Unanticipated problems can only be solved once they emerge.

Some very interesting points made in Sweden Has an Education Crisis, But It Wasn’t Caused by School Choice by Tino Sanandaji.

The issue is that PISA scores for Sweden have declined dramatically in the past decade or so. I have discussed elsewhere some of the limits on the usefulness of PISA scoring. But stipulating the PISA scores, what explains the Swedish decline? The search for an answer is made more difficult by the fact that Sweden has undertaken numerous education reforms in the same time period and has admitted much higher levels of refugees than in the past. So the decline could be explained, at a first order of analysis, by ineffective reforms, by the disruption that comes from reforming, by changing demographic/cultural mix, etc. Many in the US education establishment like to believe that the decline is a result of experimentation with two specific reforms; privatization and vouchers. Sanandaji brings that analysis into question.

I like his description of the teething process of implementing any major new social policy.

To some extent, the childhood ills of the voucher system can be viewed as part of the Hayekian learning process of building entirely new systems — unanticipated problems can only be solved once they emerge.

Sanandaji's analysis:

But in my view, the main culprit was the experiment with radically new pedagogical methods. The Swedish school system used to rely on traditional teaching methods. In recent decades, modern “individualist” or “progressive” pedagogic ideas took hold. The idea is that pupils should not be forced to learn using external incentives such as grades, and children should take responsibility for their own learning, driven by internal motivation. Rote memorization and repetition are viewed as old-fashioned relics. Teacher-led lectures have increasingly been replaced by group work and “research projects.”

Grades have been abolished below the sixth grade, and homework heavily reduced. According to TIMMS (a test similar to PISA), the average hours Swedish students spend doing mathematics homework declined from 2.1 hours per week in 1982 to 1.1 hours in the late 2000s. Despite criticism from teachers, the Swedish school board has ruled that pupils are allowed to have mobile phones and wear caps in class.

The Rousseauian experiment in pedagogic method has caused a collapse in discipline and non-cognitive skills in general. The PISA report shows that Sweden has become an outlier in terms of expressions of non-cognitive skills: “Sweden has the highest proportion of students who arrive late for school among OECD countries,” the report notes. And classroom discipline has declined along with teacher authority: “The disciplinary climate in Swedish classrooms is generally more negative than on average across OECD countries,” PISA writes.

Duke University psychologist Angela Duckworth has shown the importance of perseverance and “grit” for academic success. Students who are not challenged are less likely to develop such traits. The PISA report again: “Students in Sweden reported lower levels of perseverance than students in most other OECD countries.”

Perhaps the single biggest problem is the decline in learning tempo. Once students are used to a slow pace, it becomes hard to demand more. The accumulated effect of reducing the pace of teaching over many years is substantial. The PISA report:

A 15-year-old student in Sweden in a typical study programme receives 741 hours of intended instruction time per year, compared with 942 hours on average across OECD countries.

Non-cognitive skills, voluminous purposeful practice - these are so well known and documented, why don't we rely on them more? We shy away from anything that contradicts our Woebegonian ideal where everyone is above average.

US Education: Expensive and ineffective? Not so fast.

I came across an interesting cache of data which begins to answer a supposition I have held for a good while; that American schools are better than they are given credit for. In international educational comparisons which are conducted from time to time, the US usually comes out somewhere in the middle of the pack despite spending inordinately more money per student in both absolute dollars as well as in terms of percent of GDP. Having seen and experienced education in Europe, Asia and Australasia (and Africa) and hired people who were the product of education systems across the globe, I have long harbored the suspicion that those rankings were not telling the complete story. It is notoriously difficult to do international comparisons of performance in any field – health outcomes, productivity, crime, etc. There tend to be three sources of problems (other than outright gaming of the rankings) - 1) Definitions (ex. what constitutes a student?), 2) Ensuring like is being compared to like, and 3) Reliability and consistency of data collection. Even when these are done well and accurately, there is still a missing back-story and context.

The problems of ensuring comparability are easy to anticipate. What constitutes a 15 year old (by age, by grade, by birthday at time of test, etc.)? Are all 15 year-olds included? What about the 10% in private schools or parochial, what about home schoolers, what about charter, what about those who have already dropped out by age 15? Does it include all boys and all girls? How confident are we that the same test in different languages constitutes the same test? Etc. The issues will vary enormously by country.

My supposition has been that in addition to the known issues surrounding data normalization, there is also likely an aspect of the Wisconsin-Texas paradox as well ( see post, The Texas-Wisconsin Paradox and intergenerational income mobility). The US is unusual in many regards, particularly in that it is large, highly heterogeneous (by ethnicity, by culture, by religion, etc.), and very productive/innovative and has been so for long stretches of time. There is complexity that comes with these attributes and there are no real counterparts against which to compare it.

The OECD’s Program for International Student Assessment (PISA) has gained traction in the past couple of decades as an effort to compare international performance in education. There are three areas of performance measure, literacy, maths and science. The most recent reading assessment was performed in 2009 and the detailed results were released in Highlights From PISA 2009: Performance of U.S. 15-Year-Old Students in Reading, Mathematics, and Science Literacy in an International Context.

While I think the PISA effort is estimable, I believe we are still a few more cycles away from it being robustly reliable. I suspect that there are particular issues of Like-to-Like comparisons. For example, the most recent results show Shanghai leading the globe with an average reading score of 556 (reading scores are normed at 500). While Shanghai is an up and coming and wealthy city, I find this result implausible. It is a city of some 25 million with an inordinate percentage of the residents being emigrants from the countryside in the past twenty years, i.e. farmers drawn to the allure of manufacturing in the city. Everywhere in the world there is always a couple of generation lag in assimilation and education attainment in the migration from country to city. The PISA report would have us believe that uniquely Shanghai has broken this worldwide phenomenon. Perhaps, but more likely they have simply sampled only their best schools in the city.

Flawed as it might be, PISA will continue to be refined and the data will become more reliable over time. It has the advantage that it is now being administered to ever larger numbers of countries (sixty in this most recent survey) and is no longer restricted to the OECD.

So with all those caveats, what can the PISA scores tell us about US educational performance. I think, lurking in the data, there is a confirmation that the US does a far more effective job of educating its population than virtually anyone else in the world, despite the heterogeneity of its population and despite the very large percentage of its population that is first generation emigrant (approximately 12% now). Efficiency is a different, and no less important, an issue. However, I think perhaps one half of the accusation of ineffective and inefficient can be taken off the table.

There is an argument that when you have a heterogeneous population, that like has to be matched to like for comparison purposes. Ideally, as the primary vector for difference, one would look at self-identified cultural heritage (or religion as a proxy) as the source of differences in performance. Since that data is not easily or reliably available, an alternative is to use race as a proxy. Compare US non-Hispanic whites to western Europeans (as the primary cultural stock), Hispanic Americans to Latin American countries, Asian Americans to eastern Asia countries and Black Americans to African countries. Lest there be confusion, this is not an assertion that there is biological determinism at work. I believe that that view is well dispatched in such accessible accounts as Steve Olson’s Mapping Human History: Genes, Race, and Our Common Origins. Rather it is a recognition that an individual’s cultural framework has an inordinate influence on personal choices and actions and that culture, while mutable, can be extraordinarily long lived. While all assimilated Americans will have greater behavioral affiliation with one another than with virtually any other country, each sub-group within the US has strong traceable influences from their ancestral cultural origins.

So how do US students compare against their respective cultural cohorts across the globe? Contrary to most characterizations, pretty well.



