Showing posts with label statistics. Show all posts
Showing posts with label statistics. Show all posts

Sunday, March 29, 2026

Mark Twain Is Right. Again.

Mark Twain has already been lauded by Curmie for his observation that “In the first place God made idiots. This was for practice. Then he made school boards.”  Of course, he’s not the only luminary who, despite having been dead for decades (or even millennia), has been cited for offering more incisive commentary on America in the 2020s than the vast majority of contemporary pundits have been able to generate.  Other folks in that category include Euripides, Luigi Pirandello, George Orwell, Martin Niemöller, and Gertrude Stein.  There are others, no doubt, who will be added to the list in the weeks and months ahead.

But ol’ Mark (or “Sam,” if you prefer, Gentle Reader) gets a second nod today for his observation that “there are three kinds of lies: regular lies, damned lies, and statistics.”  Curmie could have written about that line years ago in reference to, say, border security.  If administration X turns away more unauthorized migrants than administration Y does, does that mean there were more illegal crossings during the X administration or that the Y administration didn’t do as good a job of identifying wrong-doers?  Probably a little of both, one supposes.

Some statistics relating to crimes actually matter.  If there’s a dead body with bullet holes in it, there’s a really good chance a murder has been committed, whether someone is ultimately arrested, tried, and convicted or not.  But if the crime in question is, say, sneaking across the border, we generally don’t know of the existence of the offense until someone is caught. 

And this brings us to the Safeguard American Voter Eligibility (SAVE) Act.  Cutesy title notwithstanding, the bill is an abomination, but we’ll get nowhere arguing statistics.  There’s a meme out there that says the Cleveland Browns have started more quarterbacks since 1999 than non-citizens have voted in that time period.  Curmie actually fell for that one, but it’s a lie… not just an untruth, a lie.  More to the point, that claim wasn’t justified even by the statistics supposedly cited: its evidence was that there were fewer convictions than Browns quarterbacks.  That’s different than the number of illegal votes cast by a factor of…what? 2? 10? 100? 1000?  Who knows?

One side claims that prosecutions and convictions are extremely rare; the other side points to tens of thousands of names of non-citizens on voter lists.  They’re both right.  And they’re both exaggerating.  The following statements are all true, as far as Curmie can determine: 1). The majority of cases of illegal voting involve convicted felons, false impersonations, or registering in multiple constituencies, not non-citizens.  2). Some locales allow non-citizens to vote in local elections (school board, mayor, etc.).  3). Motor/voter programs may give non-citizens the impression that they can legally vote, especially if, say, they have a green card, have petitioned for asylum, etc.  4). A significant number of people who were about to have their registrations revoked were in fact citizens, and a disproportionate percentage of them were either black or Hispanic. 

Curmie made similar points 14 years ago when Texas was about to purge thousands of voters from the rolls.  It’s actually a pretty good essay.  You should read it in its entirety, Gentle Reader, but here are a couple of selections:

This all boils down to a simple illustration: if someone named Carlos Martinez, a registered Democrat, shows up to vote, too many Republicans want to say “no,” and too many Democrats want to say “yes.” The correct answer, of course, is “yes, if…”: if you’re a citizen, if you’re registered in this district and nowhere else, if you have some sort of reasonable proof that you are who you say you are….

We need to establish some system of presumption. You can’t register to vote unless you can prove citizenship; once registered, however, the government must prove you should be removed. None of this reliance on motor vehicle registrations or jury exemption lists: they’re notoriously unreliable. And the presumption rests always with the status quo. Once you’re registered, the burden of proof shifts to the government to demonstrate to a high standard of proof that you shouldn’t be….

As regards increased demand for appropriate identification, perhaps requiring a photo ID: yes, by all means, if and only if there is a full-scale, well-funded campaign to make sure that prospective voters know not only that the laws have changed, but how they’ve changed, and how to secure, without undue hassle, a legally sufficient, free, identification card.

The same argument about presumption re-appeared a few years later when Curmie was discussing a different xenophobic exercise in a different Trump administration: literally incarcerating and rescinding passports from honest-to-God born-in-the-USA citizens who happen to be… you know… of Hispanic heritage.  Curmie concluded that piece with this: “Yes, there are leftie commentators who are comparing this issue to the internment of Japanese-Americans during World War II or even the worst excesses of Nazi Germany. Is that really where we are? Of course not. But have we taken far too many steps down a road towards a very nasty, xenophobic, and unjust nation? Oh, yes.”

