Whether it is measured in degrees obtained or years of schooling completed, education correlates with many positive life outcomes, including income, civil and political participation (including the act of voting), charitable giving, life expectancy, health and well-being. Education is also negatively correlated with many “high-risk” behaviors; the less educated you are, the more likely you are to smoke, engage in substance abuse or commit crimes. But many questions remain unanswered about the precise mechanisms of cause and effect between education and these life outcomes. Does education, for example, directly affect income, or are the traits that cause a person to pursue more education the same ones that affect income or, for that matter, all the positive (and negative) outcomes we have mentioned? Is high educational attainment a major cause of health, wealth and longevity? Or is education simply another result of a common underlying trait such as “intellectual ability”?
Theories asserting that education is the direct cause of many life outcomes currently dominate the public debate about education policy (and some academic fields in the social sciences). These theories — grouped under the rubric of “human-capital accumulation” — assert that education causes positive outcomes by creating both specific, job-related skills and a more generalized increase in cognitive capacities. They are most clearly linked to University of Chicago economist and recent Nobel laureate Gary Becker and his seminal book Human Capital: A Theoretical and Empirical Analysis, with Special Reference to Education (1964).
But in the mid-1970s, economists A. Michael Spence and Joseph E. Stiglitz, who is also a Nobel laureate, both put forward countervailing theories that might explain why some abilities can be correlated with education, even though education might not improve them in any way. While the technical details of these theories vary, the mechanism posited is the same: According to these theories, it is not that education improves one’s ability but rather that education serves as a way for one to demonstrate himself/herself worthy, in the eyes of society in general or more specifically to potential employers. People with higher ability find it easier to successfully navigate the educational labyrinth and so distinguish — or, in the terms often used in the literature, sort or signal — themselves as different from people of lower ability by acquiring more education. The end result, these theories hold, is a strong correlation between ability and education, even though education does not necessarily increase ability. (Education may serve multiple purposes — not only signaling one’s abilities but simultaneously transmitting skills and providing accreditation. But either one or any combination of these things might cause ability to be correlated with education.)
For many years, one of the authors of this piece, Norman H. Nie, had been attempting to sort out the role that education can (and cannot) play in creating greater political and social equality. The questions raised in Nie et al.’s prize-winning 1996 book, Education and Democratic Citizenship in America, led a team of researchers at the Stanford Institute for the Quantitative Study of Society to become more generally interested in education and its impact on a variety of life outcomes and to attempt to distinguish those outcomes that seem more skill-driven from those that seem to come mainly from innate intellectual traits or ability.
Our original hypothesis was that verbal ability would surely be one of those outcomes driven mainly by years in the classroom. As we looked deeper into the data on that ability, however, we came to the counterintuitive conclusion that increased education actually does not appear to cause an increase in verbal ability. This finding in no way denies that education may create and transmit many different kinds of skills and capabilities. Nor does it obviate that matriculation through an ever-advancing scientific and technical curriculum clearly has an impact on what students learn and what work they are qualified to do. And finally, we wholeheartedly agree that reading the classics is good for the soul and the mind. Our argument is much narrower, indicating only that underlying verbal abilities are most likely a facilitator of high educational achievement and not a result thereof.
As in many other areas of research, the ultimate answers to the many questions about causes and effects of education probably lie somewhere between extremes: Education probably has some independent effect on earnings (or civil participation or any other outcome) and is probably affected (or determined) by some other factors that also impact earnings, frequency of voting and (on the opposite side) smoking and other high-risk behaviors. The academic literature in economics, education, social psychology and many other disciplines is filled with works trying to quantify the effects of education on different outcomes. (The economic literature alone on the relationship between education and income could fill a library.)
But there is no consensus about the results of these studies. And the question about the effects of education on life outcomes is not purely an academic one. It has important policy implications: If increasing a person’s education has a big effect on earning power (and/or citizenship and other positive outcomes), public investments in education will reap clear benefits for society and individuals. But if — as our research suggests — the underlying relationship between education and income (or other outcomes) is not one of cause but rather a mere correlation based on some innate differences in abilities, policies aimed at increasing educational attainment will have weak economic, quality-of-citizenship and other social benefits.
Education should not only be thought of as a means to an economic end. There are other reasons to pursue education and other reasons for public investment in education. Still, our results do suggest that the government may, in some cases, be spending money on certain kinds of education that might not produce the “human-capital building” results sought. For example, our findings would call into question government programs to mandate a shift from a high school degree as the endpoint of publicly funded education to an associate degree that would require two additional years of education. The change has been advocated as a way to be more consistent and competitive with the education systems of most of Europe and parts of Asia. But our research shows that previous investments in expanding education have not resulted in improvement in American adults’ verbal abilities. That finding suggests much more research is needed before enormous resources are committed to two additional years of general education that may — but may well not — deliver the social goods promised.