Click on the above chart to enlarge and then click a second time to enlarge once more. There is a lot of information crammed into a small space. East Asian countries are light green. European countries (along with Australia, New Zealand and Canada), dark blue. Islamic countries, dark green. Latin American countries are light blue. Israel is mauve. Trinidad and Tobago is orange. T&T’s population is one third sub-continental Indian and one third of African ancestry with the balance of mixed ancestry. There are no African countries that participated in PISA 2009 and so T&T is used as the closest proxy. The solid red bar is the average American performance, 500, on reading. Right at average. The green bar with a heavy red border is the breakout for Asian Americans (541). The dark blue bar with heavy red border is non-Hispanic white Americans (535). Light blue with heavy red border is Hispanic American (466) and orange with heavy red border is USA Black (441). There are no available comparisons for Native Americans.

The culture, institutions, and human capital that the immigrant groups brought with them and pass on to their children are positively associated with local development in the US.

Some good long while ago, I looked at the PISA international education results based on continents of origin. I did this because, while I think that PISA is well intended and broadly well executed there are still some iffy data, and it seemed to me that US educational outcomes were being ill-reported. For example, the US has had a large, essentially rural and low education attainment immigration influx in the 1980-90s and yet the PISA scores hold steady. That would seem to indicate that the schooling system is assimilating a large population with no impact on PISA scores which is quite an accomplishment.

Public education is, in my own opinion, dramatically underperforming versus the resources invested but none-the-less, I think it is accomplishing much more than it is being given credit for.

The results were striking. Caucasian Americans perform better than all European countries on virtually every performance measure. Same with Asian Americans and their counterparts in Asia, African Americans and their counterparts, and Hispanics and their counterparts.

My one concern with this analysis was that it implied a decades and centuries long cultural resilience. That the values and behaviors of country of origin had very long durability, even past the point of conscious awareness. This is entirely consistent with the work of Gregory Clark (The Son Also Rises and Farewell to Alms) which reveal not just decadal persistence of values and behaviors but centuries long duration. The fact that there is a good body of evidence supporting the hypothesis still doesn't make it particularly easier to grasp. Which behaviors are being sustained and by what mechanism(s)? We don't, however, have to understand a phenomenon in order to acknowledge its reality.

This study, Does It Matter Where You Came From? Ancestry Composition and Economic Performance of U.S. Counties, 1850 - 2010 by Scott L. Fulford, Ivan Petkov, and Fabio Schiantarelli, provides some good evidence supporting the idea that culture indeed has a very long half-life. From the Abstract:

The United States provides a unique laboratory for understanding how the cultural, institutional, and human capital endowments of groups interact and shape economic outcomes. In this paper, we use census micro-sample information to reconstruct the country-of-ancestry distribution for US counties from 1850 to 2010. We also develop a county-level measure of GDP per capita over the same period. Using this novel panel data set, we investigate whether changes in the ancestry composition of a county matter for local economic development and the channels through which the cultural, institutional, and educational legacy of the country of origin affects economic outcomes in the US. Our results show that the evolution of the country-of-origin composition of a county matters. Moreover, the culture, institutions, and human capital that the immigrant groups brought with them and pass on to their children are positively associated with local development in the US. Among these factors, measures of culture that capture attitudes towards cooperation play the most important role, perhaps because they also reflect the quality and functioning of local institutions. Finally, our results suggest that while fractionalization of ancestry groups is positively related with county GDP, fractionalization in attributes such as trust, is negatively related to local economic performance.

What is the cost of educational quality

From The Value of Smarter Teachers: International Evidence on Teacher Cognitive Skills and Student Performance by Eric A. Hanushek, Marc Piopiunik, and Simon Wiederhold. An excellent topic area. From the Abstract:

International differences in teacher quality are commonly hypothesized to be a key determinant of the large international student performance gaps, but lack of consistent quality measures has precluded testing this. We construct country-level measures of teacher cognitive skills using unique assessment data for 31 countries. We find substantial differences in teacher cognitive skills across countries that are strongly related to student performance. Results are supported by fixed-effects estimation exploiting within-country between-subject variation in teacher skills. A series of robustness and placebo tests indicate a systematic influence of teacher skills as distinct from overall differences among countries in the level of cognitive skills. Moreover, observed country variations in teacher cognitive skills are significantly related to differences in women’s access to high-skill occupations outside teaching and to salary premiums for teachers.

Seems well structured and a good-faith effort to get at some important answers. It is handicapped, at a minimum, by its reliance on PISA data which has some well-noted issues.

But perhaps there is enough substance there to at least warrant some confirmation of general assumptions. The implications are interesting.

I have documented in the past that PISA scores are highly correlated with cultural groupings to the continental level at least, and sub-regional in locations such as Europe where there is sufficient data, and in the US where our ethnicity obsession allows a comparison of American PISA scores by cultural heritage (white Americans to Europe, Hispanics to Central and South America, Asian-Americans to Asia, etc.). The notable outcome of that analysis is that, while we bad-mouth our public education system, it delivers superior results when viewed from a cultural heritage perspective. White Americans perform better than all European countries other than Finland (and then only on the reading score), Asian-Americans better than any Asian country, etc. Yes, we have an inordinate number of issues with our public schools, but they are not as bad as is often made out to be.

There are almost innumerable material independent variables which affect educational outcome. A by no means exhaustive list:

Education policy
Centralization versus decentralization
Teacher training
Curriculum
Teacher quality (either years teaching and/or IQ)
Classroom management
Familial structure of students
Socioeconomic status
Per student spending
Class size
Age mixing
Gender mixing
National productivity
Career choices and compensation
Physical context of the school
Cultural valuation of education
Cultural, class, economic, ethnic, religious, etc. class heterogeneity
Relative starting point of children
Relative starting point of teachers
etc.

Disentangling which combination of all those variables has how much affect on which student populations under what circumstances is an enormous question which we have spent fifty years and more trying to understand with some progress but not near as much as we would wish. At this point, it is pretty clearly understood that 1) there are no silver bullets, and 2) one size does not fit all.

Hanushek, Piopiunik, and Wiederhold contribute some useful information.

With both approaches, we consistently find that differences in teacher cognitive skills across countries are strongly associated with international differences in student performance. In terms of magnitude, a one SD increase in teacher cognitive skills is associated with an increase in student performance of 0.10-0.15 SD. Since PISA scores represent the cumulative learning of 15-year-olds, this suggests an average learning gain of about 0.01-0.015 SD per year.

0.15 SD is about 15 IQ points. The implication is that if you lift your average teacher IQ from 100 to 115, you will raise you student performance from 100 to 101.5-102.3.

IQ is distributed on a Gaussian distribution. Only about a third of the population have an IQ above 115. Because it is Gaussian, the further along the curve you got (high and low, the disproportionately fewer there are). At two standard deviations above the mean, there are 3 people above and IQ of 130 (roughly).

Consequently, moving resources from the upper end of the distribution into schools will be tied with a reduction in supply for high cognitive jobs as well as a corresponding increase in cost, representing a loss of productivity.

If what Hanushek, Piopiunik, and Wiederhold have found is true, then the economic question becomes, "Is the loss of supply of 115 individuals for general economic activity more than made up for by increase of general population productivity associated with a move from population average IQ of 100 to 101.5-102.3. I don't know the answer. There are also, of course, all sorts of other ethical and you moral questions as well.

The economic question is hugely pertinent though, especially in a free market environment where individuals are able to make the personal choices that allow them to make decisions which optimize their well-being. If you are living in an agricultural country with low development and poor infrastructure and are out in the rural areas, a job as teacher, even if you are 1 SD above everyone else, might be worthwhile at an average wage because it keeps you close to home. If you are in a developed nation with a free market and are in the city with a 1 SD IQ, it might be the difference between an income of $65,000 as a teacher and $100,000 as a marketing manager. Who makes what choices is constrained by context of location, circumstance, economy, law, policy etc.

And also culture. Some places such as Japan and Finland keep teacher compensation low in return for significant cultural prestige. Te be a teacher is to be a cut above the average in social standing.

We are way too early in our understanding of the metrics, the causal dependencies and the supply elasticity.

