad’s perspective, I wasn’t partying.

        You might recall the Astro and Safari minivans that were on the roads back in the late 80s-early 90s? I build them. Hell, it was better than working for Parks and Recreation and emptying garbage cans that had “fermented” all weekend. It was a good day when the garbage “juice” didn’t run down to my elbow.

        Report

    • Ed4: It used to be much more common in the early days of STEM because there weren’t that many people with degrees in computer science or related fields yet. Other businesses also used to have a much higher concentration of people who studied the humanities at college rather than business majors.

      Report

      • Indeed. My ex didn’t get the job because she was good with computers, I think she got it because she was sharp, could think and write well, and and had damn good common sense.

        Report

  3. E1: It would have been nice to have some examples of renegade academic’s from the past. I also wonder if adjunctification is a cause of the demise of the radical academic (if there ever was such a thing). Today’s academics generally don’t have the luxury or security of tenure or even being tenure-track. This means they need side gigs and this tends to be explaining their work to a general audience. This seems to involve a lot of connections to pop-culture or trying to become a pop-culture figure yourself. LGM has a new poster who calls herself a digital anthropologist. A lot of her posts seem to be about what memes and gifs (two things I loathe) say about internet culture and society. Younger academics seem to have no patience for the old pop-culture hating academics of the past.

    Ed6: I think your statement is somewhat misleading. These International Schools are not taking their curriculum from U.S. public schools but there is a demand for an English-based education and this gives opportunity to American teachers. We slash our budgets and the rich of other nations get our teachers.

    Report

  4. Ed5: here is the key:

    In broad strokes, vocational/tech training helps you get a job right out of school, but hurts you as you go along later in life:

    The strength of University is that it generally inspires a lifetime love of learning. So vocational training itself isn’t bad, but it also needs to try to instill that love of learning, or at least be blunt with students that flexibility is critical & they need to keep learning new skills at every opportunity.

    Report

    • “The strength of University is that it generally inspires a lifetime love of learning.”

      Does it or does this just exist in a group of people anyway? I’m a cynic and think that there are a lot of people who go to university because they know it is necessary for a middle-class or above lifestyle. Not necessarily because they have any love of learning.

      Report

      • A friend of mine posted on FB this weekend how college was a terribly waste of time and that she learned more, what was learned was not said with specifics but probably has to do more with the school of hard knocks than book learning, outside of college than in it. This was so horribly offensive that I decided not to respond but its indicative how many people think. Most people go to college or university because they want that degree signal or a glorified vocational education to earn money. Few go because they love learning. Abstract learning for its own sake in the arts, humanities, or sciences is not that popular.

        Report

        • How is that anymore offensive than saying that schooling should only be for the love of learning, not signaling that you can jump though hoops or “a glorified vocational education to earn money”? Schooling should, ideally, be for what ever reason you want and be contemplated in any form available, formal or informal.

          Report

      • I was once told, while in grad school, “it’s the unusual students who become professors” meaning, those of us In TAships planning on going into college teaching should not be too discouraged when many of the students don’t seem that interested in a lot of the topics or only seem to care about “credentialing.” (I once had a student ask me, during a process we were doing in lab, “Is this what pathologists do?” and that seemed an awfully random question to me until a labmate showed me a magazine article ranking pay levels of careers and “pathologist” was close to the top of the biology ones.)

        Report

      • “Love of learning” happens along a gradient. It’s not the case that a person is a serial student/ravenous autodidact or a lump of unthinking flesh. A person can be eager to learn more about a narrow range of topics. For instance, I love learning more about the physical sciences and programming languages, but have very little interest in high finance or the various social studies topics, beyond a high level curiosity.

        For a tradesman, it used to be that a person could learn a trade (e.g. carpentry), and learn little else,can live a middle class life. Now, unless your skills are artisan level, you’d better be learning about new construction techniques and thinking hard about how your old skills can translate to new materials or techniques. This is what vocational schools need to include, and it’s something that the apprentice & union models were highly inconsistent with.

        Report

        • Add to that the travel between apprentice and journeyman is a formal educational experience quite as intense in many fields as a university degree. And the distance between journeyman and master takes a lifetime of learning to cross.

          One of the things most people don’t realize is that trade school only gets you to the point of being accepted to become an apprentice in many fields.

          Report

          • And not everyone can be a Master. It’s like professional athletes – what is your plan if you don’t go pro, or if you blow out your knee in the first season?

            Hell, I still run across degree holders who’ve been out of school a while who have given no thought to what they could do should the market for their current wage earning skill set dry up? Maybe my wife and I are different, but we constantly talk to each other about what we could do should we find ourselves out of work in a tight market for the current skill set? I mean, the likelihood of that in a major metro area like Seattle is pretty slim, but it’s good practice and it keeps us thinking about the skills we have and the skills we want to develop.

            Report

            • that’s actually one of my worries. I’m good at teaching and apparently good at a particular subset of research but – especially last year when my state was in a death-spiral, budget-wise, I went through a lot of anguish over “How will I earn my bread if my university closes?”

              Friends reassured me: you’re smart, you’re responsible, you have common sense, you are good at learning quickly, you could do ANYTHING. But I wasn’t so sure.

              (One friend keeps telling me I’d make an awesome patent examiner but apparently that would require me moving to the DC area – which means I’d probably need three roommates in order to be able afford housing, and I am too old for rando roommates.)

              I’d probably have to move which would have a whole ‘nother series of bad ripples and bad unintended consequences (e.g.: I own my house here outright because I bought a fixer-upper in a depressed housing market).

