What Schooling Can and Can't Do
I have been in education, in one form or another, for all of my life. I have also studied it somewhat extensively from empirical and political theoretical standpoints. As such, I might have some insight into the functions of schooling and what we ought to do with it. Perhaps no one idea I put forward in the following will be terribly original, but I hope that the synthesis of ideas that I create will be original or, more importantly, useful.
What is schooling for? Why do we demand so much schooling for our people? An obvious answer is teaching people basic skills, like how to read. As a spoiler alert, I think this is basically the correct answer. However, ask anyone in the business, and the answer you get will almost certainly be very high minded: to teach critical thinking! To expand horizons! To develop the child into the culture! Etc. In the following, I want to deal with the “teaching a student to think,” or having the student “learn how to learn,” or to “think critically” rationale. These are all pretty vague as they are, but hang around schools long enough and you will hear these phrases. Just going by my personal impressions, this is probably the number one idea that educators have about what they’re doing for their students.
How can we even test these kinds of claims? How do we know when a person has learned to learn? Knows how to think? Can think critically? And if we have measures for these traits, have we applied them to measure the difference education makes? After all, these kinds of claims mean that students forgetting literally all of the content they “learned” in their classes is hardly a mark against the teaching they received: the students did not learn history or biology, they learned how to think. (This is doubly convenient since most kids probably learn to think better as they grow up anyway; we may attribute the work of nature taking its course to the school room instead and proclaim our teaching a success, whilst having really achieved nothing.) So we cannot take any measures suggesting that people barely learn anything in schools against their thinking abilities. So how are we supposed to test this argument?
IQ Gains and Schooling: Reviewing the Causal Literature
There is at least one literature that suggests cognitive improvement from education, and that is measured in terms of IQ. In fact, much of the work is quasi-experimental in nature. (Ritchie and Tucker-Drob, 2018) This work shows that a year of education may cause anywhere from a 1 to 5 point increase in IQ per year of education. That’s far from nothing! If that is true, then 12 years of education causes a change from 70 IQ to 82 IQ, if not to 130! A mere four years—your typical time in high school—can cause a 4 to 20 point increase, potentially going from an IQ of 90 to 94 or even 110! At the smallest estimate, fifteen years of education produces a one standard deviation increase in IQ. In fact, the actual gains may be even five times larger. If this is true, then even the trillion dollars a year we are spending on education is probably worth it, for results so large.
This is exactly what raises a red flag for me. These results are simply too large to be real. Simply put, raising IQ has generally been considered incredibly difficult (see Jaeggi et al., 2008; the authors claim to have found a method for training working memory to raise fluid IQ, which I will discuss later on. They describe the consensus that existed at the time with regards to increasing fluid IQ, including much of the literature suggesting the difficulty of such a feat. Sternberg, 2008, shows how surprising their alleged discovery is in his response to the study, which he titled: “Increasing Fluid Intelligence is Possible After All.” Emphasis added.). The idea that we found the magic bullet in schooling—even schooling that does not seem to communicate specific knowledge—seems implausible to me. Is the methodology of the studies cited in Ritchie and Tucker-Drob valid?
There are three strands of evidence included in the meta-analysis. Methodologically, two of these are severely flawed, and the third is mildly flawed but plausible.
One involves the implementation of education on one side or another of an age cut off point. This is an extremely flawed methodology. It assumes that time of birth is basically random, so people on one side of an age cut off will be statistically similar as those on the other side of the cut off. But is a person’s time of birth truly random? I hate to give it to the astrologers, but time of birth indicates something about when the parents were getting busy. Parents who get busy after getting very drunk on a holiday are probably different from parents who get busy in some half-assed month like April that doesn’t have any holidays. Any causal inference that depends on time of birth being truly random is open to challenge. So I concluded just thinking from an armchair. To confirm my armchair speculation, there is an entire Wiki page dedicated to “Influence of Season Birth in Humans.” We need not identify the precise causes of the various relationships noted in this page in order to say that birth month is not random. Any causal inference based on that premise is flawed.[i]
The second strand of evidence involves controlling for IQ before taking on so many years of schooling. IQ tests are taken once, in youth, and then again in midlife. However, this methodology effectively assumes that differences of earlier and later IQ must be produced by schooling somehow, on the assumption that IQ is otherwise constant throughout the lifespan. And it is true that roughly half of individual differences in IQ remain constant throughout the lifespan (Deary, 2014); however, that implies that half of differences in IQ do not remain constant throughout the lifespan. The fact that there is a change “produced” by education may not indicate causality from education; instead, it may indicate a late bloomer, dumb as a kid, who found their stride later on for other reasons. These late bloomers may still get more education. On the other hand, there are probably also burnouts—people who have a high IQ when younger and who become dumber as they get older. These burnouts would probably also get less education. With these two combined results, we can reasonably expect IQ to cause more schooling rather than the reverse, even with a control from youthful IQ scores.