My suspicion is that in advanced economies that are highly urbanized and where civil rights are well protected so that everyone has full access to all choices, the cost of pulling in +1 SD teachers will be very high and very apparent and the system return on that investment over a generation will be lower and later and less visible than will make such a shift in resources a compelling proposition at a national level. Local - quite possibly. Nationally - I wonder.

Gun control has no impact on gun deaths.

From California's comprehensive background check and misdemeanor violence prohibition policies and firearm mortality by Alvaro Castillo-Carniglia, et al. From the Abstract:

Purpose

In 1991, California implemented a law that mandated a background check for all firearm purchases with limited exceptions (comprehensive background check or CBC policy) and prohibited firearm purchase and possession for persons convicted within the past 10 years of certain violent crimes classified as misdemeanors (MVP policy). We evaluated the population effect of the simultaneous implementation of CBC and MVP policies in California on firearm homicide and suicide.

Methods

Quasi-experimental ecological study using the synthetic control group methodology. We included annual firearm and nonfirearm mortality data for California and 32 control states for 1981–2000, with secondary analyses up to 2005.

Results

The simultaneous implementation of CBC and MVP policies was not associated with a net change in the firearm homicide rate over the ensuing 10 years in California. The decrease in firearm suicides in California was similar to the decrease in nonfirearm suicides in that state. Results were robust across multiple model specifications and methods.

Conclusions

CBC and MVP policies were not associated with changes in firearm suicide or homicide. Incomplete and missing records for background checks, incomplete compliance and enforcement, and narrowly constructed prohibitions may be among the reasons for these null findings.

Just another study among thousands, but interesting none-the-less. I frequently observe that the suggested gun control policies shrilly advocated for after some tragedy almost never are pertinent to the tragedy itself. In other words, even had the recommended policies been in place, the facts of the case are such that the tragedy would have happened anyway.

That California has among the most stringent restrictions in the nation makes this a notable study. One can conjure all sorts of scenarios why that might be, but it is interesting that that is the fact.

UPDATE: On reflection, I wonder if there might not be a further point of interest to be teased out from this study. America has a higher murder rate than most of our OECD counterparts. However, complicating the analysis, it is also true that the US has the most heterogeneous population compared to all our OECD counterparts. Given America's sheer size, in population and geography, there are simply no real benchmarks with whom to compare.

From other studies, it is observable that America tends to be average when you compare it to individual other countries but when you make the comparison based on culture of origin, the US tends to come out way ahead. Education is the one with which I am most familiar. White Americans outscore all European countries, (usually significantly), on PISA scores other than the Finns on reading. Asian Americans outscore all Asian countries on PISA scores. African Americans outscore all majority black countries on PISA scores.

The same is true of productivity and income.

So while there are differences within America between culture-of-origin groups, all people in America from every other culture do better than their home counterparts when they move to America.

My suspicion is that the same is likely true for the violence. We know that, in rounded numbers, 50% of all murders are committed by African Americans, 25% by Hispanics and 25% by European culture origin Americans despite their corresponding presence in the population of 15%, 15% and 70% respectively. My suspicion is that were you to do a rigorous comparison, one might find that the murder rate for white Americans to be near or lower than that of their European counterparts and the same for African Americans (for majority black countries) and Hispanics (for Hispanic countries of origin).

I have never seen such a rigorous study done but there are all sorts of intriguing snippets that are at least suggestive. While we are rightly dismayed in America about rare but tragic mass school shootings, it is easy to neglect that three or four of the top five global school shootings have occurred in Europe where gun rights are generally much, much more tightly circumscribed.

The overall point is that there are good reasons to believe that cultural dispositions are likely a far greater driver of base rate outcomes than are particular policies. Returning to education as an example, the broad outlines of public school education among the 50 states are reasonably similar. There are always many policy trials and experiments going on in the different states, some small number of which occasionally demonstrate effectiveness and then become widely adopted. That said, there is insufficient variance in policy to explain the consistent differentials in cultural origin differences consistent across the nation.

What we don't know, and would be useful to know, is whether America is inherently violent because of its inherent heterogeneity, because of its Second Amendment, or because of the particular, and evolving, mix of cultural origin disposition towards violence. Were we to conduct such a rigorous study, my suspicion is that we might find that the Second Amendment is far less of a driver of violent death than we have all along assumed.

So if results are driven much more by culture than by policy, it is reasonable to speculate about its impact on the propensity to violence in a state such as California which has long had a strong orientation towards limiting Second Amendment rights. Especially given that California is also the state with the greatest increase in Hispanic emigration (legal and otherwise). In an arena where there is both an increasing constriction on gun rights but also an increasing share of the population with a cultural origin in a culture with a dramatically higher manifestation of violence, perhaps the two are operating at cross-purposes to one another.

A further complicating factor is that California is also the state with among the worst school systems, worst income inequality and the highest rate of population in poverty.

Specifically, in the forty year time frame Castillo-Carniglia, et al are examining, it is conceivable that in fact the gun control laws are in fact effective but that their effect is being swamped by the increase in violent country-of-origin culture, and inequality, and poverty.

All interesting but speculative questions.

History

Roman terracotta amphorae found in Pisa, classified by type. This section of the wall has vessels used to transport goods that are dated to the 1st C BC through 1st C AD which is in the Museum of the Ancient Ships of Pisa in #Pisa, #Italy.#archaeology #tuscany #italia #ancient pic.twitter.com/lK6AM5xFTY

— Decimus Claudius (@decimusclaudius) June 8, 2024

 

The Texas-Wisconsin Paradox and intergenerational income mobility

There has been a spate of papers in the past couple of years calling into question whether the US is the highly mobile, equal opportunity nation we like to consider ourselves to be, compared to the developed countries of the Old World and OECD (See American Exceptionalism in a New Light by Markus Jantti, et al and Intergenerational Social Mobility: The United States in Comparative Perspective by Emily Beller Michael Hout). This argument has been contested by others (See Why economic mobility measures are overrated by Tyler Cowen). Having lived and worked in Europe, North America, and Asia, this new view of lower American mobility compared to others is inconsistent with my experience and perception of those countries. Anecdotal assessment does not trump empirical data but it is a catalyst to skepticism. Cowen and others raise important points. I want to explore an avenue of criticism which I have not seen discussed elsewhere; the impact of cultural heterogeneity on aggregate intergenerational income mobility statistics.

In a “perfect” world, someone in one quintile would have an equal chance of ending up in any one of the five income quintiles (i.e. the difference of the quintile of the parental home into which they are born and the quintile at which they arrive at some point in life, say retirement or death). In order to take this randomness as the standard of fairness and equality, one has to assume that the quintile in which one is positioned, both at birth and later at death, is simply a function of luck. This assumption is a common one among the glitterati and chattering classes. Anything less than random distribution is taken as evidence of structural impediments, discrimination, protectionism, etc. In some countries these structural and cultural impediments are real though I do not believe them to be in America.

However, if one accepts that people are purposeful, then the quintile to which they evolve is a function of exogenous circumstances interacting with cultural and individual attributes such as values, behaviors, knowledge, skills and patterns of decision-making. For any given set of exogenous circumstances, some individuals/cultures will be better positioned to make more productive decisions than others. Likewise, over a period of time and a range of exogenous circumstances, it is likely that some individuals/cultures will be more responsive and adaptive than others. It is demonstrable that certain cultures are more robust over time (they survive and evolve) and/or are more prone to greater productivity. In other words, values, behaviors, knowledge, skills and patterns of decision-making are the decisive determinants of productivity and continuity over time (and over a broad spectrum of random exogenous events) rather than luck.