So there’s no guarantee, on the one hand, that even having a US passport will in fact mean anything to the Mad King of Trumpistan.  But even if that particular strategy doesn’t re-appear, there’s plenty to hate about the SAVE Act.  No, we shouldn’t yammer on about how few violations there actually are (we can’t prove those assertions beyond the general observation that there aren’t many), but we sure as hell need to keep pounding the fact that this bill is all about voter suppression. 

Curmie remembers citing a quotation generally attributed to Benjamin Franklin in a blog post from four years ago: “That it is better 100 guilty Persons should escape than that one innocent Person should suffer, is a Maxim that has been long and generally approved.”  Unfortunately, however, Curmie also noted that a 2016 study by the Cato Institute found that a “terrifying 52% of Donald Trump’s ‘early core supporters’ responded to the question of which is worse, ‘having 20,000 people in prison who are actually innocent; or, having 20,000 people not in prison who are actually guilty’ by selecting the latter.”  There’s something fundamentally psychotic about that mindset, but… Trump supporters.  (Curmie antiphrastically refrains from suggesting to them that anyone whose name appears in the Epstein files should therefore be imprisoned immediately.)

The idea that it is somehow worth disenfranchising legitimate voters if by doing so we catch a few bad guys is, of course, the best case argument for supporting the SAVE Act.  The other rationales are even worse: that people who have changed their names for whatever reason—because they are trans or women who took their husband’s surname when they got married, for instance—ought to have to jump through more hoops than the rest of us just to be able to vote.  Oh, sure, a passport would work (assuming it hadn’t been arbitrarily revoked for spurious reasons) but those things cost a fair amount of money and take weeks or even months to get (especially since the Trump administration has initiated staff cuts in the agency charged with processing applications).  

Place the presumption with the status quo (if you’re currently registered to vote, you stay that way until and unless the government can prove that you should be disenfranchised) and make the process for new applicants fast, simple, and free, and we can talk.  Otherwise, it’s a poll tax intended to suppress the votes of women and poor people: those who would be more likely to vote against Republicans, in other words.

Of course, liberals are more likely to have passports than conservatives are, and liberal women are less likely to change their names when they get married, so it’s possible that this little stratagem would blow up in the smug faces of the Trumpistanian minions.  They’d certainly deserve it.  A better alternative, however, would be to have a couple of Republican Congresscritters care more about their country than about Dear Leader’s latest power grab.  As of this writing, there may just be a large enough handful of them to prevent this absurdity from becoming law.  We can but hope.  (Oh, and write, and call, and…)

Friday, May 10, 2024

The Axios Survey, For What It's Worth


Once upon a time, Curmie was something of a math whiz-kid; he’s really good at easy math.  We don’t talk about analytic geometry or natural logarithms in polite company, but various applications of algebra, from grading spread-sheets to strategies for investing the retirement nest egg, have been very much a part of Curmie’s world for a long time. 

Even keeping in mind Mark Twain’s observation that there are three kinds of lies—regular lies, damned lies, and statistics—Curmie has been fascinated by statistics since he was a little boy.  Of course, knowing the numbers is only part of the story; it’s the conclusions we draw from that raw material that matter.

First, we need to consider the possibility that the facts are not only misinterpreted, but actually wrong.  Ask anyone who paid even a little attention in high school English who the groundlings were at the premiere performance of Hamlet, and you’ll get a variation on “the butcher, the baker, and the candlestick maker.”  The evidence is two-fold: it cost only a penny to attend, and there are dirty jokes.

But the fact is that we don’t know how much it cost to attend.  We have a grand total of two contemporary sources that state what the standard admission fee was at the Globe.  It’s a reasonable surmise, but still not a certainty, that Hamlet cost the same as other plays.  But whereas we assume that a penny was an insignificant amount of money, it turns out that as a percentage of the disposable income of a journeyman laborer, the price of admission to the Globe was almost identical to a cost of a ticket to the National Theatre today.

Curmie also has it on good authority that, notwithstanding Victorian era claims to the contrary, well-educated and even wealthy people think dirty jokes are funny at least some of the time.  That whole “my head in your lap… country matters” business?  [Note: in Elizabethan England, the “t” in “country” would have been part of the first syllable.]  Oh, Will, you naughty rascal!

One more thing: the Globe was an outdoor theatre, dependent on daylight for spectators’ ability to see the stage action.  The butcher, baker, and candlestick maker were butchering, baking, and candlestick making at that hour.  Hamlet’s audience would likely have looked like that of a Broadway matinee today: people who don’t need to work or are visiting the city, or those who worked different hours (the Inns of Court often did business at night, for example).