The final type of evidence is instrumental variable analysis, involving staggered policy implementation in a given country. While the first two strands of evidence are terrible, this strand of evidence is much more persuasive. Ritchie and Tucker-Drob give an example of this kind of study. That study describes introduction of an increase in compulsory schooling years combined with funding for new schools and such when municipalities submitted their plans for implementing the new law. They describe the following methodology:
Each separate municipality was able to introduce the full compulsory schooling reform after local officials submitted a reform plan to a national committee, which, on approval of the plan, provided national funds to finance the creation of the new middle schools and the extension of compulsory schooling. The timing of the reform in different municipalities was therefore not explicitly randomized, but earlier studies of the reform have not been able to uncover strong correlations between observable characteristics of the municipalities and the timing of the reform. Extensive checks performed as part of our analysis fail to uncover evidence that implementation of the reform was not exogenous to our outcomes of interest (education/IQ). (Brinch and Galloway, 2012. References omitted.)
I find this persuasive. Strictly, this is not randomized; however, we would expect correlations to other observable facts about the municipalities of some kind if something non-random were going on in plan submission. Therefore, we might think that this is effectively quasi-randomized. Can we count on such quality randomization in the other instrumental variable studies? I do not know the answer to this; I do not have access to all of the other studies. I will work on the assumption that randomization is successful, but that is just an assumption. If we are being honest, I doubt that assumption holds. But let us move on, since I cannot assess the matter first hand.
Other than this, my main methodological objections are either minor or else major but still show that schooling matters, at least if we assume that the other studies based on instrumental variable analysis are of a similar quality to Brinch and Galloway. In instrumental variable analysis, we need to be sure that the instrument is genuinely related to the casual variable. Here, the suggestion is that the policy reform increases the number of years of education. Yet, as Brinch and Galloway note, there were already educational institutions in place for the years in question. The idea is that the extra funding in the new policy would make attendance at school easier. Are we sure that there were genuine increases in attendance (as opposed merely to enrollment)? The data is based on years of education completed (maximum of 12, since the IQ test was taken at the age of 19 for the male population in Norway). If a person completed 12 years of education, did they complete 12 years of education or did they “complete” 12 years of education? Since the increase in compulsory attendance was from 7 to 9, we might wonder if it was 9 that they “completed” instead, but the point remains.[ii]
Another assumption of instrumental variable analysis is that the instrument affects the outcome only through the proposed causal variable. However, it seems possible that the added funding for new schools could have differently influenced municipalities in other relevant policies. After all, money is fungible; if the national government suddenly gives municipalities a bunch of money to further their educational priorities, they might then have money freed up for other priorities. Since this was only a decade or so after WWII, Norway probably still needed some rebuilding, and there may well have been other programs at the municipal level that could have influenced IQ once implemented. If these programs influenced nutrition or health, they could have boosted IQ as a result. The instrumental variable may have influenced more than just schooling. This is made less likely if observable correlates between municipalities and their accepting the new mandate are small or non-existent, but it is possible that some new policies were constant upon accepting the new funding. Correlation measures variation within data, not constants. If all of the municipalities took on some similar set of new policies in response to new money coming in, those could have influenced IQ just as well as schooling.
These objections are less than compelling, and I must admit that if this were my study, I would not be persuaded by them. Things seem effectively randomized; the effect of new, compulsory education is surely not going to be to reduce the years of education; and the idea that any other, new policies municipalities might take up in response to have more money to spend on education just seems unlikely to really influence IQ so quickly after acceptance of the new schooling mandate. My final objection is not to this particular study, but to studies depending on a similar methodology in the American context. Upon reviewing that literature, Stephens Jr and Yang, 2014, show that the increased years of schooling are confounded with other variables, including increases in the quality of schooling. The increase in school quality, independent of the number of years of schooling, could be what stands behind IQ increases from schooling. To me, this seems like the strongest objection to Brinch and Galloway: the mandate does not only increase the number of years of schooling, but also literally gives municipalities new funds to build new schools. The change in physical faculties or any of the other changes attendant upon the building of those new faculties could be the real driver of IQ increases. Thus, the instrument in this case quite explicitly does not only influence the number of years of education; it effectively influences the quality of education, as well. The causal effect of the instrumental variable could work through either of these routes. Indeed, I strongly suspect this is the case, since it is consistent with the American evidence. Perhaps quality matters more than quantity. However, while this is the strongest objection I can make to Brinch and Galloway, it remains clear that these results are produced by change in schooling, not changes in things that are not schooling. Thus, schooling remains important in this objection, albeit in a slightly different way.
Only one of the strands of evidence suggesting that more schooling produces IQ gains seems legitimate.[iii] However, that strand seems relatively strong. It is not perfect, but it definitely moves my priors in favor of schooling having a positive effect on education. In the context of other literature, I strongly suspect that it is schooling quality, rather than years of schooling, that primarily contributes to the effect in question; however, even this objection still leaves me with a higher assessment of the probability that schooling changes IQ for the better.
The question becomes: what is the mechanism by which these effects occur? It could be that teachers successfully teach students to think, which shows up in increased IQ. Or, it could be that going to school changes anything else in the life of the student. Is there any evidence that it is teaching “the ability to think” or “to think critically” that causes this presumed causal relationship?