The validity of the expectation that fairness is represented by equal intergenerational mobility is belied by looking at national economies. Were one to accept the premise that fairness is defined as random quintile movement between generations, one would expect to see dramatic changes in the order of national economies each 25 years. While virtually all economies have improved in the past 250 years (ten generations), the relative order of national economies has remained relatively stable, i.e. there is very little intergenerational movement of national economies. Intuitively we understand that this a function of national cultures and institutions rather than random luck. Countries with stable law, property rights, personal and commercial freedoms, etc. combined with cultural attributes such as future-orientation, tolerance, risk aversion, personal accountability, diligence, work-ethic, trust, etc. are and remain more productive than other countries lacking these attributes. Countries that have evolved up the hierarchy of productive countries (Chile, Singapore, South Korea, Taiwan, etc.) have all been countries that have a winning combination of such elements. There is no correlation between a country’s luck (natural endowment of resources) and its hierarchical position of productivity.

Countries that are already productive and have robust cultures that predispose them towards further productivity retain their positions in the rankings. This makes sense to most people and we do not expect dramatic changes in rankings from generation to generation because we do not believe that national productivity is simply statistical luck. Nor do we believe it should be.

Why is it that we expect that dynamic to be different at the individual level? We don’t. We want the rules to stay the same for all individuals so that there is some form of manageable predictability on which to base life-decisions but we don’t expect all individuals to make equally good decisions. The reality that there are some configurations of culture (values, behaviors, knowledge, skills and decision-making) which are more adaptive to a broader range of exogenous circumstances is simply an evolutionary reality which can be transmitted via cultural structures. While we can identify a portfolio of specific cultural attributes which correlate with productivity, we are not at a point of knowledge where we can identify the exact mix and relative importance of the particular attributes. For example are personal freedoms more important to productivity than stable laws? We know they are both important but we don’t know how important each is overall or relative to each other.

At the level of individuals, there is a broad range of evidence, which individually is not particularly robust but which cumulatively is more compelling, supporting the idea that quintile placement is not a matter of random luck but is a purposeful achievement based on certain combinations of values, behaviors, knowledge, skills and patterns of decision-making (see Creating an Opportunity Society by Ron Haskins and Isabel Sawhill for empirical evidence of the significant impact of behavioral norms on life achievement as well as A Farewell to Alms by Gregory Clark for evidence regarding the impact of “middle class” cultural values). See Surnames and the laws of social mobility for evidence regarding the durability of constant re-achievement of success across multiple generations.

Once we accept that quintile placement is not a matter of luck but of purposeful achievement, then we are freed to investigate what variables in which combination are most likely to yield the highest level of robust, continuing productivity over time. In a dynamic, heterogeneous, and uncertain environment, it is difficult to pin down cause-and-effect. At the same time there are some trade-offs that emerge under many circumstances: With more freedom comes greater inequality; With greater equality comes lower productivity; With greater heterogeneity comes greater inequality, etc.

This all serves as background to what I suspect may be happening with regard to measured intergenerational income mobility between the US and other developed countries in the OECD. The first point to make is that the US is the most culturally heterogeneous large OECD country (Singapore comes close in terms of heterogeneity but is only 3 million people.) Broadly the US is made up of 65% descendant of northwestern Europeans, 15% descendants of Africans, 15% descendants of Hispanics (Mesoamericans and Spanish heritage), and 5% other. In all other large OECD countries, there is far greater cultural, ethnic, and temporal homogeneity (temporal homogeneity reflecting the degree of recent emigrants). Most other OECD countries are at 85-95% homogeneity compared to the US’s 65%. This opens up the issue of apples to oranges comparisons, the nature of which can be illustrated by the Texas–Wisconsin educational paradox where Texas does a better job of educating (as reflected in test scores) each of its ethnic/cultural subgroups compared to Wisconsin but Wisconsin has higher average test scores.

How can each of the constituent components in Texas be higher but the overall average be lower? Differential scores between the constituent groups, i.e. Asians score highest on average and then in descending order, whites, Hispanics and then African-Americans. While all groups score higher in Texas than in Wisconsin, Texas has a dramatically higher proportion of its population from lower scoring groups than does Wisconsin. Basically Texas is a White, Hispanic and African-American state whereas Wisconsin is a White state. So which state has the better education system? The one that has higher overall average scores from a high scoring homogeneous group or the state in which each ethnic/cultural group scores higher than their corresponding group in the other state?

In the field of international education comparisons, Programme for International Student Assessment (PISA) is an attempt to compare educational K-12 outcomes across the OECD. The US usually lands pretty much in the middle of the test results seeming to indicate a relatively poor performance considering the per capita expenditures on education (far higher in the US than elsewhere). However – apples and oranges. The US is basically educating three heterogeneous groups – descendants of northwestern European and heirs to Western Civilization, descendants of Africans and heirs to 250 years of subjugation, and recently arrived descendants of Mesoamerican and Spanish civilizations. I have not seen a thorough disaggregation but the elements I have seen seem to indicate that the US whites score close to the top of the PISA tests compared to European countries, US African-Americans score at the top compared to the few predominantly black countries that participate in PISA and US Hispanics score at the top compared to Latin American countries. So an overall middle performance but with every sub-group performing higher than their apples-to-apples peers. The Texas-Wisconsin paradox at work again.

There is a similar paradox, little discussed, in terms of violence. Overall, the US is a much more violent country than most European countries but this ignores the fact that there are dramatically different levels of violence within different sub-groups. Nearly fifty percent of murders in the US occur within the African-American community (which represents approximately 13% of the overall population). If, once again, you break out violence by group, you find that white America has levels of violence close to or below most of the large European countries, African-Americans better than most black majority countries and Hispanics better than most Latin American countries. The Texas-Wisconsin paradox at work again.

My suspicion is that measures of intergenerational income mobility are similarly plagued by the Texas-Wisconsin paradox. I do not have the measures to prove it, in part because this is a relatively new field where there is not a lot of reliable historical data. If and when we are able to get that data, I suspect what we will find is that there is relatively high levels of intergenerational income mobility among whites, principally driven by educational attainment and family structure. I suspect we will find relatively low intergenerational income mobility among African-Americans but with a 10-20 year kink in the trend line around 1950-1970 when labor market, legal structures and cultural practices opened up opportunities for all African-Americans but which had the net effect of burgeoning opportunities for the elite in the community and depressing opportunities for the less advantaged. Specifically, the cycle of single parenthood which emerged in the 1950s onwards will have had a powerful depressing effect on income mobility overall. Finally, the record for Hispanics will be very mixed. I suspect that culturally integrated Americans of Hispanic descent (i.e. Hispanic Americans whose families have been in the US for several generations) will show the same level of intergenerational income mobility as whites but the larger number of new and first generation immigrants will show very low intergenerational income mobility and their large numbers will swamp those of the entire group. It usually takes large immigrant groups 2-4 generations to integrate and take on the macro-culture and until that culture change has occurred, the intergenerational mobility will continue to look like that of their original cultures of origin.

Every one of those common assumptions is simplistic, misguided, or downright wrong

A week or two ago I had a post, The Texas-Wisconsin Paradox talking about how comparisons between systems is very dependent on comparing apples-to-apples.

As if to drive home the point, there is a new study examining the PISA international education results. Despite spending more per student than anybody else, the US usually ends up somewhere in the middle on math and reading scores. In my paradox post I mentioned this in passing and indicated my skepticism that PISA was adequately taking into account the cultural heterogeneity of the US.

As if to validate my point, Derek Thompson summarizes a new study in his post, Why Gloomy Pundits and Politicians Are Wrong About America's Education System .

Here's what everybody knows about education in the United States. It's broken. It's failing our poorest students and codding the richest. Americans are falling desperately behind the rest of the developed world.

But here's what a new study from the Economic Policy Institute tells us about America's education system: Every one of those common assumptions is simplistic, misguided, or downright wrong.

