Friday links: why higher education costs so much, statistics vs. knowledge, statistics vs. judges, and more

Also this week: racial and gender bias in postdoc hiring, the story of another “Hidden Figure”, the latest on Plan S, Jimmy Carter gets tenure, and more.

From Jeremy:

The US National Academy of Sciences voted overwhelmingly in favor of allowing expulsion of members who violate its code of conduct, which covers sexual harassment,  bullying, and scientific misconduct.

France has banned publication of statistical information about judge’s decisions, such as statistical models predicting future judicial decisions based on judges’ past decisions and personal attributes. Violation is punishable by up to five years in prison. Yes, really. (ht @kjhealy)

The latest on Plan S, the proposal by some European funding agencies to mandate open access publishing.

An argument that people have an irrational preference for “knowledge” (i.e. direct evidence such as eyewitness testimony) over statistical evidence, both in courts of law and in life in general. This would be an interesting piece to discuss in a stats course. Personally, I disagree with it. I think our preference for direct evidence over statistical evidence in courts of law is because in court we want to judge guilt or innocence in a particular case. Not “what fraction of defendants are guilty in cases that are like this particular case in relevant respects?” (ht @dsquareddigest)

The story of JoAnn Morgan, the only woman inside the control room for the Apollo 11 launch.

The backstory of a rogue geoengineering attempt. I’ve used this case as the basis for an exam question in aquatic ecology. Under what conditions, if any, would you expect a one-off pulse of iron enrichment in the ocean to give you both carbon sequestration and more+bigger salmon?

Why is the price of higher education in the US so damn high? Here’s a deep answer–the Baumol Effect. I’m not an economist so I’m not qualified to fully evaluate the argument. But the logic is exceptionally well-explained and it’s at least consistent with some important features of the data. So if it’s wrong, I don’t think it’s obviously wrong. Note however that the standard illustrative example of the Baumol Effect–a string quartet–is no longer the best example.

Jimmy Carter finally gets tenure. (ht @dandrezner)

I’m Pratchett Nowhere, and this is All Things Considered. 🙂

From Meghan

A new study looked at racial and gender bias in postdoc hiring (and, importantly, did so in a way that should have removed the effect of social desirability bias, by saying it was a study on CV formatting). They sent CVs that were identical except for the name (e.g., José Rodriguez vs. Bradley Miller). In biology, faculty did not show gender bias in competence and likeliness to hire, but Asian and White applicants were viewed as more competent and hireable than Black and Latino applicants.

I wonder if we’ll be able to open any windows at Evolution or ESA this summer:

10 thoughts on “Friday links: why higher education costs so much, statistics vs. knowledge, statistics vs. judges, and more

  1. I don’t understand Alex’s argument for the Baumol Effect.

    What seems strange about his argument is that there should be a decrease in cost in almost everything simply because of the economies of scale of production, even before adding new technology. If that hasn’t occurred, then it seems like there needs to be an explanation for why not, and this is true of education just as much as it is of automobiles.

    Matt Levine’s argument is great. Kinda hints that there might be a lot more problems with it than is being let on….

    • “What seems strange about his argument is that there should be a decrease in cost in almost everything simply because of the economies of scale of production, even before adding new technology.”

      No, because what matters are relative increases in productivity in different industries.

      • “what matters are relative increases in productivity in different industries.”

        ??? I’m surprised Alex made such a simple mistake.

        The comparison of education or healthcare to the auto or computer industries isn’t valid because there’s no mechanism – competition – to drive productivity in education or healthcare.

        The assumption that a car is the same today as it was in 1950 is wrong. In 1950, few automatics, rare power steering, DC electrical, absolutely no effort to economize fuel or materials, no stereos, no CD players, no defrost, crappy heating systems…In fact you’ll recall that the entire US auto industry almost disappeared – *twice* – because it refused to acknowledge consumer preferences. The fact is that cars have changed dramatically because the companies that make them are subject to competition.

        OTOH, most of the curriculum I studied as a geology major has almost no relevance in today’s job market. There’s no reason for 300 schools to be teaching igneous petrology. Yet teach it they do, because they’re not subject to competition.

        If the universities were subject to market forces, there would be maybe 10 or 20 geology programs in the country, not 300. The ten or twenty would be top notch, because they would be able to generate synergies on teaching and equipment that would keep other schools out of the game and keep costs down. this concentration of resources would also keep costs down for research.

