Statistics can be extremely hard!


Counting those (top-secret) documents: Joe Biden has cancelled some college debt. In this fairly recent report, two Forbes reporters described a related phenomenon:

MCGURRAN AND HAHN (3/28/22): In 1980, the price to attend a four-year college full-time was $10,231 annually—including tuition, fees, room and board, and adjusted for inflation—according to the National Center for Education Statistics. By 2019-20, the total price increased to $28,775. That’s a 180% increase.

In what follows, we'll assume the accuracy of those basic statistics concerning annual cost. The key takeaway is this:

Even after adjusting for inflation, the cost of attending college soared over those forty years. This is the kind of topic you won't see discussed on MSNBC, where the only known topic is this:

Trump Trump Trump Trump Trump Trump Trump Trump Trump Jail Jail Jail Jail Jail!

For the record, discussing statistics is often extremely hard, even for our top journalists. In the example from Forbes, consider this possible problem:

Even after adjusting for inflation, the cost of college almost tripled over those forty years! But the Forbes reporters chose to describe that as "a 180% increase."

Their statement is accurate—but the cost had almost tripled! Had we been their editor, we would have replaced their (accurate) account—their "180% increase"—with the other (accurate) account of that state of affairs.

On MSNBC, you'll never see anyone reporting the way those costs have risen, or asking why they've risen so much. The channel's hosts have only one topic, and it's entertaining: 

Trump Trump Trump Trump Prison!

The current news about Trump is important, of course—but it isn't the nation's only significant topic. That said, you're routinely denied information about the topic of college costs as you get fed that one pleasing topic:

Trump Trump Trump Trump Jail.

The news about Trump is important. This morning, it occurred to us that some of our leading journalists are having a hard time with a basic statistical question. That question goes like this:

How many top-secret documents did Trump take to Mar-a-Lago?

That's a very basic question. It seems to us that a misleading account is being widely offered in the wake of last Friday's release of the DOJ affidavit.

It seems to us that this misleading account tends to understate the actual number. Statistics are often extremely hard, especially for those at the top.

Tomorrow: How many such items in all?


  1. Current tuition = 2.8 x old tuition.

    1. Yes, I think Somerby would prefer this expression too, because 2.8X sounds like less than a 180% increase. It has the advantage of being more accurate than Somerby's 3X, but it would be harder to understand than simply saying three times bigger. Somerby's quibble is about what is easier for the reader to understand, but he ignores the literature on how innumerate people actually read numbers and think about quantities.

      Here is what Somerby needs to read first: Michael Lewis's book about Kahneman & Tversky's work, or better yet, read their work itself. Followed by books on judgment by Daniel Ariely.

  2. "Even after adjusting for inflation, the cost of college almost tripled over those forty years! But the Forbes reporters chose to describe that as "a 180% increase."

    Somerby says he prefers to call this "almost tripled" instead of the more precise actual figures. He states explicitly that both descriptions are accurate. But are they?

    The "almost tripled" description rounds up and glosses the more exact, more precise average amount that is 180% higher (not 200% higher). Is a rounded off figure equally accurate to a more precise figure? I don't think so.

    Does the difference matter? It depends on whether a reader will construe almost triple to be larger or smaller than 180%. I suspect that they will consider the 180% to be larger because triple is 3X and 3 is a smaller number than 180. Psychologically speaking, and especially for the innumerate or those not reading carefully, the way Forbes wrote it is likely to seem like a bigger increase than Somerby's preferred way. But that is an empirical question that can be tested by questioning actual readers and asking them which is larger.

    Propagandists know the best ways to present figures in order to achieve a desired shifting of opinion or consumption of misinformation. Do Republicans want the public to understand how much college costs have increased? It would create more sympathy for those having their loans forgiven, if the public understood such things. Somerby's preferred editorial change would be in the direction of minimizing the increase in costs (3X just doesn't sound as big as 180%, even if it is bigger, just as people will choose a 1/4 slice of pie over 1/3 because it too sounds bigger, even if it is not). But why is Somerby even offering this nitpick and why on earth is he calling it equally accurate when it manifestly is not?