When you break down student performance by social class, a more complicated, yet more hopeful, picture emerges, highlighted by two pieces of good news. First, our most disadvantaged students have improved their math scores faster than most comparable countries. Second, our most advantaged students are world-class readers.

Why break down international test scores by social class? In just about every country, poor students do worse than rich students. America's yawning income inequality means our international test sample has a higher share of low-income students, and their scores depress our national average. An apples-to-apples comparison of Americans students to their international peers requires us to control for social class and compare the performances of kids from similarly advantaged and disadvantaged homes.

It doesn't come out all that clearly in the post but apparently most other countries are oversampling participants in the tests from the higher income/social classes, groups who normally perform better on such tests. So, if I am reading the study correctly, the US had 20% of its test taking participants from the bottom quintile of income; what you would expect. On the other hand, Finland only had 6% of test takers from the bottom quintile. Of course their average is going to be higher.

This reminds me of the first time I encountered this apples-and-oranges issue as it relates to education measurement. I was in college and had only been back in the US three or four years and was still learning much information which my peers took for granted. I was particularly struck by the fact that Iowa had such an outstanding educational system based on a comparison of their SAT results to those achieved by other states. It was a couple of years before I discovered that in addition to SATs there was another standardized test, the ACT. In Iowa, as it turned out, everyone took the ACT but only those going on to college, and in particular those going out of state to university, would take the SAT. The consequence was that Iowa was reporting an SAT average for the top 5% of their students and was comparing it to states where all students took the SAT. One of my early lessons in making sure that apples are being compared to apples.

The report that Thompson discusses doesn't answer all the questions but it at least sheds light on some of the underlying issues of race, emigrant status, class, income, etc. that hinder like-to-like comparisons.

An Interior View of the North Gallery at the Camposanto in Pisa, 1834 by Antoine-Marie Perrot

An Interior View of the North Gallery at the Camposanto in Pisa, 1834 by Antoine-Marie Perrot 

Click to enlarge.

Obfuscating ideologically unacceptable research findings

Woof. I am pretty well educated and test pretty highly on reading comprehension. I still had to work hard to interpret the results of Long live your ancestors’ American dream: The self-selection and multigenerational mobility of American immigrants by Joakim Ruist. The abstract is:

This paper aims to explain the high intergenerational persistence of inequality between groups of different ancestries in the US. Initial inequality between immigrant groups is interpreted as largely due to differently strong self-selection on unobservable productive characteristics that are also highly persistent across generations. If these characteristics are responsible for a larger share of total inequality between immigrant groups than between individuals generally, the former inequality will be more persistent. This explanation implies the additional testable hypothesis that the correlation between home country characteristics that influence the self- selection pattern – such as the distance to the US – and migrants’ and their descendants’ outcomes will increase with every new generation of descendants. This prediction receives strong empirical support: The migration distance of those who moved to the US around the turn of the 20th century has risen from explaining only 14% of inequality between ancestry groups in the immigrant generation itself, to a full 49% in the generation of their great-grandchildren today.

Perhaps I am dense but it is not crystal clear what Ruist is saying. Academic obfuscation tends to occur under any one of three cases - when the academic is in a complex field (philosophy or physics), when the academic's findings are socially or ideologically unacceptable, or when the academic is in a bereft field (gender studies or ethnic studies).

Given that Ruist's native tongue is likely Swedish, this is an impressive level of structured obfuscation. But is it due to A) complexity, B) unacceptability, or c) mediocrity? I am going with B.

Ruist is testing a hypothesis I first heard from my father (a chemical engineer with an inquiring mind in all fields) sometime in the 1970s.

The hypothesis is that American exceptionalism was in part based on the how expensive and dangerous it has always been to immigrate to America. Only those who were the brightest and most motivated would elect to undertake the expense of immigration and the risk of death and/or destitution. This natural process of self-selection ensured that the US immigrant population constituted the best and the brightest. The corollary hypothesis was that the cultural attributes and behavioral motivation which brought such self-selected immigrants to the US would also persist through familial culture over time. We got the best people and their values persist over time.

I think this is what Ruist is testing and it is what he is finding support for. Interestingly, I did some research on this a few years ago using Program for International Student Assessment (PISA). The results are at US Education: Expensive and ineffective? Not so fast. I found the same thing as the Ruist and consistent with the hypothesis above: American cultural/ethnic groups outperform all their cultural peers. White Americans outscored Europeans, Asian Americans outscored everyone in Asia, Hispanic Americans outscored everyone in Central and South America, and African Americans outscored everyone with large Black populations.

Why do I suspect that this finding is unacceptable and therefore needs to be cloaked in obfuscating language? The implication is that culture matters, that some cultures are superior to others, and that culture is persistent over time. The dominant (and violently insistent) belief in academia is postmodernist critical theory blank slatism. That no culture is better than another, that all people can be engineered to be identically gifted, and that the state policy is better than familial behaviors.

Ruist's research, and the PISA numbers, stand in support of the hypothesis that life outcomes are heavily influenced by cultural values, that there is a founder effect related to the self-selection of immigrants and that the cultural values of those immigrants persist over time.

Cultural valuation of education

From How upwardly mobile are Hispanic children? Depends how you look at it. by Nathan Joo and Richard V. Reeves.

I have in the past made the argument that culture has a lot more impact on current circumstances than we otherwise acknowledge. For example, while the US performs only at a middling level on the international education testing PISA, this can be attributed to some degree to the cultural hodgepodge which is the US. When you look at PISA scores by race (as a crude proxy for cultural origin), Americans do much better than every corresponding group. In other words, white Americans outscore all Europeans, Asian Americans outscore all Asians, Hispanic Americans outscore all Central and South Americans and African Americans outscore those countries with a black plurality.

These cultural influences can echo on for generations. In Great Britain, Scots are famous for their carefulness with money (says I as a Scottish descendant). In The Millionaire Next Door or one of his later books, Thomas J. Stanley identifies which groups in the US are the top of the wealth league. Among the top are Scots Americans even though we are centuries since the last big immigrant surge from Scotland and long since one would have guessed that they had assimilated culturally.

While I make the argument that culture has an underappreciated impact over long durations, I have also been cautious about just how certain that reading is. Joo and Reeves seem to provide some confirming evidence about the importance of culture over appreciable time frames.

They look at education attainment of different groups of immigrants in comparison to the first generation of immigrants. As might be expected, second-generation groups attain greater education than first generation. One could anticipate that the amount of education immigrants bring is highly circumstantial to their country of origin. However, once here, all their children likely have roughly the same degree of access to higher education as one another. There really shouldn't be much difference between groups unless cultural valuation of education is 1) different between groups and 2) transmitted across generations. And that is what Joo and Reeves find. Look at the differences between groups, i.e. the differences in cultural orientations towards education based on region of origin.

Fascinating.

School Board centralization and declining school performance

From Links to Consider, 9/1 by Arnold Kling.  The subheading is Ray Kurzweil on the future; Peter Gray really dislikes Common Core; Alice Evans and Oliver Kim on East Asian exceptionalism; Aaron Renn on elite university connections. 

Commenting on Gray's essay, Gray ascribes public school decline to Common Core and Kling says no.  Kling observes:

I think that the deterioration of public schools has many causes, and it began before Common Core. Among the causes are the better opportunities for gifted women to go into fields other than teaching, the consolidation of school districts (making them much more bureaucratic and insulated from parents), the increased power of teachers’ unions, and the Woke takeover of schools of education.

Kling's explanation for the perceived causes of public school deterioration is well established, broadly discussed and most likely substantially contributive.  Other than this one:

The consolidation of school districts (making them much more bureaucratic and insulated from parents)

Maybe it is also a well worn proposed cause but I don't think I have seen it proposed and I spent the better part of a decade closely involved in a business immediately adjacent to education (children reading).  Regardless of whether it is well circulated, it is an intriguing idea.