Curmie was thinking about all these things when he read the Generation Lab poll commissioned by Axios about the concerns of today’s undergraduates.  Titled “Exclusive poll: Most college students shrug at nationwide protests,” the article suggests, among other things, that perhaps the protests on American campuses won’t hurt President Biden’s re-election chances as much as might have previously been believed.  Students don’t blame him for the situation in the Middle East as much as they do Hamas, you see. 

Curmie apologizes for bluntness, but the idea that literally anyone would suggest that Biden was more responsible than the actual combatants has crossed the line into full-blown lunacy.  If you want to say he hasn’t handled the situation as well as he might have, you’ll get no argument from Curmie.  But to say that he’s more culpable than Hamas (or Likud, for that matter) is begging for a nice padded cell.

Curmie does stop short of declaring the entire poll and its interpretation redolent of bovine fecal matter, but that one eyebrow did indeed shoot upward and the opposite eye did close to a squint.  Believing that it definitely cost exactly a penny to attend Hamlet or indeed that the groundlings were who your high school English teacher said they were requires considerably less of a leap of faith than trusting these poll results.

For one thing, the questions are incompetent.  Notice that neither antisemitism nor freedom of speech and assembly is an option.  Of course “the conflict in the Middle East” didn’t generate as many responses as issues that affect students directly: healthcare, education funding, etc.  

The Vietnam War, the source of student protests in Curmie’s youth, was a different matter: college-aged young American men were being drafted and used as cannon fodder for a war many of them thought was unjust.  It’s hardly surprising that there were a lot of protests on university campuses.  What’s happening now in Gaza has little direct relevance except to those with relatives in either Israel or Gaza.

More to the point, given that there are an estimated 15.8 million undergraduates in the country, that “small minority (8%)” of students who have taken part in a demonstration on one side or the other represents over one and a quarter million students.  That’s a lot of folks in Curmie’s neighborhood.  There are multiple unknowns here: the percentage of demonstrators on each side of the dispute, for example.  It’s evident that a significant majority support the Palestinian (not to say Hamas) position: we’ve certainly heard more about them, and one doesn’t often stage a demonstration in favor of the status quo.  “Don’t divest” comes up rather short as a rallying cry.

We also don’t know why those students were involved in protests.  For some, no doubt, the demonstrations were more social than political.  For others, they were an excuse to skip class.  And to suggest that there was a fair amount of naïveté in evidence is only to state the obvious.

But even granting that the Venn diagram of those who see the conflict as a major concern and those who have actively taken part in protests isn’t exactly concentric circles, there is a good deal of overlap.  The statistic that jumped off the charts to Curmie but apparently not to Axios’s Sareen Habeshian was this: the number of students saying they’d demonstrated about the situation in Gaza and Israel was over 60% as high as the number who consider the Middle East a “most important” topic. 

That’s an extraordinarily high number.  Even if half of the demonstrators weren’t true believers, that leaves a third or so of those who regard the conflict as of high importance willing to demonstrate on behalf of their beliefs.  When was the last time, Gentle Reader, that you saw a demonstration about those issues more students allegedly care more about: better healthcare, education funding, or gun control? 

Even demonstrations about abortion access (speaking of an absurd omission in a poll about what matters to college students!) have faded as the Dodds decision moves further into the past.  (Perhaps abortion issues may have been folded into the umbrella of “healthcare,” explaining why it was the number one response to the survey?  If so, why?)

One of Curmie’s professor friends suggests that today’s college students, a group we might call the post-COVID generation, are angry at the world at large, and the conflict in Gaza and Israel is as good a flashpoint as anything else.  Perhaps.  But Curmie, old fart that he is, revisits Buffalo Springfield’s biggest hit, linked above, which is chillingly relevant 57 years after it was recorded.  It’s difficult to decide what the most telling line is.  “There’s battle lines being drawn / Nobody’s right if everybody’s wrong”?  “Young people speaking their minds / Getting so much resistance from behind”?  Or maybe the opening line, “There’s something happening here / What it is ain’t exactly clear”?

The one thing that is clear is that this survey tells us more about Generation Lab and Axios than it does about what today’s undergrads are thinking.


Monday, July 19, 2010

Global Warming, the Charlie James Syndrome, and a Very, Very, Very Big Number

Ever since I was a little kid, I have been aware of what I refer to as the Charlie James Syndrome. My father, although he grew up in New England, was a fan of the St. Louis Cardinals. He started following the “Gashouse Gang” of the 1930s, and remained a loyal Cardinals supporter until the day he died.

