Month: February 2023

A couple of years ago I became pretty skeptical of the panic over social media. No matter how obvious the harms of social media might seem, especially to us oldsters, the evidence just wasn't there. This became especially obvious to me in late 2021 when the media wildly misreported the leaked Instagram survey of teen boys and girls. In fact, in my year-end list of bad trends, I included this one:

Blaming everything on social media. This really ought to stop. There's very little rigorous evidence to back it up, and quite a bit to suggest that social media is a net positive.

Jonathan Haidt has been on the opposite side of this argument for a while, and yesterday he criticized the kinds of objections people like me raise about the evidence of social media harms—primarily that the actual research is thin and inconclusive:

Those were reasonable things to write in 2019, but not in 2023. We now have dozens of experiments, plus some very consistent and incriminating patterns in the hundreds of correlational studies.

Maybe so! This is not a subject I follow obsessively, so maybe things have changed. Haidt referred me to his congressional testimony from last year for more, so I read it.

For a while I was nodding along as Haidt produced a bunch of charts showing a steady rise in anxiety and depression among teens starting around 2012. But then there was this:

“The associations between social media use and well-being therefore range from about r = − 0.15 to r = − 0.10.” I agree with this assessment, for both sexes combined....A ballpark figure for the correlation just for girls is roughly r = .15 to r = .22. The effect size is even larger for girls going through puberty....For them, the size of the correlation with poor mental health could be well above r = .20.

This requires a translation from math to English. Here it is:

• Overall, social media can explain about 1-2% of the difference in well-being among teens.
• Among girls, it explains 2-4%.
• For girls going through puberty it "could" explain more than 4%.

As Haidt points out, 2% is low, but it's not nothing. And as he also points out, his concern is solely about the effect of social media on mood disorders like anxiety and depression. However, even limited to mood disorders it's hard to square this with the infamous Instagram survey that got so much attention in 2021:

Teen girls report that Instagram had a net +39% effect on loneliness; +29% on anxiety; and +49% on sadness. On all three measures, only about 10% of teen girls reported that Instagram made things worse.

So . . . I dunno. I'd offer up a few tentative conclusions:

• Even Haidt limits his criticism to depression and anxiety, mainly among girls, but in the popular press this too often gets translated into a generalized panic about social media having a bad effect on nearly everything. Kids these days don't know how to use a salad fork! Blame Facebook! This is just wrong.
• Even if Haidt is correct, the size of the effect on teen depression is quite small.
• It also appears that social media has a negative effect only in large quantities. A couple of hours a day is unlikely to change anything.
• Most likely, the biggest effect is on teen girls who already suffer from depression and then begin to obsess over their social media accounts, which makes them even worse off. If this is the case, it's counterproductive to panic over social media in general. The thing to watch for is depressed teens who suddenly start spending five or six hours a day on social media. That may be the sign of a dangerous downward spiral.

You can see from this why I called yesterday's CDC report kind of mysterious. It's one thing to look at trends in teen depression since 2012 and conclude that social media has played a role. It's quite another to explain a huge jump starting around 2019. The effect of social media is simply too small to account for it, and the timeframe makes no sense anyway. So it remains a mystery.

POSTSCRIPT: All this said, my gut agrees more with Haidt than my brain. For what little it's worth, here's a few pieces of advice for parents:

• It's not really feasible to keep teens away from social media, but it's probably a good idea to keep them away at least until age 14.
• Absolutely have no guilt about insisting that you have access to their accounts and will check in on them periodically.
• That said, don't check in all that often.
• Try to persuade your kids not to follow too many people. This is what causes phones to demand attention constantly.
• Take their phones away at night. That goes for you too.

This is Charlie examining some dead leaves hanging from the underside of a little teak table in our backyard. I believe he finally had to eat one before being convinced that they were really pretty yucky.

Here's the growth rate of outstanding credit card balances in the United States:

Growth has been steadily accelerating over the past eight quarters, which suggests that consumers are keeping up their high spending levels only by relying on plastic. That can't continue unless wages start to grow significantly, something the Fed is determined to prevent.