We understand that the area we are researching is a controversial one and that some may misapprehend our preliminary results as being somehow anti-education. Nothing could be further from the truth. Rather, we support additional research into the causes and effects of education so policymakers can make more informed decisions on how to spend limited educational funds to maximum positive effect for both individuals and society.
We set out to study the relationship between education and verbal ability because, among other things, verbal ability serves as the foundation for interpersonal communication and the transmission of ideas, both of which are necessary for success in the job market, and for the other positive outcomes we mentioned. There is also substantial literature asserting that verbal ability is the single most important indicator of general intelligence. Most important for our purposes, verbal ability is more directly related to education than many of the life outcomes we mentioned; it is part of the curriculum as well as a crucial tool for acquiring education.
Our initial hypothesis was that if amount of schooling causally affects any outcome, it would be verbal ability. The vast expansion of the American education system over the course of the 20th century served as our test bed. We expected that the huge increase in educational attainment in the U.S. across the decades would be accompanied by a substantial improvement in verbal abilities. To our initial amazement, we found no evidence for such improvement.
We started our investigation by showing that there is, indeed, a strong correlation between education and verbal ability. The data on which our analyses are based came from the General Social Survey, a program of in-person interviews that has been conducted regularly since 1972 by the National Opinion Research Center at the University of Chicago. While the samples were nationally representative, to avoid complications caused by changing demographics and questions about the validity of such tests with minority and immigrant populations, we included only the native-born, white American population 30 to 65 years of age, using information collected over the last 35 years of parallel surveys. (We used only those 30 years or older to ensure that we were dealing only with people who had completed their education; we stopped at age 65, lest we contaminate the analysis by differential mortality rates.)
Education levels and scores on a vocabulary test given to subjects are indeed correlated (see Figure 1). Over the three-plus decades studied, those with more education got better vocabulary scores, and vice versa.
Those results, however, do not necessarily imply that education causes increased verbal ability. If education did increase verbal ability, we would expect increasing levels of education over time to bring about measurably higher levels of verbal ability. During the 20th century, there was an unprecedented expansion in the levels of educational attainment in the U.S. The average American born between 1910 and 1914 received a bit more than 10 years of education. The average American born between 1970 and 1974 received 14 years of education. In 60 years, the “average American” went from being a high school dropout to having two years of college — a remarkable increase. The increase in education is across the board. A person born between 1910 and 1914 who obtained some postgraduate education was in the top 6 percent of his or her cohort in terms of education. By the 1970s, nearly 16 percent of the birth cohort had some postgraduate education. The percentage of college graduates or beyond has almost quadrupled over the same period, from just over 10 percent to almost 40 percent.
But, as Figure 2 shows, even though education has increased considerably through the decades, and even though education is correlated with verbal ability, verbal ability has stayed practically constant over time. The lack of change in the average vocabulary score of Americans, despite the large increase in the population’s average years of schooling, is an intriguing finding. There may be several explanations for it:
• Education does not necessarily improve verbal ability, and, therefore, the large investment in expanding the education system did not result in any change in vocabulary scores.
• Education does improve verbal ability, but the expansion of the education system came at a cost of harming the quality of education. The deterioration in quality of education was so large that the net effect of more education of lower quality was unchanged vocabulary scores.
• The test itself is not a valid measure of verbal ability, specifically vocabulary. During the 35 years of surveying and interviewing people born over a 70-year period, the language changed, rendering the test useless.
Without the ability to run long-term experiments, producing conclusive proof that education does not improve verbal ability is a formidable feat. It’s hard to prove that A does not affect B, if A and B are correlated. There will always be claims that you’re either doing A the wrong way (as in lowering the quality of education to educate a larger portion of the population for more years) or measuring B the wrong way (as in using standardized vocabulary tests that are not a good measure of verbal ability).
So to test our theory further, we took a two-pronged approach: We looked for more evidence supporting the notion that education does not affect verbal ability, and we investigated whether alternative explanations for our results could explain the full picture of education’s correlation to verbal ability. (For a detailed analysis and full methodological details, view the complete article on the Stanford University Web site.
If, as we believe, verbal ability is not improved by education, certain results should follow: We should not, for example, see any increase in verbal-ability test scores as educational levels rise over time. In fact, given a stable pool and distribution of verbal ability, if our theory is true, as more and more people obtain ever-higher levels of education, the pool of the most verbally able should become evermore depleted, leading the increased number of educated people to have lower and lower average scores on verbal-ability tests as the 20th-century education explosion unfolded.
And this is precisely what the data show. As the higher education system of the U.S. expanded, taking in and graduating more and more people, we dug deeper and deeper into the pool of talent, at least as far as verbal ability goes. If a college degree was once something only attained by the top 10 percent of each age cohort, today the top 40-plus percent of recent cohorts have received a bachelor’s degree. Clearly, the top 40 percent of a birth cohort are not as able as its top 10 percent. If education does little to increase verbal ability, one would expect that average scores for the larger proportion of less-able college students of today would be lower than for the smaller proportion of more-able college students of yesteryear.