It explains something I have considered in the past and have had little explanation for.  My observation has been that School Principles seem to have been much more well regarded and held in higher social status in the past than today.  Reading local histories, along with Church leaders, School Leaders seem to have been more intimately involved in town and county leadership than is my impression today.

I may be misreading the history.  And maybe my impression of the present is also false.  I live inside a large city (and always have done so) and maybe school leaders in the suburbs and smaller towns still have that old status.  But that's not my impression.

Much assumption has been that School Principles have suffered a degradation in social status over the past fifty years.  Why?  

My explanation to date has been that perhaps the degradation in social status was related to 1) declining fertility - fewer children overall, smaller families (and therefore less exposure to schools for a reduced amount of time), 2) fewer people having children at all, and 3) possibly the feminization education (more persuasion and tending of emotions and less executive function).  I am somewhat dubious of all three.  Possible candidates; even plausible, but . . .

School board centralization - to the extent it has really happened - makes quite a bit of sense.  If power is centralized the School Principle has less decision-making power, and therefore respect.  Further, with bureaucratic centralization, the School Principle also has less ability to respond to parents' expectations,  therefore further degrading respect.  Centralization seems an entirely plausible explanation if school boards have indeed centralized power and decision-making.  Empirically I have not seen any studies on the issue at all but my sense is that that centralization has indeed occurred.  

Kling's suggestion about school board centralization goes some way towards explaining the fall in School Principle status.

But what about the larger issue of overall school performance?

There are two further explanations for degradation which Kling does not mention but which I suspect are material causes.  

1)  The increasing foreign born, English as a second language (ESL) population

2)  The declining core of married biological parents.

Both of these trends substantially increase the demand on schools for non-educational services and capabilities, absorbing resources and forcing schools into functions that they cannot be as good at, i.e. substitute parenting.  

Both trends are indisputably true.  In 1950 the foreign born population in the US was 6.9%.  In 2020, it was nearly double at 13.7% of the population.  To the extent that foreign born, non-English speaking students demand more services (language instruction, assimilation instruction, increased security, etc.), it is a drain on school resources and a demand on teacher skills for which they are usually ill-prepared.  Both would drive down school performance and therefore reputation and social prestige of the School Principle.

I am confident that this is at least part of the explanation.  America's PISA (Programme for International Student Assessment) scores have always been middling in the developed world pack which is astonishing given the financial resources we pour into the sector compared to every other country.  

But if you break out the test scores by race (as a proxy for culture of origin and a very crude proxy for country of origin) then you find a Simpson's Paradox.  The US national average is in the middle of the pack among developed nations.  But every racial group in the US is at the top of the performance for its international corresponding groups.  White Americans score materially higher than every European nation.  Hispanic Americans score higher than every Latin American nation.  Asian Americans score higher than every Asian nation.  Black Americans score higher than every majority black nation.  

Clearly American teachers are doing something right.  The middling performance arises from the changing demographics of America, not necessarily the declining performance of teachers.  

Similarly, even if there were no change in demographics, the changes in American social norms has added further burdens and responsibilities on American schools which did not exist earlier and which divert resources from education to social policies.  As an example, in 1950, 5% of children were born out of wedlock.  In 2020, 39.8% were born out of wedlock.  It is a well established finding that children in stable, married families have dramatically better socio-economic outcomes across the board, including education.  

Schools used to serve a population of children who had greater stability and more focus on education accomplishments than they do now.  With the weakening of the family structure, schools have had to become more an institution of social policy than purely of education.  Yet a further erosion of status.

This is clearly seen within states.  Schools serving populations which are stable and married (and low foreign-born ESL) have materially higher test outcomes, and lower costs and better facilities.  They are not middling of the PISA pack.  We have something of a bi-modal distribution.  Our above average schools are way above average and our below average schools are way below average. 

Is the deterioration of public schools real?  Yes, conditionally.  Public schools in urban centers have probably become measurably worse on most socio-econometric measures in terms of both inputs and outcomes.  But public schools in non-urban centers have probably become measurably better on most socio-econometric measures in terms of both inputs and outcomes.  

Melding the two causal lists, my assessment as to causes of decline in public school performance would look something like:

Increasing percentage of the student body without a stable two biological married family structure

Increasing foreign born ESL student population

Consolidation of school districts (making them much more bureaucratic and insulated from parents)

Woke takeover of schools of education

Increased power of teachers’ unions

Better opportunities for gifted women to go into fields other than teaching

The drain of talented young women away from schools owing to better opportunities elsewhere was likely a real phenomenon but that change was pretty much effected in the 1960s-80s.  Hard to see that it is an explanation of any trends in the past twenty or thirty years.

Patrician Pathological Altruists

From Countries with Higher Levels of Gender Equality Show Larger National Sex Differences in Mathematics Anxiety and Relatively Lower Parental Mathematics Valuation for Girls by Gijsbert Stoet, et al.

This hurts some of the progressive memes.

Despite international advancements in gender equality across a variety of societal domains, the underrepresentation of girls and women in Science, Technology, Engineering, and Mathematics (STEM) related fields persists. In this study, we explored the possibility that the sex difference in mathematics anxiety contributes to this disparity. More specifically, we tested a number of predictions from the prominent gender stratification model, which is the leading psychological theory of cross-national patterns of sex differences in mathematics anxiety and performance. To this end, we analyzed data from 761,655 15-year old students across 68 nations who participated in the Programme for International Student Assessment (PISA). Most importantly and contra predictions, we showed that economically developed and more gender equal countries have a lower overall level of mathematics anxiety, and yet a larger national sex difference in mathematics anxiety relative to less developed countries. Further, although relatively more mothers work in STEM fields in more developed countries, these parents valued, on average, mathematical competence more in their sons than their daughters. The proportion of mothers working in STEM was unrelated to sex differences in mathematics anxiety or performance. We propose that the gender stratification model fails to account for these national patterns and that an alternative model is needed. In the discussion, we suggest how an interaction between socio-cultural values and sex-specific psychological traits can better explain these patterns. We also discuss implications for policies aiming to increase girls’ STEM participation.

So the most progressive nations have the highest degree of gender difference in mathematics anxiety compared to all other countries.

"We also discuss implications for policies aiming to increase girls’ STEM participation." So the Patrician Pathological Altruists strike once again, imposing solutions to solve problems that don't exist. They have decided that there should be proportional representation in the field regardless of what actual women actually want.

Trade-offs. Will they never go away?

From What the Prescription Drug Debate Gets Wrong by John Tierney.

The American pharmaceutical industry is the most innovative in the world and saves more lives than any other institution. So, of course, it is also the national villain.

In this autumn’s election, once again, voters say that one of the top issues—the top issue, in some polls—is lowering the price of prescription drugs. Politicians of both parties ritually denounce Big Pharma for profiteering. In his first press conference as president, Donald Trump accused drug companies of “getting away with murder,” and Bernie Sanders has called the industry’s greed a “public-health hazard to the American people.” A central plank in the “Better Deal” that Democrats are promising in the midterm elections is for the federal government to “negotiate” drug prices, and some progressives don’t even make that semantical pretense. They call for outright price controls, if not the “deprivatization” of the industry, on the grounds that Big Pharma is too powerful to be constrained by market forces.

[snip]

At one level, this is just political opportunism. Big Pharma is easy to resent because its products are so essential, and it’s easy to attack because it’s actually not so big. Of every dollar that Americans spend on health, only a dime goes for prescription drugs. The lion’s share of health spending goes to hospitals and people in the health-care professions, whose relatively high fees and salaries are largely responsible for Americans bearing the world’s highest health-care costs. But how many politicians want to go after doctors and nurses? What Democrat would dare arouse the ire of the health-care unions? Much easier to scapegoat the greedy drug companies.