              I was actually threatening at one point that I’d learn to deal blackjack and work at the casino, but given how I feel after a day’s interaction with students, I think that sort of career would kill me faster than using up all my savings and slowly starving to death would.

              Right now things are looking up a bit so I’m back to plan A: riding this career out until I can retire. Whether that’s at 72 like my dad did or 60 (when I minimally can) will depend on a lot of things, some of them external to me.

              Report

      • “The strength of University is that it generally inspires a lifetime love of learning.”

        I’m a cynic, and think that this has never been true. Sure, some students at the University of Paris in the Middle Ages would rather buy a book than eat, but these were always the exception. Higher education has always been a career path or a way to get out of the house for many students.

        This was certainly my observation as an undergrad in the early 1980s. There were some of us that regarded the course catalog as a candy shop, but we were a small fraction of the student body. With a student body in the five figures there were rather a lot of us, and we tended to find each other and congregate, but the drink-and-fornicate crowd was far larger.

        Report

        • And it was actively discouraged until fairly recently. The Gentleman’s C was a thing because doing any better showed you were an ungraceful grind (and probably a poor scholarship student without social standing) or a dullard.

          Now this has changed but I still run into people (including smart people) who find the smallness of undergrad experience (SLAC with around 525 in my year) and grad school experience (1 of 9 directors for three years and this is large for a directing cohort) to be too small because they can’t just be a face in the crowd. Large lectures are a feature and not a bug for many.

          Report

          • You even had this in the United Kingdom but the slang for a student excessively worried about their grades was swot rather than grind. Prime Minister Wilson and Health were not popular in their college days at Oxford because they were swots. Good grades were not mandatory at many elite colleges and universities in the Anglo-world until after World War II.

            Report

    • Constant love of learning is great, and for those who benefit from it, good for them. There is a significant number of us that have problems with memory interference. Anything new we learn runs the risk of paralleling something else in memory, running the risk of cross over. It’s not a problem for those whos mind works to parse memory cleanly, but not all brains work the same.

      from Wired:

      “Though memory interference is well documented, researchers like Cowan are still guessing at the phenomenon’s neural mechanics.

      “I assume it happens because two different ideas that are similar have similarities in the patterns of brain activity,” he says. “Your brain has to settle into the right pattern, and if you are confused, your memory can fail when you settle on the wrong pattern.” If you’re learning two related languages at the same time, say Portuguese and Spanish, you may find words from one invading the sovereign territory of the other. It’s not that you’re out of hard drive space; you’re just learning to sort and group the information you’ve newly acquired. Theoretically, your storage capacity for long-term memories is endless.

      You possess a different kind of memory, though, known as working or short-term memory—and that kind easily fills to capacity and overloads. Juggling more than just a few pieces of information in your head at once is really hard. Throw one item too many into the mix and you’ve forgotten the name of the person you were just introduced to, or lost the idea you had before you got that phone call.

      Researchers can count the items people can retain in short-term memory. And it’s not a lot. When asked to remember colored spots on a computer screen, for example, most people can only remember three or four of them. Tasked with remembering random letters, most people max out at seven. “But if I ask you to remember the letters ‘CIA, FBI and IRS,'” says Cowan, “you can remember those nine letters no problem. But because they’re not meaningless, but are grouped in your mind, it’s really like remembering just three items.””

      Report

        • I’d say part of learning is coming to understand your own cognitive makeup, which all seems to fall under “learning to learn.”

          Now, the degree that modern universities actually understand and foster multiple learning strategies — so far as I can see (from anecdote), they don’t do a very good job.

          Report

          • And even they are making an effort compared to primary schools. As I’ve said before, I was in college before I figured out ‘how’ I learn, rather than just deciding I had some mild learning disability.

            Report

            • — Well, I suppose my anecdotes are worth what you’re paying for them. That said, I have a number of neuro-diverse friends in university, and basically professors know their subject matter, but they have no real insight on how to teach it to a varied population. So if you’re “weird-brained,” you’ll have to figure out your own methods.

              Report

              • Yep, and Universities are better situated to identify and attempt to work with “weird-brained” students.

                Primary schools (even private ones), not so much. If parents aren’t aware and able to get some outside help to try and figure something out, you get a kid that struggles.

                Report

              • then again, the professoriate tends to attract “weird-brained” people. I’ve known profs who were diagnosed as high-functioning Aspergers. And I myself am kind of compulsive and overly literal-minded.

                What seems to happen, esp. at the upper-level-undergrad or grad-student level, is you seek out profs who are weird-brained in the same way that you are, and you learn well from them.

                That said: I wish I had been required to get some background in psychology of one sort or another before becoming a prof. I’ve filled in where I can with programs (most recently one on “de-escalating when people are angry”) but I’ve had students I just had a difficult time working with.

                I also think it’s easier in situations of small class size – it’s hard to adjust for different styles in a class of 150, much simpler in a class of 15.

                Report

  5. Ed2 – I don’t really see anything alarming about any of those comments from admissions officers. I’m 100% certain they are taken out of context. There is no evidence that applicants are being punished for not being sterotypical minorities, merely that some discretion and skepticism is applied towards claims of minority status on college admissions applications.

    Ed4 – I’ve written about this before. The ability to write and communicate well is paramount to nearly every occupation. This is why it is de facto the most important skill we learn in primary school. Majoring in petroleum engineering for instance is less widely applicable, unless petroleum engineering is all you do for the rest of your life.