Two plausible interpretations of the casual results suggest that this may not be the case. First: test-retest effects seem to exist in IQ tests (see Scharfen et al., 2018). Now, the instrumental variable analysis evidence does not depend upon retesting per se; however, so many teachers complain about “teaching to the test.” It seems highly plausible that the IQ gains produced by increased schooling could just be a result of students having a better idea of how to take a test, including an IQ test, in general. Even if teachers do not “teach to the test,” traditional methods of schooling would obviously tend to make people better at testing. Instead of an increase in actual cognitive ability, schooling could just produce an increase testing ability. If this is the reason for increases in IQ, then the increases in IQ are illusory; instead of increases in cognitive ability, we are instead observing increases in the ability to take a test. This is the cynical interpretation, and I have to imagine it is at least partially true.
Nonetheless, there is another possibility, if we believe that the IQ gains are produced by actual learning in schools. In general, evidence for increasing cognitive ability is rare. However, it is possible that improvements in working memory can cause increases in fluid IQ. (This was the argument of Jaeggi et al., 2008.) This is based in evidence that fluid IQ is based, in large part, on working memory. By this logic, increasing working memory should increase fluid IQ. Now, the evidence for this proposition (and for the proposition that we can improve working memory at all) is marred by various forms of bias (of measurement, publication, etc; see Rodas et al., 2024). However, unlike many other theories of IQ increase, this theory is well-grounded theoretically and empirically, except that the direct evidence for a connection is lacking. While the positive evidence in favor of the theory is marred and of poor quality, that does not mean we have proven that there is no connection: that means our current evidence, one way or another, just does not point in any particular direction. Our priors should remain more or less unchanged by the existing evidence, but I would suggest that our priors should favor this specific argument about IQ increases, at least as compared to other arguments about how to increase IQ (insofar as they represent real gains in cognitive ability).
Why is this relevant? There is evidence that reading ability is related to working memory. (See Peng et al., 2018) Added years of education might directly improve reading ability—even Caplan, probably the staunchest critic of schooling short of an unschooler, grants this (2018, pp 40-49)—which would, in turn, improve working memory.[iv] More broadly, a variety of differences between groups in IQ can be explained via literacy rates, as can the increases over time and the fact that increases over time primarily accrue on the lower end of the IQ spectrum rather than the upper end (Marks, 2010). A reasonable theory is that, instead of IQ increasing because we teach students to think, IQ increases because we teach students to read (first to read the test, and then to understand the complex ideas on the test.) As Marks puts it, “an IQ test requires a high level of literacy because it entails the understanding and application of complex information. An illiterate person is therefore ill equipped to complete an IQ test.”[v] If this is true, then we should focus less on teaching students to think and more on teaching them to read.
Teaching Students to Think: Do the Mechanisms Make Sense?
When we have two theories, we need to look at the evidence that would distinguish one from the other. Is there evidence supporting the “teaching students to think improves IQ” theory, as against the “teaching students to read improves IQ” theory?
To my knowledge, there is not. This is what a substantial proportion of Bryan Caplan’s The Case Against Education is about, as he explores the transfer of learning literature and discovers very little there to justify that idea that schools make a big difference.[vi]
Transfer of learning involves three different kinds of transfer to be properly worthwhile in the “teaching students to think” kind of way: the learning 1. transfers outside of the classroom, 2. a significant amount of time after class is over, 3. to subject matter at least slightly different than what was presented in the classroom. In other words, we remember the material when we are going about our daily lives outside of school, and can think about it in problems different from those of the classroom.[vii] The most impressive result from the literature involves statistical learning by relatively elite (University of Michigan) students after a semester of statistics, while answering problems that were written to be easy (though different from the original statistical material). Of four questions, two involved no gains in quality of statistical reasoning, but two involved some gains—from 16% to 37% of the respondents giving a statistical answer on one and from 50% to 70% on the other, between the beginning and end of the semester. These modest gains were measured at the very last week of class. Thus, it is arguable whether the second requirement for useful transfer even applies here. And yet, most students did not improve. The other studies of transfer are even less impressive. (Caplan, 50-59) At most, people in classrooms learn what they are directly taught: they do not learn broad skills of critical thinking, or anything else that requires transfer. If there are IQ gains produced by more schooling, they probably do not come from teaching students to think.
Simply put, knowledge does not transfer; the only knowledge that we can really count on students learning is that which they directly use in their lives, if they even learn that. “Teaching students to think” is not a good justification for current pedagogical methods, certainly not as they currently exist.
Schooling, Obedience, and Civilization
If the function of schooling as it exists is not this high minded ideal, then what is it? It is possible that we have just made a mistake as a society; after all, the idea that a person could learn “how to learn” by regularly learning is intuitive enough. The difficulty of raising IQ is actually super counterintuitive; why wouldn’t training the mind on more intellectual problems make a person smarter? Even Karate Kid makes the mistake of assuming that transfer of learning will rather naturally happen, and we make other mistakes based on the same premise.[viii] In my judgment, it is perfectly reasonable to assume that most people genuinely do believe in this high-minded ideal, and are simply mistaken. This is the motivation behind our current pedagogy as we find it in the hearts of the public. However, there is another possibility, and I find it instructive to grapple with that possibility.