The critics do get one thing right: the pharmaceutical industry is no paragon of free-market capitalism. Companies spend much of their time appeasing regulators instead of satisfying customers. The bureaucratic delays and complexities discourage innovation and competition, allowing some firms to profit by gaming the rules rather than developing new drugs. The system is so opaque and convoluted that both parties agree that it needs to be reformed.

[snip]

The obvious explanation, at least to the academics and journalists and politicians who have dominated the debate for decades, is that America lacks the nationalized health systems of other developed countries. Too much of the money goes into the coffers of drug companies, private insurers, advertisers, and other profiteers. To make health care affordable and spread the benefits equitably, America must emulate the systems in Europe and countries like Canada and Japan.

The conclusion sounds plausible until you take a closer look at the numbers. Yes, Americans do spend more, but they can afford more because their incomes are higher, and they’re getting more in return. Yes, the statistics are better in other countries, but it’s not because of their price-controlled drugs and national health services. If America adopted their policies, the gaps would only widen.

The gap in life expectancy is due to demographic and cultural differences, starting with the much greater ethnic and economic diversity in the United States. Fans of government health care like to point to Sweden and Norway, where people live three years longer than Americans, but that gap shrinks to one year if you consider just Minnesota, a state with relatively little poverty or racial diversity and lots of well-educated Scandinavian-Americans. The U.S. life expectancy at birth, 79 years, looks bad compared with Singapore’s 83 or Japan’s 84 (the world’s highest), but people from those countries live even longer in the United States. The life expectancy for Asian-Americans is at least 85, and some estimates put it at 87. Some of that added longevity is presumably due to the American health-care system, though it’s hard to quantify because so many variables affect life expectancy: genes, income, education, diet, exercise, lifestyle, and other cultural factors.

Indeed, this latter point is critical. It is the Simpson's Paradox which undermines almost every conclusion based on drawing comparisons between the US and any other OECD country.

When you compare American lifespan, morbidity, PISA scores in education, etc. the US comes in at the mid-range or upper third of the OECD average, rarely at the top. There are a lot of potential reasons for this.

The one that explains most or all of the gaps is the Simpson's Paradox arising from the fact that the US is not only a huge continental country, but because of its founding principles (treating people as individuals with God given rights) means that the US is multicultural to a degree virtually absent anywhere else (Singapore being a, literally, tiny, exception).

On all measures the Simpson's Paradox is that while the US as a whole performs in the upper middle of the OECD pact, when you look at the subsets of multiculturalism (such as race, ethnicity, culture, religion, class, etc.) the US outstrips virtually everyone on all virtually all measures.

Take the crudest matchings of race for continent, in other words European Americans compared to Europeans, African Americans compared to Africa, Asian Americans compared to Asia, etc. In virtually all measures the American groups far outperform the origin continent comparisons. For example, European Americans are longer lived, richer, have less crime, are better educated, healthier, etc. than any country in Europe. Than same with African Americans and Africa, Asian Americans and any country in Asia, Hispanics and any country in Latin America.

I elaborate on this in Simpson's Paradox and The Texas-Wisconsin Paradox and intergenerational income mobility and US Education: Expensive and ineffective? Not so fast.

The Simpson's Paradox applies beyond race/ethnicity/culture to religion as well as indicated in that last link on American education.

Healthcare provision remains an area of the economy desperately requiring improvement. We won't achieve that through the naive solutions which seem to be the primary choices advanced by policy makers. It is more complicated and we have to keep an eye on the trade-offs which are really in front-of us instead of assuming away costs and uncertainties.

There are reasons for optimism

McKinsey & Company have recently released a new report, The Economic Impact of the Achievement Gap in America's Schools. McKinsey is one of those organizations that generally invests a good deal of time, money and intellectual effort into producing insightful reports (for example, see Education Knowledge Highlights.) In this instance their primary conclusions are not particularly controversial: there is large variation in performance between schools and between school systems across the US, there are large variations in performance based on demographics, that the US education system is inefficient in terms of resources consumed for results produced compared to other countries, and that there is a high economic cost to the nation based on scholastic underperformance. Heroically, they attempt to quantify that cost (in the hundreds of billions up to the low trillions depending on assumptions and baselines chosen.)

While their numbers are certainly wrong, given the weakness of the underlying data, they at least provide a ballpark estimate. I am also reasonably sceptical of the comparability of the data they use for comparing US versus other countries (Program for International Student Assessment, PISA).

McKinsey's themes are limited, undramatic and straight-forward: better data is required, funding is an issue, minorities are ill-served and schools are important. They properly emphasize that there are reasons for optimism and that there are plenty of avenues to explore that hold out the prospect for material progress.

The weakness of the report, from my perspective, relates to the international comparisons as mentioned above; the disconnect between the supposed comparisons and the economic productivity of the US (if the US has been performing so dismally on education for the past 30 years, how is it that the US economy has performed so comparatively well versus the nominal educational superstars?); and finally, the dearth of recommendations (how do you optimize results within the constraints of the US circumstances?) The report is a worthwhile addition to the overall literacy and education literature. There were a couple of insights that you won't see anywhere else such as:

As a rule, schools in poor neighborhoods spend far less per pupil than schools in their nearby affluent suburbs. Since teacher salaries are one of the biggest components of district cost structures, affluent districts routinely outbid poorer ones for the best teaching talent (in addition to offering typically better working conditions and easier-to-teach children).

Breeding Leonardo

From Peter Farb's Humankind. I love unexpected historical surprises. Who would have guessed that this ever occurred.

Shortly after Leonardo da Vinci died in 1519 at the age of 67, his younger half-brother Bartolommeo set out to reproduce a living duplicate of the great painter, sculptor, engineer, and author. Since he and Leonardo were related, the father that Bartolommeo chose was himself. He chose as his wife a woman whose background was similar to that of Leonardo's mother; She was young and came of peasant stock, and had also grown up in the village of Vinci. The couple produced a son, Piero, who was then carefully reared in the same region of the Tuscan countryside, between Florence and Pisa, that had nurtured Leonardo. Little Piero soon displayed an artistic talent, and at the age of twelve he was taken to Florence, where he served as an apprentice to several leading artists, at least one of whom had worked with Leonardo. According to Giorgio Vasari, the leading art historian of the period, young Piero "made everyone marvel . . . and had made in five years of study that proficiency in art which others do not achieve save after length of life and great experience of many things." In fact, Piero was often referred to as the second Leonardo.

At the age of 23, however, Piero died of a fever and so it is impossible to predict with certainty what he might have gone on to achieve - though there is some indication in that Piero's works have often been attributed to the great Michelangelo. Nor is it possible to say positively how much of Piero's genius was due to heredity and how much to environment. Full brothers share, on average, fifty per cent of their genes, but Bartelemmeo and Leonardo were half-brothers and so would have had only about a quarter of their genes in common. Piero's mother and Leonardo's mother do not appear to have been related, but in the closely knit peasant village of Vinci it is quite possible that they had ancestors in common and thus shared genes. On the other hand, a strong environmental influence cannot be ruled out. The young Piero was undoubtedly aware of his acclaimed uncle; and certainly his father, Bartolommeo, provided every opportunity that money could buy for the boy to emulate him. But Bartolommeo's efforts to give the world a second Leonardo by providing a particular heredity and environment might, after all, have had little influence. Piero possibly was just another of the numerous talented Florentines of his time.

Lots of thoughts. How early people thought of themselves as breeding stock. Whether sheer expectation might have played a role. How amazing it is that we know this story half a millennium later. What the nature of the courtship might have been. How do you even open that conversation - "I want to marry you to breed a second Leonardo"? Finally - today's helicopter parents don't hold a candle to those of yore.

Give credit where credit is due.

Yes, but. From Achievement Gap Between Rich and Poor Public School Students Unchanged Over 50 Years
Spending four times more in real dollars per pupil doesn't compensate for low-quality teaching. by Ronald Bailey.