    Report

  6. Ed5: The thing about college vs. trade school is that the kind of kid that’s interested in a trade isn’t the kind of kid that’s interested in the liberal arts, or at all likely to be successful in that kind of environment anyway. It seems both pointless and counterproductive to push them in that direction.

    What’s needed, seems to me, is for the trade schools themselves to be a bit less specialized to possibly make it easier for the graduate to shift gears as necessary.

    Report

  7. [Ed4] Of the two pieces offered on the value of liberal arts in “Silicon Valley”, the first offers stories, telling them and understanding them, as a key skill that has value. I agree. I’m not surprised at the admission that Kristin Peterson spent a fair bit of time crying in the bathroom after taking a job with a bank. But she learned that milieu and brought her value to it.

    The second claims, it seems to me, that liberal arts give one moral superiority, by virtue of being trained to be more critical. As you might guess, I’m not all that friendly to it.

    To be fair to Tom Slee, what he does is read books and write about them, and he’s responding to stuff written in a couple of books, the first of which, The Fuzzy and the Techie seems to take as its thesis that the fuzzy is morally superior to the techie. Slee rightly identifies the “post-political” attitude that I’ve seen here. However, I’ve seen it elsewhere, too.

    That he writes paragraphs like this is irritating:

    If you think, as Scott Hartley does, that making the U.S. military more efficient is making the world a better place, and that the only problem with the U.S./Israeli Stuxnet cyberwarfare virus is that “the code could be repurposed by our enemies and turned against our own infrastructure,” then we have different views on what “making our technology more ethical” involves.

    Apparently he thinks there’s some other problem with a virus that takes down machines that are of use mainly to build nuclear weapons is a bad idea? I can’t tell, he assumes that I know what he thinks are the other problems with Stuxnet, and his failure to mention it comes across as moral snobbishness. Perhaps he was thinking, “Well of course everyone knows what the moral critique is, so I’m not going to insult them (or waste my word count) by telling them.” But that just makes it into a substanceless backhanded swipe. It’s not the only one.

    Then this:

    For many of us, Silicon Valley’s “Don’t Be Evil” proclamations have lost credibility as the technology industry has grown from scrappy underdog to seemingly unaccountable colossus. Hartley interprets the industry’s stumbles only as well-meaning mistakes rather than any sign of deeper problems.

    “Don’t Be Evil” is, first and foremost, an internally-directed manifesto at Google. It wasn’t meant to be a marketing slogan, but as an aspiration, and a central value of the company. I know someone who recently changed jobs from Google and Uber, and who reported a bracing change in culture between the two. Set against Google’s “Don’t Be Evil”, Uber’s cultural slogan might be characterized as “Kill, Kill, Kill”.

    Does The Fuzzy and the Techie fail to engage with these questions in a valuable and meaningful way? I’m ready to accept the reviewer’s word that that’s the case. But then, neither has he.

    The portion of the review focused on “What Do Algorithm’s Want” is much better, though the anthropomorphism is annoying. I wonder though, if the author realizes, while describing how algorithms can become gatekeepers, that he’s making the case for Net Neutrality?

    Report

  8. Interesting abstract on the correlation between Right To Carry laws and violent crime. Not a good result for the more guns = less violence argument.

    Ten years after the adoption of RTC [Right to Carry] laws, violent crime is estimated to be 13-15% percent higher than it would have been without the RTC law.

    Unfortunately, the body of the paper is behind a pay-wall.

    Report

    • Again your making a stand on social objectivity.

      Right to carry is based more closely on the individual construct of self defense, which is likely a strong preference in rule of law.

      The concept of banning a defensive weapon for reasons of ‘social objective ideals’ creates a conflict between individual constructs and social constructs.

      I’m not saying one has to buy their burial plot and make all the arrangments, if your going to die on the hill of social objectivity, but it helps.

      Report

      • That’s how liberals usually play their game. If it’s a right they like then its deeply enshrined in our constitution and is a fundamental right. If not, they talk about how it’s a social burden.

        Report

      • Again your making a stand on social objectivity.

        I’m not making a stand on anything, Joe, except that our best understanding of how the world works is empirically based. Evidence, the scientific method, all that jazz. And in this case, the bare evidence of this study shows a correlation between RTC and increased violent crime rates.

        I mean, you can explain it away as a social construct and so on, but doing so amounts to using an ideological filter to determine reality, which – ironically – is exactly what you’re accusing me of doing. But I’m not. I’m just citing the paper.

        Or think of it this way: had the writers of the writers of the paper understood your distinction between an individual and a social construct the contents of the paper wouldn’t have been any different, since the domains between those two things are radically different. The first is empirical, the second ideological.

        Report

            • No, it doesn’t but somehow I suspect that in the next round of attempts at gun control whether by the legislature or the courts, an enterprising lawyer will try and use such a study. I would as a lawyer. The plaintiff in the Brown v. Board of Ed used psychology studies, so why not something like this?

              Report

        • As before, I said it is a good idea to seperate political science and non political science. That was not heeded for fears of obscurification.

          To say that social objectivity is scientifically empirical, and that individual constructs is ideological is a nice trick. Prove it.

          Report

          • To say that social objectivity is scientifically empirical, and that individual constructs is ideological is a nice trick. Prove it.

            Evidence and ideology are two different domains, Joe. Evidence is the presentation of a fact – that Trump went to Camp David this weekend for the first time as President – while ideology is an explanatory account derived from basic principles (usually first principles) about how the world does or should work.