Instead of broad learning, Simler and Hanson suggest that schooling is about discipline and, as they put it, becoming civilized, including indoctrination into the ideology of society (e.g., patriotism). School is “a mixed bag. Schools help prepare us for the modern workplace and perhaps for society at large. But in order to do that, they have to break our forager spirits and train us to submit to our place in a modern hierarchy. And while there are many social and economic benefits to this enterprise, one of the first casualties is learning.” (Simler and Hanson, 240) In other words, there is a mismatch between hunter gatherer freedom and the modern need for discipline. The modern school, according to Simler and Hanson, corrects that mismatch.
This argument suggests that the point of schooling is the so called “hidden curriculum.” Thus, according to Bowles and Gintis,
We must consider schools in the light of the social relationships of economic life.... We suggest that major aspects of educational organization replicate the relationships of dominance and subordinancy in the economic sphere. The correspondence between the social relation of schooling and work accounts for the ability of the educational system to produce an amenable and fragmented labor force. The experience of schooling, and not merely the content of formal learning, is central to this process. (Bowles and Gintis, 125)
As Bowles and Gintis show, those who do better in school also have a variety of traits that are much better for workers, and they do not have a variety of traits that are bad for workers. Simply put, people who get high grades are submissive; people who get low grades are not. Submissiveness is more important than actual ability. (Ibid., 131-140) The point is not the pointless education: it is teaching students to obey pointless commands, so that they can go into jobs with an expectation of very little autonomy, and so they can be civilized by being disciplined. For Bowles and Gintis, this is the civilization process required by the capitalist economy, while Simler and Hanson take it to be the civilizing process demanded by civilization as such. In both cases, the assumption is that pointless schooling teaches people obedience and willingness to accept hierarchy, in the workplace or elsewhere.
Ideally, we would experimentally test this by randomizing who goes to school and who does not, and then see if the kids who went to school end up more disciplined for the workplace than kids who did not. We do not have such randomized data. Therefore, we should instead take a close look and see if the mechanism actually makes sense in terms of who gets sorted into what kind of work. If students are undisciplined in school and proceed to work on the factory floor, then the theory that they need school discipline to learn how to do factory work becomes untenable.
Enter Paul Willis' Leaning to Labour. Willis' work was a response to Capitalist Schooling in America. His critique of the work was simple: Bowles and Gintis failed to demonstrate that the hidden curriculum actually had an influence on working class youths. In Willis' analysis, what is relevant is not just the content of the hidden curriculum, but also how the youths react to that curriculum. Working class youths resist the hidden curriculum. Therefore, working class youths are not sorted into working class work via the hidden curriculum. The opposite is true: working class youths take on their difficult, manual work precisely because of how undisciplined they are in school. Their discipline on the factory floor was very clearly not made possible by their discipline in school. In fact, many of the youths held factory jobs while in school. The idea that school civilized these students, broke their spirits and made them capable of factory work in the first place is laughable given their resistance to school authority at the very time that they worked factory jobs. But the schools did contribute something in Willis' analysis: it legitimated their place as working class citizens to them and helped them believe that their work was honest, decent work, unlike the bourgeois upper class types who make their living by slavish servility in school. In other words, Willis' working class youths did not become capable of factory work because of school: they became capable of factory work because they positively assessed it compared to school and the kinds of work schooling could lead to. Many believed that those who did well in school trained themselves up to be dishonest scum, and that working with their hands was true, honest work. Insofar as school contributed to the willingness to labor, its contribution was to morally legitimate manual labor compared to intellectual labor.[ix]
This fits more broadly into the transfer of learning literature. Simply put, why would students learn obedience in school and then transfer it outside of school? Surely they can see that school is not the same as the rest of the world. Whether they choose to resist or obey in school, they can make a different choice outside of school. What comes to bear is not whether they learn to obey in school: it is whether they are the kind of person to obey or resist authority, wherever they are. Schooling may distinguish between the disciplined and the undisciplined, but it does not cause the difference. As with intelligence, the disciplined and the undisciplined are such going in. Whatever specific skills—washing hands, cleaning clothes, and so on—might be taught by schools to civilize people, the idea that school breaks the spirit to make adult work possible is not plausible.
The school clearly produces certain moral standards in the minds of the students. The working-class youths in Learning to Labour adopt their moral stance of being honest workers as opposed to the dishonest upper class types as a response to schooling and its demands for submission. Minimally, the work of “civilizing” students is a lot more complicated than just introducing them to schools that break their spirits.[x]
Simler and Hanson cited one work especially to make their case: Eugen Weber’s Peasants into Frenchman: The Modernization of Rural France, 1870-1914. A closer reading of the history Weber presents shows that Weber’s work does not support the “civilization” thesis nearly so well as Simler and Hanson take it to. Weber pays special attention to the development and effects of schooling in rural France in chapter eighteen of his work, “Civilizing in Earnest: Schools and Schooling.” (303-338) Without a doubt, schools have some influence on civilizing the rural French public in this history, but it is not such a one-to-one relationship as the hidden curriculum thesis would suggest.