      • Sorry Jim, you’re misunderstanding the linked post. No, he’s not assuming that a car today is the same as a car in1950.

        Not that Alex Tabarrok is infallible, obviously–no one is. But since I’m not a PhD-holding economist myself, I’m going to rely on my own understanding of the linked post, plus Tabarrok’s credentials. We’ll have to agree to disagree. Thank you for commenting. I’m going to step away from this conversation.

      • The rest of the schools would be content to teach intro and historical geology with perhaps a mineralogy, paleontology, hydrology or even the occasional ore deposits course. But outside the top schools, there would be few or no grad programs, no grad level courses, and no upper level undergrad courses.

  2. Fair enough J, just a couple thoughts:

    might also check out this post on Baumol:*mn
    some info on the quantitative relative productivity gains of various industries.

    I accept that a wage premium in a highly productive industry might be partly forced on a less productive industry and thus raise prices. Teachers often claim they’re worth as much as programmers. 🙂

    Yet it’s also a fact that there’s no mechanism to eradicate inefficient education institutions like there is for auto manufacturers. Since there’s no selection mechanism it would be *very* surprising if educational institutions had achieved any gains in productivity at all, much less achieved some maximized level of efficiency.

  3. re: today’s post by Alex on the Baumol effect

    His comparison of car/car repair and shoe/shoe repair is valid but irrelevant because it misses the competitive effect: aggregate spending and wages of care and shoe repair have likely fallen dramatically as a proportion of the overall economy. So, yes, the cost has risen but the competitive impact of productivity is still there. Alex just choses a data viz that hides it.

    No one is complaining about car and shoe repair because the overall budget impact is small because of the declining demand.

    Everyone complains about education because the overall budget impact is large because prices are increasing even as demand increases. There’s no productivity impact because there’s no mechanism for productivity.

  4. Hi J,
    Sorry, I didn’t explain that very well.

    And don’t get me wrong, I love MR and I think Alex and Tyler are smart guys.

    My understanding is that Alex is trying to explain why costs rise in industries with little or no productivity growth. Alex attributes the rise in costs to rising wages, which rise in unproductive industries rise because of the wage growth in productive industries.

    I accept that explanation. But it’s trivial and obviously true. Regardless of productivity, no one will spend $X0,000 to get an education degree if the wages aren’t roughly comparable to a computer science degree, which costs the same.

    In Alex’ original post “Bloat Does Not Explain the Rising Cost of Education,” he says:

    “The bloat theory is superficially plausible….But the bloat theory requires longer and lazier rivers every year, which is less plausible.”

    The second claim is unsupported and likely incorrect, on two counts.

    First, the *exact opposite* is actually the case: because there is no selection mechanism in education we should *expect* longer and lazier rivers every year.

    Second, we don’t actually need “longer and lazier” rivers – that is, we don’t need new waste to be added – to drive a continuous rise in price over time. All we need is that existing waste never gets eliminated. Because even if the product is static what constitutes “existing waste” grows as technology and knowledge change.

    Alex’ assessment of what constitutes “bloat” is a strawman. Of course “bloat” includes too much administration and too many perks for college students. But the real “bloat”, again, isn’t “new” waste. It’s in the waste that never gets eliminated. It’s in *relative productivity loss*. In the private economy, companies merge to consolidate operations and reduce costs; change products to meet consumer demand; and eliminate products when demand drops. None of this happens in education. It doesn’t have to. There is no selection.

    So, yes, OK, the Baumol effect is real. But it’s irrelevant. It doesn’t eliminate bloat as a source of increasing cost because bloat is mostly *relative productivity loss*.

  5. can I beat this horse any more? 🙂

    One of the big reasons education costs keep going up is because education is missing out on massive potential efficiencies. It’s not clear to me at all why every one of thousands of professors and instructors write and prepare their own introductory courses. This is a massive waste of effort and a ridiculous cost for virtually no benefit.

    The vastly more sensible approach is to create a standardized courses each with a book, set of assignments, labs and tests, then eliminate the lecture and run the course exclusively through the book and labs, adding extra lab time if necessary. For non-lab courses, students need only buy the book and pass the test.

    If there were a competitive system for education this would have been the case a long time ago. It may yet emerge with modern technology, but it just as well could have happened without the internet. Modern technology does offer additional efficiencies but from what I’ve seen they’re not being exploited very well. The last thing any university wants to do is come in under budget.

Leave a Comment

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.