            The whole theory of individual constructs – by definition – is in the domain of theory and ideology: it’s a conceptual tool used, in conjunction with the concept of social constructs, to explain and account for social dynamics. It’s no different in that sense than a purely materialistic account of social dynamics which (tries to) reduce all human social activity to (eg) wealth acquisition. But it is different in one important way: while material objects are observable, and the human desire for them is also observable, social and individual constructs are not observable. They’re inferred (insofar as they exist) from observation but only by applying those concepts to observable reality. So even on your own theory, social reality is objective and distinct from the explanatory account you invoke to account for it.

            Report

              • Would it be ideology based or evidence based if police wanted to work in a public that was disarmed?

                Insofar as the claim is true, it’s empirical. It’s just a fact (supposing it is) that police want a disarmed public.

                Report

      • The concept of banning a defensive weapon for reasons of ‘social objective ideals’ creates a conflict between individual constructs and social constructs.

        Who said anything about banning defensive weapons? Or banning anything?

        The policy-relevance of the paper is that one of the ideologically-based arguments in favor of RTC is undermined: that increasing the number of carriers will decrease violent crime rates. If the paper is correct, it doesn’t.

        Add: it’s an example of evidence running counter to ideological presuppositions. Whether people take that evidence into account when forming their beliefs is another matter, but to the extent they don’t they’re anti-empirical/anti-evidence. Which, in my view, pretty well defines US conservatism these days.

        Report

          • Right with facts and evidence about humans but wrong about humans??? That makes no sense. If the facts and evidence about humans is correct than it is correct. Period.

            This is one of those gob smacking debates that comes up every now and then ( on all sides) where people just ideologically refuse to even accept the existence of evidence that they don’t’ want to see. That is always bad. It doesn’t even mean your values about self-defense are wrong or that RTC isn’t’ the best policy. But if some evidence doesn’t support your preferred policy than shunning it as wrong think is a bad road to go down.

            Just about every complex policy has at least some mixed outcomes. There are no perfect policies, nothing is all sparkles and rainbows ( except a bag full of sparkles and rainbows, but i’m sure someone would complain the bag is paper or plastic or something).

            Report

            • What really pisses me off is someone telling me it’s sparkles and rainbows to ten decimal places. At some point I get in a yelling match to let go of my leg.

              Let’s for a minute talk about crime. Do you know empirically that a crime occurred, or was it even determined empirically? Was it left up to arresting officers in a human environment dealing with other humans? What kind of empirical controls over data do you got there? Maybe a officer is more likely to make an arrest if some other parameter is present.

              I continue to question your astounding certainty.

              Report

              • Ummm huh??? This is meaningless. Data and results can certainly be questioned. That is the scientific process. Denying results you don’t like because they gave results you don’t’ like isn’t a good look for anyone.

                Report

                  • The reasoning behind performing the study isn’t questionable – people made a claim, the study examines that claim.

                    I question how they determined a causal link. Did they make some pretty reaching assumptions regarding the statistics they used, or did they have some pretty concrete evidence (it sounds like it’s statistical assumptions, but how weak those assumptions are is something I’d need to see the study itself to determine, and I’m not going to spend $5 when someone else will in pretty short order).

                    Report

                    • The reason for the study is most likely to attempt to produce social policy leading to social control.

                      If people didn’t think they could control other people with social policy, this study wouldn’t exist.

                      Report

                    • If I’m interpreting the abstract correctly they’re using a difference-in-differences approach. Basically they look at the crime rates in an area that passes RTC laws before and after the change, and compare that difference with the difference in national crime rates over that period. The logic being that if an area’s crime rates change in a way that’s different to the national picture that implies somethign about RTC, maybe not in a single case, but if you look at many cases you will get a picture of what is happening.

                      It’s not perfect (after all, what is?), but its a decent method for something like this.

                      Report

                      • So the strength of the claim using such a method would depend on size and variety of areas examined plus the ability of the researchers to control for confounding factors (as well as definitions of what constitutes an event of interest). And the method would offer no information regarding positive causal links, only strong or weak correlation.

                        Report

                        • The reason I am pushing back on this isn’t the study as such, it is the claim that RTC carry itself is to blame. If they had just said, “look, your claim is RTC would reduce violent crime, and it’s clearly not”, I’d be OK with that. But claim causation based upon a difference-in-difference study would require a much closer review of the crimes in question.

                          Report

                        • In other words, they controlled for general trends, but can still only show correlation.

                          Perhaps the cultural factors that lead an area to adopt “RTC” laws also cause such areas to respond differently to changing social conditions. I dunno. We can try to identify and test. It is possible to infer causation from observational data, although it is non-trivial to do. However, if they didn’t do that, they didn’t do that.

                          Correlation ain’t causation. Repeat. Repeat again. This is stats 101.

                          Report

                          • Yeah, I don’t know, since I don’t have the paper, just the summary. It’s possible they did something that let’s them claim strong evidence for causation.

                            That said, one of the authors, John Donohue, is a principle author on a number of studies that make claims against firearm carry, and has penned a number of op-eds against firearm carry, so I can’t help but wonder if bias & some healthy p-hacking isn’t also playing a part in this.

                            Report

                              • NoOoOoOo, never!

                                You’d think that despite all the press the issue with p-hacking has gotten over the past year or so, people would stop doing it, or be more mindful of it, but what I hear, it’s still a big problem. Because the scientific publishing incentives continue to be buggered all to hell.