Instead, the tale Weber tells is a tale of schools failing to bring “civilization” until the peasants bought into the need for it. Before then, the schools made no leeway; only after the peasants accepted the need for schooling (with the values and learning it brought) did the schools have significant efficacy. “It was only when what the schools taught made sense that they became important to those they had to teach. It was only when what the schools said became relevant to recently created needs and demands that people listened to them; and listening, also heeded the rest of their offerings. People went to school not because school was offered or imposed, but because it was useful. The world had to change before this came about.” (303; emphasis added.) Thus, in the old schools—putting aside the lack of attendance that Weber notes throughout the chapter (see e.g., 308)—the learning was only surface deep. Thus from Brittany, a report reads: “no child can give an account of what he has read [in French] or translate it into Breton; hence there is no proof that anything is understood.” (Quoted at 306) The patriotism that Simler and Hanson discuss also went unnoticed: when did children even hear the French language? (310-312) Indeed, teachers themselves were often ignorant and less than morally praiseworthy people. Of the subjects that that they allegedly needed to learn to teach their students, they typically mastered only “arithmetic and land-surveying… recognizing, as did villagers and schoolchildren, that they related to practical needs.” (317) The peasants did not accept the “civilization” brought by schools because that civilization had no value in rural France. What was valuable was work: children would leave school at the earliest possible age to work (assuming they ever went to school), and “once they had gone to work, learning was out of the question: they were at their tasks ‘by candlelight in winter, and in summer from one twilight to the other.’ There was no time or energy for anything else.” (320-21)
The essential reforms were those of Jules Ferry in the 1880s, which massively expanded the reach of schooling in rural France (308-309). Even with these, however, it is clear that the schools would have been rejected if not for the changes in the surrounding society that made those schools valuable and useful to the peasants whose children attended them. Thus one report, after noting the difficulties with producing rural attendance at school, nonetheless argues: “the remedy to this state of things lies in public opinion. Even the most ignorant portion of the masses begins to understand that instruction is useful to all [and not just to their betters]. Country people know now that reading, writing, and arithmetic are means of rising in the world.” (Quoted at 325) As Weber notes, this was the “fundamental cause” of the “indifference” to schooling that existed in rural France: while the urban poor might use some of the skills taught in the old parish schools, the rural poor could not use them, and found schooling a waste compared to whatever else the child could be doing. “School was perceived as useless and what it taught had little relation to local life and needs.” (326) While some French reformers might have hoped that the peasants would sacrifice material concerns in order to go to school, it was instead “when the school mobilized those interests that men began to care.” (326) It was when the rural poor began to recognize that their learning could be useful, to get them work and let them rise beyond being a peasant, or to let them carry out their own immediate needs without depending on others, that they started to respect schooling and make sacrifices for it (327-29).
It was only then, after the peasants had bought into the utility of schooling, that schools were able to build on this and reform their habits and feelings. Now it was recognized that “the only escape” from brutality “was education, which taught order, cleanliness, efficiency, success, and civilization.” (329. Emphasis in original.) Children learned “how to greet strangers, how to knock on doors, how to behave in decent company.” (330) It was here, as well, that children were taught patriotism for France (331-336). They were “to understand and appreciate the beneficence of the government.” (Quoted at 331) More broadly, “schools taught potent lessons of morality focused on duty, effort, and seriousness of purpose.” (331) Schools accomplished this, according to Weber, but only after they had already produced buy in from the peasant populations.
Thus, Weber’s work does not suggest that schools simply impose civilization on people who would otherwise be uncivilized. These people have to agree that civilization is needed and good before the schools can succeed in their task. Once the peasants agreed, then schools could take on the process of civilization, because the peasants agreed with what the schools were attempting to do and did not resist. Far from resisting, they pushed the effort forward themselves, in their own persons and in their children.