I, on the other hand, followed the utterly hapless New York Mets. So it was that one day late in the 1963 season, my Dad and I were listening to a Mets-Cardinals game on the radio (we didn’t have a TV). Despite my loyalties to the Mets, I was aware that the Cardinals were in a pennant race (they ended up in second place, 6 games behind the Dodgers), whereas the Mets finished 15 games out of 9th place in a 10-team league. So my loyalties were divided.

Anyway, on that particular day a journeyman outfielder for the Cardinals named Charlie James hit two home runs, giving St. Louis its margin of victory. In that same game, Stan “The Man” Musial, a future Hall-of-Famer, went something like 0-for-5 with three strikeouts. “Clearly,” thought I, roughly a week or so from my 8th birthday, “Charlie James is a vastly superior baseball player to this Musial guy my Dad keeps talking about.” Needless to say, he wasn’t. James’s career OPS (On-Base Percentage Plus Slugging Average) was .652. Musial’s lifetime OPS was .976; even in 1963, when he was 42 years old and James had the best year of his career, Musial was still the better hitter by a .729-.697 margin.

But on that one particular September day in 1963, Charlie James sure had a better game than Stan Musial did. For some reason, that game stuck in my mind, and it has become an important memory—not merely because a little boy got to share a baseball game broadcast with his Dad, but for the lesson of Charlie James. It wasn’t too long before I realized that anomalies happen, but that doesn’t make picking Stan Musial instead of Charlie James a bad idea.

So with all the hoo-ha about climate change, I have remained ever-so-slightly skeptical. True, I trust scientists more than I trust politicians, and I have found the protestations of the right more than a little foam-flecked. But the fact that the arguments being adduced against the whole idea of climate change aren’t convincing doesn’t mean that those in favor of the theory are much better.

And, of course, there is ample anecdotal evidence that the doomsayers are exaggerating. All of us have a friend in northern climes who jokes about having to shovel 14” of global warming in the winter. I just returned from Ireland, where the reason we couldn’t go inside Thoor Ballylee, W.B. Yeats’s tower, was that the coldest winter in nearly a half a century had led to flood damage that hadn’t been cleaned up yet.

Both sides of the debate are guilty of selective interpretation of evidence. Some readers will recall the O.J. Simpson trial, in which prosecutors made a big show of making the defendant try on a pair of gloves, ready to strut their stuff when one more nail was added to Simpson’s coffin. Of course, the gloves were too small, leaving the prosecutors scrambling to babble—accurately but with palpable desperation—that, reasonably, the gloves would have shrunk because of the exposure to blood. But all Johnny Cochran needed was one little chink in the armor; when, with rhetorical flourish, he proclaimed, “If the gloves don’t fit, you must acquit,” the trial was effectively over.

I am always reminded of that when I see someone claiming that, for example, an increase in snowfall is actually attributable to global warming. I follow the scientific explanation, and it makes sense, but there’s still that nagging feeling that I’m buying into a scam. So what would be a statistic that would prove unequivocally that global warming is a real and present danger?

How about this, from the official National Oceanic and Atmospheric Administration (NOAA) report on the Global State of the Climate:
June was the fourth consecutive month that was the warmest on record for the combined global land and surface temperatures (March, April, and May were also the warmest). This was the 304th consecutive month with a combined global land and surface temperature above the 20th century average. The last month with below average temperatures was February 1985.
I first saw this passage, by the way, on the Guardian’s website. [N.B. I find it interesting that the mainstream US media—even such “liberal” bastions as MSNBC and the Washington Post—ignore the most significant statistic of the story; I can’t find the story at all on the New York Times website.] OK, let’s look at that 304 consecutive month number. Logically, there’s a 50-50 chance that each month will be warmer than or cooler than the 20th century average. So, the odds of having 304 straight months above that average based purely on chance are 1 in 2 to the 304th power, or roughly 1 in 3.25 times 10 to the 91st power. That is, 1 in 32,500,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000. Those strike me as rather long odds. You know the expression “one in a million”? The odds of 27 consecutive months being above the 20th century average are less than one in a million. That gives you a rough idea of what 304 would be.

Or look at it this way: Charlie James hit a homer about every 48.5 at bats. 304 consecutive months isn’t James homering twice in one game: it’s his hitting a home run in 53 consecutive at bats. (The major league record is 4.)

Now, global warming deniers… you were saying?