                                Report

                          • That is a distinct possibility, for example if areas with rising crime rates were more inclined to adopt RTC, that could bias the result. That could be controlled for, but that’s where Researcher Degrees of Freedom come in.

                            Report

                  • So explain your issues with the research methodology then. But saying you object to the question gives away the game. As Oscar said the question is completely valid. It is assessing a claim about how a policy works in the world. That is the kind of question research is used for. But if you don’t even believe the question can be asked your mind is shut to even asking a question that might lead to evidence you don’t’ want to see. That isn’t’ the way to finding workable polices for anyone. No one has the main line to The One True Policy that has all the answers.

                    Report

                    • Man I’m giving you more a hard time than I probably should. Me being who I am, I question whether the final policy will go something like this:

                      We found man to be violent.

                      Ok, let’s ban violence.

                      Damn, that didn’t work.

                      Ok, how many men can we throw in a cage?

                      Damn near all of them if we write enough policy that no man is innocent.

                      Make it so.

                      Report

                      • You think this is giving me a hard time. (shrugs) What “final policy” are we talking about here. Some research peeps looked into a claim about RTC and you are going to somewhere else to talk about something else. Look if there is data that doesn’t’ fit your beliefs ignoring it only make you look bad. It’s what ideologues of all stripes do. And they all look bad when they do it.

                        Report

          • Unfortunately for those humans, they’re bouncing around in a real physical world where objective reality interacts with them whether they like it or not.

            Report

            • I’ve been thinking more about gender distribution in a population limiting families to not having kids after a boy is born and I’m still puzzled about the answer. So help me explain where I’m wrong. Here’s the part I’m still stuck on: as long as we include limiting factors on procreation like age or desire, the tail ends of the distribution will have a higher probability for girls than for boys since having more girls is open whereas having more boys is closed. So consider those tail ends: the four mothers who’ve had (say) 12 kids, all girls. The most likely outcome on the next round of births is 2 boys and two girls (and iterated from there). But (and this is where it gets sticky for me), the odds of having having four boys equals the odds of having for girls. In the first case, the process ends with boys at a +4 and in the second with the process continuing with girls temporarily at +4. But at that point, the likelihood of 4 boys and 4 girls is equal, and two of each merely maintains girls at +4. Which suggests to me that at the tail end there’s a higher likelihood of more girls than boys. IOW, insofar as the restriction is “no more kids after a boy”
              rather than “must have kids till a boy”, we’ll get an open end on girls.

              Report

              • Part of the problem seems to be that as people look at the far out tail possibilities (40 girls and one boy), they forget that at iteration zero, fully 1/2 of the population has one boy and zero girls. You could do the probability calculation: 50% of households have zero girls, 25% have one girl, 12.5% have two girls and so on, while all households have exactly one boy (assuming for the sake of the model that moms live for ever and can crank out a million kids). On one hand, you have the households with a long series of girls. On the other hand you have half of households having zero girls and one boy.

                A more intuitive way to think about it is to line them all up in time and pretend everybody has kids in “rounds” at the same time. The first round will produce 50/50 boys and girls. The boy households will stop. The other half of the population will try the game again. The existing population of kids is 50/50 looking back. The next iteration, it’s again 50/50 with 25% of the households continuing. But no matter how far you go, each generation is 50/50 and looking back, the existing pool is 50/50. The number of iterations makes no difference because it’s always 50/50 up to that iteration and the expected result of each iteration is 50/50. Going the intuitive route kind of shows via poor man’s induction that the late iterations don’t change anything because they’re recursively the same as the ones before them.

                Oddly, I find the Monty Hall problem easier to summarize than this one for some reason.

                Report

                • Let me try again. Consider a distribution which conforms to probability over the course of all iterations until only 8 birth chains are still in play. We’d predict that iteration (It0) to result in four boys and four girls. Let’s suppose that happens. The next iteration – It1 – will take place over four chains. In that chain we could have a distribution conforming to the odds of 2 boys and 2 girls, or we could 1,3, or 3,1 or 4,0 or 0,4.

                  So from a point prior to It1 we know that the we have equal probabilities of that iteration resulting in +4 boys and +4 girls. If four boys, the iterations end (since every chain has terminated), and +4 is the largest differential boys can achieve over girls. If four girls, tho, the iterations don’t end even tho girls now outnumber boys by +4.

                  In the case that it’s all girls, run iteration It2 over a range of four births again with the same probabilities regarding the distributions. If all boys, then the chains terminate with equal number of boys and girls (boys were -4 at It1). If all girls, girls now outnumber boys by +8 (girls were +4 at It1).

                  It seems to me that out edges of the distribution where the sample size is very small the normal probabilities allow for more girls than more boys, not in any particular set but as a function lower probability events occurring in an iterated game. That is, at the tail end, comprised of only 4 active chains (where the distribution conforms to expectation), the greatest boy differential cannot be more than +4 whereas it could be higher than that for girls.

                  Report

                  • This is a case where your intuition is wrong, and so the correct course of action (besides just letting it go) should be to explore why your intuition is wrong. As long as you’re clinging to the idea that your intuition is right, you’re not going to resolve it, and you’re not going to accept any of the explanations anyone offers — so it’s a waste of time for anyone to try. This is a good analogue to a lot of the political discussion that happens here.

                    But the obvious next step, if you were genuinely interested, would be to explode out your example for all the different possibilities and see what happens.

                    Report

                    • The real problem here is you can make an argument for a slight advantage either way depending on your choice of simplifying assumptions.