Weber describes what happened in a way that I find truly extraordinary:
If we are prepared to set up categories with well-drawn limits, society educates and school instructs. The school imparts particular kinds of learned knowledge, society inculcates the conclusions of experience assimilated over a span of time. But such a view, applicable to specific skills and subjects, has to be altered when the instruction offered by the school directs itself to realms that are at variance with social education (as in the case of language or measures), or that social education ignores (as in the case of patriotism). In other words, the schools provide a complementary, even a counter-education, because the education of the local society does not coincide with that needed to create a national one. This is where schooling becomes a major agent of acculturation: shaping individuals to fit into societies and cultures broader than their own, and persuading them that these broader realms are their own, as much as the pays they really know and more so. (330-31)
My only objection to this is suggesting that the successful use of schooling was a “counter-education.” Based on the evidence that Weber cites, when the schools offered a counter-education, they effectively offered no education for lack of buy in. Only when the peasants bought into the education did it become both useful and “complementary.” At that point, the schools could use the utility the peasants saw in the education as a basis for expanding peasant notions about the society in which they lived, making them more attached to France and its customs than their locality. And in this, the schools did not offer a counter-education; instead, it taught peasants that they should learn the specific materials before them because it could make them rich. Indeed, patriotism itself was advocated for, not merely via descriptions of heroics, but also because “the fatherland was a source of funds for road repairs, subsidies, school scholarships, and police protection against thieves—‘one great family of which we are all a part, and which we must defend always.’” (332)
The French, in every region, were all a part of “one great family.” It is clear from Weber’s evidence that this family was not meant to replace local affections. Bruno’s Tour de France made clear to its millions of readers “French patriotism was a natural complement to their own: ‘France is a garden, the provinces are the flowers in it.’” (335) This makes it perfectly clear: Weber is right that “society educates and school instructs.” All that must be added is that the instructions of school will be resisted and washed away until they are consistent with the education of society. Once that consistency exists, then schools can successfully expand the education of society, and show how pre-existing social values and beliefs are consistent with other, perhaps broader concerns, while also instructing in the specific habits and subjects that will allow for the execution of existing social values and beliefs. Weber’s history shows that schooling can successfully 1. Instruct in specific habits and knowledge that further pre-existing values and goals, and 2. Point out how pre-existing social values and beliefs are related to broader values and beliefs, and thus help populations see possibilities they otherwise would not have seen. In the instances where Weber says there was a counter-education—in language and measures—it was a counter-education in terms of very specific mores, mores that were contrary to the broader French values that the peasants now accepted. The “counter-education” was a part of the complimentary instruction, and acceptable as such. Schools cannot impose civilization: they can instruct in civilization once a population is open to it and already striving for it. Then, schools become the means for people to strive towards a civilization that they already want.
Thus, in general, schooling can teach something like discipline—if there is buy in for discipline. Schooling can teach something like a language—if there is buy in for learning the language. Schooling can promote patriotism—if that patriotism is connected, in the mind of the learner, to the needs and wants that they actually have. With buy in, much can be accomplished, albeit over decades; without buy in, nothing can be accomplished, and we will be lucky not to produce reaction.
Conclusion
I would suggest that this is why the transfer of learning results are so dismal. Do people genuinely buy in to their school learning? Do they really believe in the schooling itself? I think that obviously is not the case. Instead, students buy into getting a diploma, a certification. The learning is not a part of any great goal; it is not a part of who they are; it does not really matter to them. What matters is getting their degree so they can move on. Without buy in, our schools do not produce results—except for the results that everyone knows matter: basic reading, writing, and math skills (Caplan, 40-49).
The current schools probably accomplish very little. The increase in IQ they seem to produce[xi] is likely a result of teaching to the test, and even if it is the result of genuine learning, it is likely that the learning is only literacy, not the broader thinking skills that educators claim to teach. Based on Weber’s history, I suggest that this is because Americans lack buy in for learning. American anti-intellectualism has been long remarked; it is a part of American culture to be suspicious of book learning.[xii] One casualty is the efficacy of our school system. Thus, the main value of schooling in America is indeed signaling—separating the intelligent and submissive from the stupid and the unsubmissive.
So signaling is the primary function of schooling in America in real practice—but why are we not more eager to recognize that? Well, one obvious reason is because that sucks. But why does that suck? Because we are eager to believe in the “teaching kids to think,” transfer of learning based tale. Not only eager; it is a super intuitive tale. The signaling tale, compared to the “teaching kids of think” tale, is kind of crappy, and not quite as obvious.
But this refusal to change ideas comes at a high price. The fact is, we could probably do something about this situation. The historical record suggests that the value of education could be more than just signaling and basic cognitive skills. It could be greater than it currently is. It taught the French patriotism! However, so long as we hold on to the myth that we are currently accomplishing wonderful things in education—things that really are not in accord with what American culture genuinely values—we will accomplish nothing. We will ensure that book learning is just book learning, that learning in the classroom is considered largely irrelevant to life outside the classroom.
What can be done? How can schools become more than they currently are, and become genuinely valuable for more than just sorting people for business?
Well, maybe I can take that on in a future post.
Works Cited:
Bavelier, Daphne, C. Shawn Green, Alexandre Pouget, and Paul Schrater. 2012. “Brain Plasticity Through the Life Span: Learning to Learn and Action Video Games.” Annual Review of Neuroscience. 35: 391-416.
Bowles, Samuel, and Herbert Gintis. 1976. Schooling in Capitalist America: Educational Reform and the Contradictions of Economic Life. Basic Books.
Caplan, Bryan. 2018. The Case Against Education: Why the Education System Is A Waste of Time and Money. Princeton University Press.
Caplan, Bryan. 2012. Selfish Reasons to Have More Kids. Princeton University Press.
Case, Anne and Angus Deaton. 2017. “Mortality and Morbidity in the 21st Century.” Brookings Papers on Economic Activity. 397-439.
Dale, Gillian and C. Shawn Green. 2016 “Video Games and Cognitive Performance.” in The Video Game Debate: Unravelling the Physical, Social, and Psychological Effects of Digital Games. Ed. Rachel Kowert and Thorsten Quandt. Routledge: New York. 131-153.