I empathize with the sentiment but I think it is also, to a degree, unfair. We have made an unconscious trade-off decisions which in part explains some of the data. From the article.

Half a century of trying hasn't closed one of schooling's most vexing achievement gaps. According to a new paper, the gap in educational achievement between public school students in the bottom 10th socioeconomic status (SES) percentile and those in the top 90th SES percentile has remained essentially unchanged over the last 50 years.

"In terms of learning, students at the 10th SES percentile remain some three to four years behind those in the 90th percentile," report a team of researchers led by the Stanford economist Eric Hanushek in their disheartening new National Bureau of Economic Research study, "The Unwavering SES Achievement Gap."

It would be one thing, the researchers note, that "if all achievement were rising, i.e., if a rising tide was lifting all boats." But that's not what's happening. Young adolescents' performance has risen over the past 50 years, but their scores drift downward once they reach high school. The upshot is that there has been no significant improvement in the overall education achievement scores of American high school student cohorts born since the 1950s.

The researchers draw upon data from four periodically administered assessments of U.S. student performance: the Program for International Student Assessment, the Trends in International Mathematics and Science Survey, and two versions of the National Assessment of Educational Progress. They then divvy up the student cohorts based their parental socioeconomic status.

The researchers calculate the standard deviation between the scores of each socioeconomic cohort to compare how far the average achievement scores of students clustered in the top 10 and 25 percent of SES percentiles are from the scores of those students aggregated into the bottom 10 and 25 percentiles of SES. A declining standard deviation would mean that the gaps between the cohorts' scores are closing. That is not what they find.

As they report at Education Next, the socioeconomic achievement gap among the 1950s birth cohorts is very large—about 1.0 standard deviations between those in the top and bottom deciles of the socioeconomic distribution (the 90–10 gap) and around 0.8 standard deviations between those in the top and bottom quartiles (the 75–25 gap). Measuring cohorts of students born since the 1950s, the SES gap closes by about 0.5 standard deviations for students under age 14. But those gains among young adolescents disappear almost entirely by the time students reach age 17.

The persistence of the SES gap remains when the researchers compare only white students over time, and they take into account such factors as the changing ethnic makeup of American school children.

The researchers note that these disappointing results occurred despite the fact that "overall school funding increased dramatically on a per pupil basis, quadrupling in real dollars between 1960 and 2015." In addition, pupil-teacher ratios declined from 22.3 in 1970 to 16.1 in 2014.

I cannot prove it but my suspicion is that part of the explanation is that we have had a prolonged period of immigration from countries with lower national IQ averages and lower education attainment countries to the US. This change of immigration policy occurred in the 1960s, matching the time frame being examined. We have gone from 5% foreign born to 15% foreign born in that time frame. That is for the population as a whole. I think the numbers are higher for the younger school grades.

If you accept that it is going to cost more to educate children in the English language, social norms of America, deal with any accumulated education deficits from wherever they originated, etc., then it makes sense the cost of education might rise while overall results remain flat.

In addition, the fact that Americans of respective broad cultural backgrounds, when compared to home origins (white Americans to Europe, Hispanics to Central and South America, Asian Americans to Asia) materially outperform their home countries' international PISA scores suggests that perhaps American education is doing a good job but it is dealing with a much more heterogeneous population by culture and by immigrant status.

It is easy to slap teachers around and lord knows there are plenty of instances of ineffective teachers and ineffective schools, but we should also acknowledge what to me appears to be a real accomplishment.

One thing I never see in these studies and which I would dearly like to know, is whether schools are effective at serving as mobility platforms. To do so, we would need to look at IQ as well as SES. My suspicion is that we do a pretty good job of finding and redirecting top quartile IQ talent from low quartile SES. Low SES status may have many vectors influencing school accomplishment but if high IQ talent from low quartile SES are actually accelerating past their circumstances, that would be great news. I suspect that is happening but I do not know it is.

Measurement is not the same as reality

Decades ago, I learned to be careful about indices ranking countries. Not only is the quality of data among countries highly variable, making the rankings questionable, but the indices themselves sometimes surprise you in unexpected ways.

My first eye-opening came in my teens or early twenties. Iowa was held up as an example of superior public education at least in part based on their very high SAT scores. Having been impressed by that evidence, some months later I was disillusioned to discover that the reality was more ordinary. Most Iowans took the comparable ACT test for college application purposes. Only those seeking to go to prestigious coastal universities bothered to take the SAT exam. As a consequence, the SAT takers were in no way representative of Iowans in general. They represented Ivy League aspirant Iowans, an entirely different group and a group, not surprisingly, with superior SAT scores.

In college, I did a lot of research using World Bank information. It was in that period when I came across some WB ranking of countries based on an index of something, let's say Liberty. I have no recollection of the details but this reconstructed example suffices to illustrate the issue. Looking at the rankings, there were a lot of anomalies and overall the index did not feel compatible with what I knew from having lived in some of those countries or having had friends that had.

For example Freedom of the Press as a variable in the index, might have shown the US doing poorly but Eastern European countries (then behind the Iron Curtain) doing well. Why? Because the index valued government funding of news reporting and therefore the US rated poorly on that whereas communist East Germany rated highly. You had to get three layers deep into the index to discover these anomalies.

We likely have something similar going on right now with PISA, the international comparison of educational outcomes. The last round or two have shown high results for China but also extraordinarily high results for Shanghai as region within China. Shanghai is a huge city of near 25 million people. Statistically it is virtually impossible for such a large population to score so far from the mean of the country as a whole. Almost certainly these high results are for some subset of Shanghai students from the best schools. We don't know the truth yet. Data reported by China says one thing, the Chinese like it that Shanghai students are the best in the world, and statistical probability points to a different conclusion. Time will tell.

All this brought to mind by this article, Questions About the Global Gender Gap by Elaine Schwartz.

The World Economic Forum (WEF) tells us that Iceland leads the world in gender equality while the women in the Philippines are #7.

[snip]

To rank gender equity, the WEF uses 14 indicators. When they measure economic participation and opportunity, wage equality for similar work is one variable. Their score for education is partially based on primary and secondary enrollment. In the health component, life expectancy counts. And with political empowerment, one area is women in government.

You can see that the metrics are pretty standard and predictable.

But some of the scores appeared surprising. Do note that Iceland’s health and survival rank is 104 out of 144 countries:

[snip]

Meanwhile, the Philippines were #1…the best possible:

[snip]

However, the life expectancy of a little girl born today in Iceland is 85.3 years and a boy, 80.9. In the Philippines, the estimate is 72.9 years for women and Filipino men, 65.7.

Why then is Iceland ranked so far below the Philippines?

The answer is the WEF methodology. Because the scores are based on a male female comparison within each country, a Filipino woman fares beautifully in terms of her male counterpart.

Similarly with education, a country like Iceland has close to 100% secondary education enrollment for males and females. In the Philippines, the numbers are closer to 60%. And yet both have the same score because we are recording a male female ratio.

To put it more bluntly, because Filipinos die younger but more equally (between the sexes), the WEF views them as a more gender equal country.

Well, perhaps, but I don't think that most people would see that as a positive outcome meriting a #1 ranking.

The more time you have to invest understanding anomalies in an index, the less valuable you often discover it to be.

Data Talks

New article shows that PISA (an international test designed to measure academic achievement in reading, math, and science) is mostly a measure of general intelligence. After controlling for g, specific domains have low reliability and little meaning.https://t.co/gNOJI3iUhE pic.twitter.com/QA6eXoGada

— Russell T. Warne 🇺🇸🇨🇱 (@Russwarne) April 11, 2022

 

I see wonderful things

Great picture from Pisa cathedral pic.twitter.com/x5yuH8lZ5c

— Stained Glass Zealot † (@glass_zealot) March 17, 2022