                      The true sex ratio at birth isn’t actually 50-50 but slightly favors boys. Menopause exists and there’s a variance on the age of onset. Likewise, death exists, as does homosexuality, variations in fecundity, etc.

                      So you have all these complexifications that you assume away for the sake of tractibility of analysis but then get your tire wrapped around the axle worrying about the “real” probability distribution out at the tippy tail of a completely unrealistic scenario.

                      I mean, if you’re gonna abstractify this thing out to the point of assuming essentially unlimited rounds of child-bearing then why assume a finite population? With an infinite population pool and assuming immortality and non-diminishing fecundity and all the rest, you’re just looking at a classic infinite series, limit to infinity, problem and the correct answer is… 50-50.

                      Report

                      • The true sex ratio at birth isn’t actually 50-50 but slightly favors boys

                        I think the creator of this puzzle assumed it was equivalent to a coin flip. It’s a different variation of the “a woman has had four babies and they’ve all been boys – what’s the probability that the next one will also be a boy” puzzle. Theoretically you’d need to account for a bunch of biological possibilities, but the puzzle is assuming coin flip and just trying to trick you into saying 1:2^5 instead of 1:2.

                        Report

                        • I think the creator of this puzzle assumed it was equivalent to a coin flip

                          Yes, that’s right. Let me try (yet) again to make the point which you say is an correct intuition and (seriously) tell me where I’m wrong, since this intuition seems blindingly obvious to me.

                          The limitation on having more kids after a boy implies that for any single birth sequence, the maximum boy differential is 1:0 while the maximum girl differential is potentially infinity:0. That means it’s possible for girls to outnumber boys by a larger margin than boys outnumber girls in that sequence. So, go to the limiting case in which there’s only one live chain and the rest of the distribution aligns with expectation: that one sequence tilts in favor of more girls than boys. That is, there’s a 50% chance the next birth tilts the overall numbers to +1 for boys (and a 50% chance that it tilts + 1 for girls!). Nothing interesting there. But there’s also a 25% chance that the subsequent iteration tilts +2 for girls, and 6.25% chance the third iteration tilts + 3 for girls, etc. Boys can never catch up! So, seriously, tell me where I’m confused about this.

                          Report

                    • This is a case where your intuition is wrong,

                      I’m not sure how, tho. The thought experiment is governed by the following two conditions:

                      1. There’s a 50/50 probability of boys/girls.
                      2. Parents stop having kids when a boy is born.

                      Does the restriction on having more kids after having a boy effect the total distribution? My contention is that it does in a small enough sample size. Consider the following probabilities of when a boy shows up in a single trial:

                      It1: B 50%
                      It2: B 25%
                      It3: B 12.5%
                      It4: B 6.25%

                      So, for any single trial sequence there’s a 50% chance B +1, 25% chance B = G, 12.5% chance G +1, 6.25% chance G +2, 3.125% chance of G +3, etc.

                      For any single trial the odds are that that trial will result in more G > B.* So, go to the tail end of distribution, even to the limiting case of a single trial. At that point, the above odds seem to apply. Why does that constitute an incorrect intuition?

                      * If I flipped an evenly weighted coin in sequence till a head lands up and then stop, would you bet that the number of times it landed heads up was greater than, equal to, or less than the number of tails?

                      Report

                      • 1. There’s a 50/50 probability of boys/girls

                        Exactly — that’s all you need to know to answer the question. It doesn’t matter which women decide to have more kids or not — every individual event is independent and has 50/50 odds, so the average distribution will be 50/50. The rest is just a distraction.

                        If I flipped an evenly weighted coin in sequence till a head lands up and then stop, would you bet that the number of times it landed heads up was greater than, equal to, or less than the number of tails?

                        Over multiple iterations, 50/50 – what else? You clearly don’t believe this, so you should open up, say, Excel and plot out all the different possibilities.

                        Report

                        • Guys. You’re both right.

                          Stillwater: “My contention is that it does in a small enough sample size.”

                          This is true.

                          Ken: “Over multiple iterations, 50/50 – what else?”

                          This is also true.

                          However, women do not have infinite numbers of children, and it takes infinite numbers for a 50/50 chance combined with “stop when you get a boy” to not result in a higher proportion of boys. (You need to get to four or five “tries” before you get higher than four-to-three.)

                          Report

                        • Over multiple iterations, 50/50 – what else?

                          There’s a 50% chance that head comes up on the first flip. That’s the only scenario where H > T.

                          There’s a 25% chance that a head comes up after a tail, where H = T.

                          There’s a 25% chance that a tail comes up after a tail, so 25% chance T > H.

                          Report

                          • Yeah, so what? The more T in the given result, the less probable it is, so on average, all the iterations in your last 25% bucket will net out to +50 Tails, balancing out the +50 Heads from your first group.

                            If you’re capable of using a spreadsheet program and adding formulas, plot it out — have columns for Probability, Score, and Total. Start Probability at .5 and set a formula for successive rows to be previous value/2. Start Score at -1 and set a formula for successive rows to be previous value + 1. Set Total to be Probability * Score for all rows. Then expand it out for as many rows as you want, and do a grand total (or running total if you prefer). Watch as the total approaches 0. Isn’t probability amazing??

                            Report

                            • Ken,

                              What the distribution of the little game I mentioned shows is that for a collection of trial sequences – not flips – the 50% will have more heads than tails, 25% will have H=T, and 25% will have more tails than heads. That’s just normal distribution, which applies to any single trial, given any coin flip has an equal probability of coming up H or T. That is, the probability assignment of any particular flip (or birth) is stipulated by the rules of the game.