Deary, Ian J. 2014. “The Stability of Intelligence From Childhood to Old Age.” Current Directions in Psychological Science. 23:4.
Haier, Richard. 2014. “Increased Intelligence is a Myth (So Far).” Frontiers in Systems Neuroscience. 8:34.
Harris, Judith. 2009. The Nurture Assumption: Why Children Turn Out the Way They Do. Free Press.
Jaeggi, Susanne M, Martin Buschkuehl, John Jonides, and Walter J. Perrig. 2008. “Improving Fluid Itelligence with Training on Working Memory.” PNAS. 105:19. 6829-6833.
Marks, David F. 2010. “IQ Variations Across Time, Race, and Nationality: An Artifact of Differences in Literacy Skills.” Psychological Reports. 106:3. 643-64.
Olsen, Kirsten. 2009. Wounded by School: Recapturing the Joy in Learning and Standing Up to Old School Culture. Teachers College Press.
Peng, Peng, CuiCui Wang, Shan Li, William Dardick, Marcia Barnes, Wei Wang, H. Lee Swanson, and Sha Tao. 2018. “A Meta-Analysis on the Relation Between Reading and Working Memory.” Psychological Bulletin. 144:1. 48-76.
Ritchie, Stuart J, and Elliot M. Tucker-Drob. 2018. “How Much Does Education Improve Intelligence? A Meta-Analysis.” Psychological Science. 29:8. 1359-1369.
Rodas, Jose A., Afroditi A. Asimakopoulou, and Ciara M. Greene. 2024. “Can We Enhance Working Memory? Bias and Effectiveness in Cognitive Training Studies.” Psychonomic Bulletin & Review.
Scharfen, Jana, Judith Marie Peters, and Heinz Holling. 2018. “Retest Effects in Cognitive Ability Tests: A Meta-Analysis.” Intelligence. 67:March/April. 44-66.
Schmidt, Emily. 2022. “Reading the Numbers: 130 Million American Adults Have Low Literacy Skills, But Funding Differs Drastically by State.” APM Research Lab.
Simler, Kevin, and Robin Hanson. 2018. The Elephant in the Brain: Hidden Motives in Everyday Life. Oxford University Press.
Sternberg, Robert J. 2008. “Increasing Fluid Intelligence is Possible After All.” PNAS. 105:19. 6971-6972.
Stevens Jr., Melvin, and Dou-Yan Yang. 2014. “Compulsory Education and the Benefits of Schooling.” American Economic Review. 104:6. 1777-92.
Waris, Otto, Susanne M. Jaeggi, Aaron R. Seitz, Minna Lehtonen, Anna Soveri, Karolina M. Lukaski, Ulrika Soderstrom, Russell C. Hoffing, and Matti Laine. 2019. “Video Gaming and Working Memory: A Large-Scale Cross-Sectional Correlative Study.” Computers in Human Behavior. 97. 94-103.
Weber, Eugen. 1976. Peasants Into Frenchmen: The Modernization of Rural France, 1870-1914. Stanford University Press.
Willis, Paul. 1981. Learning to Labour: How Working Class Kids Get Working Class Jobs. Columbia University Press.
[i] Yes, I cited Wikipedia, because this should be an incredibly obvious objection. The failure to immediately see this suggests to me that the peer review process is more enamored of playing at the most recent development in statistical methodology than in being assured that the use of such a methodology actually makes any sense.
[ii] To be clear, I think this objection is extremely weak. I doubt the effect of a new, compulsory education law would be to reduce years of education; this is quibbling over the exact amount of increased attendance. And if there was a lower increase in attendance as I am implying, then the actual effect of schooling on IQ would be even larger than estimated. While I am trying to steel man the argument that this study is poorly done—and by implication, that all of the instrumental variable analyses may be mistaken—the direction of the mistake may not always be against schooling causing IQ increases. My goal in this section is steel manning the case against this type of study; I need to do this because I think this study is pretty good. If the other instrumental variable studies are of similar quality, then this is a good literature.
[iii] We should also challenge the idea that we can measure increases in IQ gains legitimately, in general. See Haier, 2014. In essence, the problem is that IQ does not measure absolute units. A one inch increase is a one inch increase, whether the increase is from one to two inches, or from six feet to six feet and one inch. However, a 5 point IQ gain means something very different if it is from 70 to 75 compared to 130 to 135. A one “unit” increase in IQ is not one unit. It is a certain increase in percentiles. Inference for a rank ordered measure like IQ is not the same as inference for an absolute measurement like inches. That said, it is clear from Haier’s analysis that instrumental variable designs are less subject to difficulties than other analyses, and the most plausible analysis for IQ-schooling increases clearly was the instrumental variable design evidence.
[iv] The non-domain specific increase in working memory seems to end at a fourth-grade reading level, which is five years of education (Kindergarten to fourth grade). (Peng et al., 2018) However, not everyone gets five years of education in five years. Example: slightly more than half of Americans 18-74 read below a 6th grade reading level (Schmidt, 2022). We must imagine that increasing the years of education (say, from 7 to 9, as in the Norway case) would increase the proportion of individuals reading at a fourth-grade level or higher. This, in turn, would increase working memory, and thus fluid IQ.