    Is it perhaps because Somerby doesn't mind it when journalists put their thumbs on the scales in ways that support his own opinions? And why is Somerby against student loans? He was a teacher and should, of all people, know why this is a good thing for our society, but lately he has been parroting conservative talking points, and the conservatives are against helping students, because it will make Biden look good and help Democrats win. So who cares what helps kids if you're a Republican? It's what helps the party that counts.

    Statistics is one of the easiest branches of mathematics -- it isn't hard at all, if you studied high school math. Today, Somerby shows us how the sausage gets made on the right when conservatives talk to those who didn't study statistics in college.


  3. Why, yes, dear Bob, this is exactly how consumer subsidies work. The state provides cheap student loans, financial aid, etc. -- and the prices rise accordingly.

    Same thing, incidentally, with electric vehicles: the state institutes consumer subsidies -- and car companies raise prices. Who benefits? Why, the car companies, obviously. Who loses? Why, obviously the taxpayers.

    And so it goes...

    1. Actual studies of rises in tuition show that there has been a huge increase in salaries of college administrators, under the assumption that they should be paid comparably to those working in private industry. Because there has been even greater inflation in pay of CEOs and other business administrators, the rise in academia parallels what has occurred in corporations. At the same time, public colleges and universities which got their funding from state budgets have shifted those costs onto students and abruptly reduced the amount of college budgets that come from states. Public colleges have not been as adept at seeking donations and gifts from wealthy alums so tuition hikes have been disproportionately high among those institutions that previously served working class applicants and those who are the first in their families to attend college. Without loans, many such students wouldn't have gone to college, given the rising costs. To see why this happened, you have to go back and look at what was squeezing state budgets.

    2. Why this happened can be traced to Ronald fucking Reagan, like most things wrong with this country.

  4. "On MSNBC, you'll never see anyone reporting the way those costs have risen, or asking why they've risen so much. The channel's hosts have only one topic, and it's entertaining:

    Trump Trump Trump Trump Prison!

    The current news about Trump is important, of course—but it isn't the nation's only significant topic. "

    Meanwhile, I put in the search terms "msnbc" and "student loan forgiveness" and found these two from the past few days:

    Plus two more talking about MSNBC's coverage:

    (This one argues that cable news has been accepting the debate framing by conservatives when discussing the impact of loan forgiveness, instead of putting their own spin on the topic.)

    Oddly, I found no mention of, much less analysis of, the increases in college tuition at Fox cable news when I used Fox as a search term.

    Somerby is incredibly nit picky when examining stats but shows no caution about accuracy when he issues his blanket statements about what happens on MSNBC.

    1. You didn’t google very thoroughly. Surprise!

      Scads more via YouTube.

    2. The task set by Somerby was to find a mention of tuition increases not just criticism of Biden’s action.

    3. Anonymouse 7:45pm, I don’t think that MSNBC opining about the discussion being fraught with right wing talking points would qualify then.

  5. Two quick ones off the top: 1) MSNBC has covered the College Loan bail out a lot, and represented other views than just the White House. They certainly would be pro on the issue, but Bob's contention that they would never show a counter view is false. 2) MSNBC has never chanted for Trump to go to jail, unlike the crowd at a Republican convention chanting "lock her up." They have strongly contented Trump has broken the law again, something
    that Bob has no interest in.

  6. He goes downhill from there. Bob needs to badly misrepresent MSNBC and what they are putting them out
    to dismiss them. Bob is a dishonest guy.
    Then he admits this is a meaningful story, and
    defects attaching any importance to it by quibbling
    over understandably unclear specifics.

  7. You cannot call a statement accurate and then discuss it as if it had been a mistake. That is dishonest. Most people who read Forbes will be OK with the way the reporter said it.

  8. Are we talking about private schools here? You'd be hard-pressed to find a reasonably good school where you can get away with $28,775 annual expenses. Maybe a state school.

    1. It's an average. Probably across all types of schools.

  9. In the good old days, tuition was 36% of what it is today.