Statistical Methods Links, 2/17/2025

Lyman Stone on national average IQ; Jean Twenge issues a correction; Scott Winship on Americans' satisfaction; Jonathan Haidt on a Lancet study of phone-free schools

Lyman Stone has doubts that national average IQ’s in Africa are as low as some measures indicate. He points out that the estimated averages for some countries are so low that they imply that the people one standard deviation below normal there would not be able to take care of themselves.

He points out that these estimates have changed a lot over relatively short time periods. To me, that says that casts doubt on their reliability.

He says,

By the way— a pure hereditarian would insist that African IQ is around 80-90. Why? Because in the highest-quality data on African-ancestry people, namely, tests in the U.S., African-ancestry people tend to score around 80-90 on average. If you’re really a strong hereditarian, the U.S. environment should not boost IQ very much, which implies that the national IQ data is incorrect

This argument works better rhetorically than statistically. “African-ancestry” is not a single, homogeneous population. In fact, there is considerable heterogeneity among African populations. A better comparison would be Nigerian-Americans with Nigerians, or Malawi-Americans with Malawians. But there would still be selection effects to deal with.

It strikes me that one might be able to trust the implications of major estimated differences in estimated IQ across countries. But the margin of error is too wide to attempt to answer finer questions, such as how these differences might be changing over time. It reminds me of trying to analyze changes in productivity trends, when the measured changes are within the margin of error in the data.

Jean Twenge writes,

the numbers in the figure above for “can’t do anything right” and “my life is not useful” included those responding agree, mostly agree, or neither (and for “enjoy life,” they included those responding disagree, mostly disagree, or neither). Goldsmith-Pinkham responded that that was probably right, but he strongly disagreed with analyzing the data that way.

I now think Goldsmith-Pinkham is correct: It’s more informative and fair to show only the percentage who agreed, leaving out those who responded “neither.”

After conceding this point, she goes on,

These are stunning changes in teens’ views. Compared to 2012, twice as many teens in 2023 did not enjoy life (a 107% increase). Fifty-three percent more think their life isn’t useful, and 39% more believe they can’t do anything right.

If you read this quickly and decide that fifty-three percent think that their life is not useful, you need to look at figure 2 in her post. The chart shows that a little less than 20 percent think that their life is not useful, and this compares to less than 14 percent in 2012. Her 53 percent is numerically correct for the change from 14 to 20, but to me it sounds like more of an epidemic than what is really there.

This reminds me of what I wrote about her book.

I was not always pleased by Twenge’s manner of presentation. She frequently augments a data point by adding “perspective.”

It is a great book, and the new edition promises to be even better. But I wish she would develop a habit of just sitting back and letting the data speak for themselves.

Scott Winship writes,

Whenever you read that 80 percent of Americans are dissatisfied with the way things are going in the US, you should remember that 80 percent are satisfied with the way things are going in their own life. Large majorities of Americans are satisfied with a variety of aspects of their lives. Gallup found in January 2023 that 90 percent were satisfied with their family life, 88 percent with their housing, 87 percent with their education, 87 percent with their work, 84 percent with their community, 81 percent with their health, 77 percent with their leisure time, 76 percent with their standard of living, and 71 percent with their household income. These are mostly down from early 2019 but up from 1995.

His point is that if you look at how people see their own economic condition, their perceptions are consistent with economic data showing low unemployment and high real wages. Their negative views of the economy seem to come from what they believe about how others are doing.

This tendency for people to rate things worse as the subject moves from their own personal lives (which they perceive accurately) to the nation is so pervasive that it has a name: “local positivity bias.” People dislike Congress but reelect their legislators. They think the nation’s schools are failing, but they like their kids’ local school. The phenomenon has also been labeled the “I’m OK—They’re Not Syndrome.”

In summary,

are Americans “right to believe their lyin’ eyes,” as [Oren] Cass claimed in a recent op-ed titled, “Three Cheers for Economic Pessimism”? This formulation begs the question of whether American beliefs about the economy conflict with objective measures. Cass and the declensionists are no more reliable guides to those beliefs than accurate interpreters of economic data. What Americans tell surveyors is consistent with the objective data, for the most part. There has been no long-term decline in economic conditions.

When it comes to analyzing trends in economic well-being, you can trust Scott Winship. Not Oren Cass.

Jonathan Haidt, Zach Rausch, and Alex McClean write,

Within each school, the heavy users of smartphones and social media are doing worse than light users, across multiple important outcomes including mental health and educational attainment. However, the fact that the two groups of schools did not differ on average leads the authors to conclude that while there are consistent associations with harm, phone-free policies alone don’t reduce those harms.

They did into the flaws of the study, and I believe that they are correct. But there is a general human tendency to search for flaws only in a study that contradicts your point of view, and when a study that supports your point of view comes along you pass it along without giving its methods such careful scrutiny.

The effect of phone use policy in schools is not an easy thing to test empirically. Are you sure that policies are being enforced and are making a difference in phone use in the school? Are you comparing different schools, in which case other differences come into play? Are you using reliable indicators of mental health or other outcome measures?

substacks referenced above: @

Flaws in a Recent Lancet Study on Phone Use in Schools

@

Twice as many teens say they do not enjoy life