[v] This suggests that not all years of education are created equal. The causal literature assumes that any increase in the number of years of schooling is the same, but the increases in the number of years tend to be around a middle range—say, from 7 to 9. The assumption of linearity seems like an extremely implausible one; did my last year as a PhD student really increase my IQ as much as my first year of learning to read? I doubt it. In general, it seems reasonable to assume that years of schooling have diminishing marginal returns to IQ, if there are any real gains.
[vi] Transfer of learning is extremely rare for any kind of intervention; however, there is one known method to promote transfer of learning. That is video games. (Dale and Green, 2016. Bavelier et al., 2012, explicitly identify the transfer results of video gaming as “learning to learn.”) Dale and Green discuss the reason that video games lead to transfer, and suggest this is so because they are fun, dynamic, and promote “learning to learn” by learning rules that can apply in many scenarios. (145-46) Based on the previous discussion, I would add the possibility that some of the transfer of learning that occurs from video games may be produced by the way in which they practice working memory. Cross sectional work confirms a small correlation between video game play and working memory (see Waris et al., 2019), though this is not experimental work; gamers who play a lot of video games have substantially more working memory than those who do not play at all. If the relationship between video games and working memory is causal, then this suggests that video games might increase fluid IQ as a form of working memory training.
[vii] Learning can be useful if we get rid of 3. Maybe I cannot do anything creative with my knowledge of a carburetor, but I do know my carburetor. I have taken it apart and put it back together again to clean it. Did that help me understand my engine? Honestly, not really. Learning how my carburetor works did not “teach me to think,” not even just about motorcycles in general. It taught me to diagnose and fix specific issues on my motorcycle, related to the carburetor. This has been extremely useful, but I doubt it increased my g. If the first two elements of transfer occur—bringing learning outside the classroom long after learning occurs—then schooling has succeeded at something. It has just not taught a person how to think; the success is narrow. But narrow domain knowledge can be extremely useful, nonetheless. Narrow knowledge of Shakespeare is no better than narrow knowledge of Christopher Nolan, but narrow knowledge of how to make good eggs has increased my happiness substantially.
[viii] We make this same mistake about learning transfer in at least one other important area: we often assume that parents have a massive influence on their children by raising them differently. (See Caplan, 2012, to summarize the evidence to the contrary.) In The Nurture Assumption, Judith Harris explicitly explains her findings that rearing styles are not as important as we typically believe by explaining that transfer of learning is rare, including transfer from parents and the household to other parts of society. Clearly, we as a society tend to assume greater levels of learning transfer than actually occur. This mistake shows itself in more than just the schooling context.
[ix] Willis explains that the working-class lads' own culture prepares them for working class work, so that there is an “element of self-damnation in the taking on of subordinate roles in Western capitalism. However, this damnation is experienced, paradoxically, as true learning, affirmation, appropriation, and as a form of resistance.” (Willis, 3)
[x] I would also suggest that this mechanism by which even the resistant are “civilized” cannot be effective in the same way anymore. I was shocked when I found these kids mouthing off Boomer slogans about how getting a job just involves going up and asking for one! One of the rebellious students talks about a teacher who constantly explains that school prepares students to get a job later: “He’s always on about if you go for a job, you’ve got to do this, you’ve got to do that. I’ve done it. You don’t have to do none of that. Just go to a place, ask for the man in charge, nothing like what he says.” (92) Just ask for the guy in charge!? I wish! Then I remembered that this comes from a study from the 1960s. These are the Boomers who could get a job by just asking. That world no longer exists, nor does the world where getting a factory job without a high school education could still result in a productive life taking care of an entire family. The popular explanation for the “deaths of despair” today is that the economy has not been good for uneducated whites. (Case and Deaton, 2017) Surely, one result of this changed economy is that there are fewer options of rebellion against schooling than there used to be. The boys in Learning to Labour could still live good lives, and their families could expect them to live good lives as factory workers. This is no longer true, and obviously so. That must change the whole relationship of the rebel to school, but it must also change how precisely the rebel is civilized, if at all. It used to be the working-class rebel kid could say: “there is a way to live honestly in the world, even if I won’t be a world leader in the process.” Now, the kid can see no other way than to be dishonest and submissive as the school demands, and the parent sees the same reality. The choice now is “live as a scumbag or live in penury.” It is easy to see how that choice could lead to despair. (This point dovetails with that made by Olsen, 26-29; schooling is now more important than ever, making it more dominant over the individual than ever.)
[xi] If they still produce it. It is worth noting that the reforms cited generally are not especially recent, and many of the studies are performed outside the United States. We have changed schooling a lot since then, and the US has always been different from the rest of the world. We might question the external validity of the results. Are we still increasing IQ today, assuming we were when these school reforms occurred?
[xii] It is well outside of my current subject to elaborate now, but I should be clear that I do not consider this distrust entirely unjustified at all.