Eventually credit card spending will have to stop growing, and with it consumer spending will stop growing too as long as real wages are flat. And then last year's interest rate hikes will kick in.

And then everyone will be shocked when economic growth sputters and dies out. How could that have happened?

NOTE: Credit card spending is highly seasonal, which is why I used year-over-year growth in the chart above.

With today's capitulation by Sen. Rick Scott, Joe Biden has now managed to trick the entire Republican Party into becoming Democrats on the subject of Social Security and Medicare. All of Congress is now completely opposed to any cuts in these programs.

This doesn't mean anything in the long run, of course, since Republicans will change their minds soon enough. It will be cloaked in the language of retirement age, inflation adjustments, "nobody currently over 65," and other subterfuges, but it will be cuts nonetheless.

Still, for a brief moment we can all shake our heads at the sight of doddering, nearly senile Joe Biden using the crudest kind of reverse psychology to dupe Republicans into becoming Social Security's loudest defender. How does he do it?

I would like to ask a question that is rarely asked: How are our 40-year-olds doing?

Here's the thing. We've been obsessing forever about how our kids are doing. Every generation is convinced that schools are failing our children and our kids are growing up unable to read, do math, or understand civics. It is a national crisis.

The reason we're able to obsess like this is because we can test our kids. Why? Because they're kids. And they're in school. If we tell them to take a test, they take it.

But adults? Forget it. They're not going to sit around taking reading or math tests just because someone wants them too. They've got better things to do.

So once again: How are our 40-year-olds doing? The answer is that we don't know and, really, no one cares.

But! We can infer a few things. If Johnny couldn't read in 1955, then 45-year-old John was probably a poor reader in 1995. If our nation's kids were at risk in 1983, then our nation's adults should be dullards today.

Is either of these things true? In general, do we think that the basic educational level of American adults has been collapsing for the past half century? I see no evidence for that, and I never really hear anyone complaining about it. It's only the kids they're worried about.

Bottom line: I suspect that panic and culture wars aside, today's kids are doing about as well as they did 20, 40, or 60 years ago. For my money, there's only one really big problem in American K-12 education:

We've made essentially no progress on the Black-white education gap over the past four decades. And guess what? That does make a difference in adulthood. If we're going to worry about our kids' education, this is what we should worry about.

According to a legal filing from Dominion Voting Systems, which is suing Fox News, many of Fox's top stars were contemptuous of claims that Donald Trump had been cheated out of the 2020 election. Among those stars was Tucker Carlson. But that didn't matter to him:

On Nov. 12, in a text chain with Ms. Ingraham and Mr. Hannity, Mr. Carlson pointed to a tweet in which a Fox reporter, Jacqui Heinrich, fact-checked a tweet from Mr. Trump referring to Fox broadcasts and said there was no evidence of voter fraud from Dominion.

“Please get her fired,” Mr. Carlson said. He added: “It needs to stop immediately, like tonight. It’s measurably hurting the company. The stock price is down. Not a joke.” Ms. Heinrich had deleted her tweet by the next morning.

They sure know their audience.

I've written approvingly about phonics before, and a few days ago Nick Kristof did too. This started up a bit of chatter about how we teach our kids to read. The background for this is twofold:

• In the longer term, it's part of the Reading Wars, which began in its modern incarnation with the publication of Why Johnny Can't Read in 1955. I'll spare you the gruesome details, but suffice it to say that over the next few decades phonics (learning to sound out letters) became a conservative darling while "whole language" (learning entire words and their meanings) became a liberal darling.
• More recently, Kristof noted that in the latest NAEP test two-thirds of fourth graders were not proficient in reading. He suggested that the answer was to ditch whole language instruction and switch to phonics nationwide. I was nodding along when I read this.