                              So consider a single trial and ask “what’s the greatest differential in a single trial – governed by the stipulated rules – between heads and tails?” Clearly, in a single trial, the greatest differential of H over T (or B over G) is +1. But the differential of H over T (or G over B) is not so constrained. It can be 8 or 15 or 132. Those results are possible, perfectly consistently with each flip (or birth) being 50% for each time the coin is flipped. And at the tail end of a chain, those probabilities allow for tails/girls to radically outpace heads/boys. Not on average, but in a specific isolated case in which low probability events run up the score for girls in a way the it can’t be run up for boys.

                              Report

                              • My chart was showing you the probabilities and outcomes of sequences of flips, not individual flips.

                                If all you’re saying is that a single sequence could end up with a large discrepancy of T>H, then sure, no one’s arguing against that, but that doesn’t contradict the answer of the original puzzle — ultimately the total of T vs H balances out.

                                If you’re arguing that the fact that you can once in a while get a large T>H discrepancy for a single sequence means that the overall 50/50 average for H/T doesn’t apply, then that’s the part that’s incorrect.

                                Report

                                • Hmmm. Thanks for the reply. It’s the second option, but not that the 50/50 average doesn’t apply. It’s something else. I’ma think about this some more and figure out where my mistake is.

                                  Report

                                      • I had to explain that problem to my boss when her column came out. I said by switching, you’re actually picking two doors instead of just one. He didn’t get it, so I said you started out picking one of three doors, so there’s a 66% chance the car is behind the doors you did not pick. Monty says some mumbo jumbo and essentially asks if you’d like what’s behind both of the other doors, the one of them he opened being obviously wrong, but that’s irrelevant. What you’re betting is by switching is that your first guess was wrong, which is true 2/3rds of the time.

                                        Report

                                        • Don’t explain it with doors, explain it with playing cards. You’re looking for the Ace of Spades, all 52 are face down. You pick a face down card, the dealer flips over 50 of the 51. Do you want to switch?

                                          3 doors runs into the human instinct to keep what we’ve got and not risk a change. That’s a huge benefit for endurance hunters.

                                          Report

                                  • Fair enough. FWIW, this part of your 11:12 comment was exactly what I was describing for my proposed spreadsheet:

                                    So, for any single trial sequence there’s a 50% chance B +1, 25% chance B = G, 12.5% chance G +1, 6.25% chance G +2, 3.125% chance of G +3, etc.

                                    i.e. .5 * ( -1) + .25 * 0 + 12.5 * 1 + 6.25 * 2 + …. — although your next statement is that this means there are more G, it actually all adds up to 0.

                                    Report

                                • “If you’re arguing that the fact that you can once in a while get a large T>H discrepancy for a single sequence means that the overall 50/50 average for H/T doesn’t apply, then that’s the part that’s incorrect.”

                                  How long does it take to get to “overall”, though?

                                  Stillwater is correct that if you go with “each participant stops when the coin comes up heads” and truncate the number of coin flips at a finite value, your data will indeed have more heads than tails.

                                  Report

                                  • Stillwater is correct that if you go with “each participant stops when the coin comes up heads” and truncate the number of coin flips at a finite value, your data will indeed have more heads than tails.

                                    No. Incorrect.

                                    Group of 100. Stop at two trials. Assume average results.
                                    50/50 (50 stop there, 50 try again).
                                    25/25 (25 stop there) (and so do the other 25 because that’s 2).

                                    Total Results: 75/75

                                    Report

                • I was confused by blowing the probabilities back out to a large sample size. All along my intuition (let’s call it that) was based on applying the limiting principle of no kids after a boy to the last remaining live chains. Say single chain. So, yeah, I agreed with you at the time because I got confused about where I was in the argument. :) Basically, it’s that in that last chain, with the distribution of B:G conforming to expectation, the possibility exists for girls to outnumber boys by more than boys to outnumber girls. (And that’s true for any specific chain, of course. Seems to me that possibility comes into play at the tale end of the game.)

                  Report


        • I would say I have a narrow definition of empirical. In the american justice system crime usually occurs without the police present. So in a vast majority of cases the police have no empirically witnessed knowledge of the event. That is where the process of crime starts. We can start talking human error after that. We can start talking witness bias after that. We can start talking police bias after that. We can start talking judge bias after that. We can talk jury bias after that.

          Report

    • Ten years after the adoption of RTC laws, violent crime is estimated to be 13-15% percent higher than it would have been without the RTC law.

      I’d be very skeptical of how they determined a causal link (just as I’d be skeptical of a positive claim in the other direction). A claim that RTC failed to demonstrate a marked reduction in violence seems a perfectly valid claim to make; but the claim made in the summary, that there is a positive causal link between the passage of RTC laws and an uptick in violence… Unless they can show that the uptick is a direct result of people with carry permits or otherwise claiming to be exercising their rights per the RTC committing violent crime.

      Report

  9. Interesting study released showing a few things about American voters –

    https://www.voterstudygroup.org/reports/2016-elections/political-divisions-in-2016-and-beyond

    1. Basically, Clinton & Sanders votes aren’t that far apart on economics.
    2. Fiscal conservatives / Social Liberals are heavily overrepresented in both parties, due to the fact the richest members of both parties are more economically conservative.
    3. Basically, libertarians don’t exist. Hell, there aren’t even that many economically conservative people.

    Report

    • As I said on Twitter, a lot of Republicans spent an awful lot of time trying to define those in the upper left quadrant as beyond their consideration. And a lot of Democrats are doing that right now.

      Report

    • The most defining break between HRC and Sanders voters seems to be how “rigged” you see the system with Sanders voters being a lot more cynical on it and HRC voters being Leslie Knope/Vox essentially with a belief in good-faith and compromise. I am curious about what kind of politician could bridge both worlds. The last ones to do so were named Bill Clinton and Barack Obama. The Democrats might be in trouble if we always need to nominated a Bill Clinton or Barack Obama to win the Presidency. Previously elected Ds also managed to do this in their own ways like FDR whom I’ve heard older libertarians defend on the basis of “He gave us hope.”

      Lots of libertarians are being butt hurt right

      Report

      • I watched a recent speech by Victor Davis Hanson on Trumpism and where it came from, and he had an interesting insight.

        From JFK until Barrack Obama, the only way a Democrat got elected as President was to have a Southern accent. Obama may have been a unique anomaly, which means Democrats will have trouble finding someone who can win. Now if there’s something to it, and it’s not just a weird fluke, then it points to the way Southerners talk slower, simpler, and more directly. That might be important because Democrats favor complex big-government policies, and that can be a hard sell from someone that can’t build a lot of trust because they sound like a fast-talking, two-faced policy wonk from some northern machine city.

        Report

        • From JFK until Barrack Obama, the only way a Democrat got elected as President was to have a Southern accent. Obama may have been a unique anomaly…

          *(This message brought to you by Southern Democrats for a secure future.)

          More seriously, I’m deeply sceptical of tiny data sets, and even more so of tiny data sets with “unique anomalies”.

          Report

          • Well, the full data set, though small, would be:

            Johnson (Texas) beats Goldwater
            Humphrey (Minnesota) loses to Nixon
            McGovern (South Dakota) loses to Nixon
            Carter (Georgia) beats Ford
            Carter (Georgia) loses to Reagan.
            Mondale (Minnesota) loses to Reagan
            Dukakis (Mass) loses to George HW Bush
            Clinton (Ark) beats George HW Bush
            Clinton (Ark) beats Bob Dole
            Al Gore (Gaia) loses to George W Bush
            John Kerry (Mass) loses to George W Bush
            Obama (Illinois) beats John McCain
            Obama (Illinois) beats Mitt Romney
            Hillary (New York) loses to Donald J Trump

            Johnson, Carter, and Bill Clinton spoke simply, whereas Al Gore speaks like a policy wonk from the church of global warming, and the rest spoke like Washington insiders. It may be that the real mistake Democrats were making is often running candidates that the working class can’t connect with, but the press and chattering classes can.

            Report

            • It may be that the real mistake Democrats were making is often running candidates that the working class can’t connect with, but the press and chattering classes can.

              I think that’s a problem… but also that’s it’s one factor in noisy data. It’s not just one thing.

              2nd terms normally win if the economy is fine, 3rd terms normally do not.

              Some scandals are so bad the next guy gets blamed too. Nixon poisoned the well for Ford. Ditto Bush 2 for McCain. Clinton fatigue played a role in W’s win, although that was less decisive.

              Trump… as bombastic as he is, has a touch for connecting to people that HRC did not. No doubt he’ll screw things up so bad that Pence won’t have a chance.

              Report

          • In that VDH speech at the top of the sub thread (which was really enjoyable) he says at 11:35 minutes.

            He also understood a couple things. We’ve all been told demography is destiny, but is it really in the electoral college? Who cares who wins the popular vote if certain states are more important than others? A vote in Iowa is worth a lot more, or Ohio, or Wisconsin than my vote in California, at least at the national level.

            And Trump seemed to… I remember watching CNN when somebody said, I think they were talking about Jorge Ramos, “He seems to be oblivious to the ramifications of the Latino vote!” I said, “yeah, he is.” Because most of the Latino vote (and I’m living somewhere that’s 90% Mexican American) was going to go in two states. California that was never going to go red, and Texas that was never going to go blue. And the states that probably it turned, maybe Colorado and New Mexico, were going to be turned no matter who. Now if he had made the argument that there were half a million Latinos who hate Trump in Ohio, Michigan, and Wisconsin, I would have listened to him. But the point I’m making is that he looked at the electoral college and he understood that demography is destiny as it applies to particular states, and not the popular vote.

            At 14:00, VDH says:

            I talked to a New York developer once, and he called me and said, “You know, I watch from my tower.” He has a big high rise. I shouldn’t say high rise. I guess it’s a skyscraper. “I see Donald Trump, and it is true that he deviates from his planned walk when he gets out of his limo, and he talks to cement people, and I see people clapping that are on construction sites. So whatever it is, it’s genuine that people like him.”

            And I thought about that. When Hillary was telling coal miners they had to learn to build solar panels, he was using the first person plural pronoun “our”. Did you notice that? I’ve never heard a candidate in history say “our miners”, “our farmers”, “our soldiers.” Nobody told him to do that.

            The whole speech is chock full of insight.

            Report

        • Libertarians are Republicans who get no credit for agreeing on the handful of things where they overlap with Democrats because, duh, that’s the reasonable position to take and you shouldn’t get a cookie for agreeing that:

          Gay people are Human
          Women are Human
          Marijuana is Human

          And so on.

          Since they still think things like “teacher unions are bad”, it doesn’t matter if they’re good on gay marriage, or abortion, or the drug war.

          They’re still the butt hurt right.

          Report

Comments are closed.