But then Bob Somerby reminded me that, contrary to popular opinion, kids read better today than they ever have before:

If we restrict ourselves to white kids, the NAEP tells us two things:

• The reading performance of today's white 9-year-olds is an entire grade level higher than in 1975.
• Nonetheless, only 42% of today's white 9-year-olds are proficient in reading.

That . . . seems a little odd, no? As it happens, I think one problem is that NAEP's definition of "proficient" is more stringent than you'd think, though it's hard to decode what they really mean by it. Still, any reading goal that requires "complex inferences about the characters’ actions, motivations, or feelings" is expecting a fair amount out of 9-year-olds.

If, instead, you look at the NAEP's "basic" reading level—which I suspect matches the popular notion of ordinary literacy—suddenly 63% of all kids and 73% of white kids can read.

Combine this with the fact that America's school kids of all races have been getting steadily better at reading for the past four decades, and it makes you wonder if maybe we don't really have a big reading problem in the first place. And while I still think the evidence favors a strongly phonics-based approach for young readers, perhaps it matter less than we all think?

POSTSCRIPT: I'll add something that I've mentioned before. Our educational problems are, for the most part, not in primary grades. Academic performance in grades K-4 has made strong gains over the past few decades.

However, the gains mostly wash out by age 18. It's middle school and high school that we really ought to focus on, not primary grades. More about this tomorrow.

Earlier today I put up a chart showing that feelings of sadness and hopelessness in teens had jumped dramatically over the past four years. Keeping in mind that this jump happened in 2019—i.e., after social media and before COVID—I wondered what was going on.

Here's another chart—this time just for girls, who display much higher levels of depression than boys. I would like this one to go through 2021, but unfortunately the boffins at the CDC are able to produce a lengthy and beautifully formatted report about the 2021 survey but were not able to add a couple of columns to their database during that time. So this goes only through 2019:

I've broken this down by grade level. Don't worry: the sample sizes are large and the results are plenty accurate even when broken down by age and gender.

I drew a trendline through the 1999-2015 results to see how things have gone since then. Here's the upshot:

• 9th grade girls just followed their usual trendline.
• 10th and 11th grade girls suddenly spiked upward in 2017 and again in 2019.
• 12th grade girls showed little movement until 2019, when their results skyrocketed.

I don't have anything special to say about this because I can't figure out anything to say. It just seems odd that 2017 and 2019 results are so different for teen girls separated in age by only a year or two.

It's possible that these are statistical anomalies and all the grade levels experienced pretty similar changes. If that's the case, we're back to our old question: What happened in 2017? And 2019?

This is the Paris Opera House as seen from the roof of the Galleries Lafayette department store. The Eiffel Tower is in the far distance, photo bombing everything in its path as usual.

Yesterday I was reading a column in the Washington Post by Kate Woodsome about the mental health of American teenagers, and it was, as usual, startling. American teens seem to be in terrible shape. In particular, Woodsome shared a CDC chart showing that more than half of teen girls experience "persistent feelings of sadness or hopelessness." Half! That seems hardly believable.

I traced this back to its source in order to get a longer-term view, and then, instead of looking at raw levels I looked at the change from year to year. Here it is:

For the most part, the numbers go up and down and don't change a great deal—until 2019 and 2021. In the space of just four years, it jumps by a third for both boys and girls.

This is jaw dropping. Nothing changes that much unless there's some kind of serious external change driving it. But what?

It's not social media since everything was fairly stable for the entire century until 2019. (Though maybe the culprit is one specific bit of social media that started up around 2017?)

It's not COVID since it started before the pandemic.

It's not George Floyd, which happened in 2020.

It's not video games.

But what is it? The only thing that comes to mind is mass shootings, which do seem to have gotten considerably worse lately. Another possibility, as it always is, is some kind of statistical malfunction.

Any other ideas? What could have happened around 2018-19 that would have sent American teens literally off the deep end?

POSTSCRIPT: The full CDC report is here. For what it's worth, bullying has been stable lately. Threats of violence and sexual violence were steady. Sexual behavior was unchanged. Drug and alcohol use was mostly either stable or down. The answer probably lies somewhere else.