It is obvious from looking at Figure 3 (and confirmed by formal statistical tests) that the verbal ability of graduates of all levels of education has decreased over time. At the same time, the graph also validates that the strong correlation between verbal ability and education has been maintained. At any given point in the time studied, graduates of higher levels of education always have, on average, higher vocabulary scores than those of lower educational levels.
Another piece of evidence regarding the noncausal correlation between education and verbal ability relates to gender differences. Men and women obtain different levels of education for various social, cultural and economic reasons at different points in time. One interesting point: As the 20th century progressed, changes in education levels were not gender equal. At the beginning of the century, women obtained substantially lower levels of education (on average) than men born in the same years. In the past, only the most able women were given the social and economic support to matriculate to college. But as the time passed, women closed the gap and even passed men in the number of years of schooling they achieved.
When we think about what should happen to the mean verbal abilities of both men and women, we expect to see two things:
1) If education does not affect verbal ability, we should see no significant increase in verbal ability for either men or women over the 20th century, even though women have gained, on average, 0.45 more years of education. This is shown in the bottom panel of Figure 4.
2) Assuming that men and women have similar verbal abilities in terms of the distribution of those abilities among men and among women, when fewer women obtained higher education than men, their verbal performance should have been higher. This result would have been due not to an innate verbal difference between men and women but to the fact that the top 7 percent of the population has higher verbal acuity than the top 18 percent of the population. And this was, in fact, the situation at the beginning of the century when only 21 percent of women obtained some college education, compared with 32 percent of the men. At that time, females with more than a high school education had higher verbal abilities on average than males with more than a high school education. At a later period (men and women born in the 1950s and beyond in our data), the education gap closed and even reversed. By 1955, more women than men had attained 12 or more years of education, and the gap in verbal ability had not only disappeared but in fact reversed — just as the ability-sorting theory of education would expect (see Figure 5). It is interesting to see that the crossover (from more men than women attaining highest education to more women than men) happened at the same time that their mean vocabulary scores also crossed over.
The evidence presented in the previous paragraphs, while not conclusive, is in line with theories positing that the relation between education and verbal ability is not one of cause. But are there alternative explanations for the data trends we’ve described? Could it be that a deteriorating education system and/or unreliable verbal tests obscured a reality in which education does, in fact, increase verbal ability?
It may very well be the case that the education system today is not as effective as it was 80 years ago, at least in the way it imparts verbal abilities and skills to the individuals who happen to go through it. But to explain the data we’ve mentioned, the system’s performance would have to be reduced by precisely the amount that negated all the effects of increased number of years of schooling — but not too much to lead to a declining overall average vocabulary score. Striking that precise balance continuously over a period of decades would seem an implausible ongoing coincidence.
Also, there is the gender problem. When we compare men and women who went through practically the same education system, only in different proportions, we find that their test scores at specific levels of education decreased by different amounts: As more women were getting more education (gaining more education than men born in the same years), their mean vocabulary scores deteriorated (even faster than mean vocabulary scores of men). This point is striking: If education did increase verbal abilities, to explain the lack of change in women’s mean vocabulary scores as they gained more education than their male counterparts, we’d have to believe that the education system has deteriorated more for women than it has for men. We’d also have to believe that the equality and progress that led to the increase in the quantity of education women received were accompanied by exactly the right amount of quality degradation to negate change in the final outcome.
The same finding appears when verbal Scholastic Aptitude Test scores of men and women are compared. Traditionally, women had higher verbal SAT scores than men. This fact plus similar findings from the General Social Service data and numerous other one-time studies actually made many social scientists believe that women had innately higher verbal ability than men at any given level of education. However, at the year women caught up to men in terms of the proportion taking the SATs, women’s mean average score matched men’s. In the following years, as the proportion of women taking the test well surpassed the proportion of men taking it, women’s mean scores fell well below men’s. This evidence once again suggests that the key to understanding the true causal relationship between education and verbal ability is through the sorting model.
To assert that education does not affect verbal ability, we also need to address doubts about the validity of the ability test itself: Aren’t we trying to squeeze a bit too much from one old lemon? After all, the measure we used is a 10-item vocabulary test, and even though it has been conducted regularly since 1973, it was constructed in the 1940s. The language might well have changed considerably during that period, rendering the test invalid.
These doubts are reasonable but can be addressed and, we feel, discounted. First, the GSS is the largest source of data that includes sociodemographic characteristics and information about education levels and that has been compiled regularly and for a significant time period. There are no better sources of data for the study we wanted to conduct. Second, other sources of data and other tests of verbal abilities and skills — such as the National Assessment of Adult Literacy, a study conducted by the National Center for Education Statistics, and the verbal parts of SAT scores — yielded very similar results, even though each of these sources has its own weaknesses.
Third, a series of rigorous tests shows the results would be similar even if the words least recognized in recent tests were excluded. That’s to say, a potential deterioration in the current recognition of any single word (or combination of words) used in the longstanding GSS vocabulary test is not the driving force behind the constant average in verbal test scores.
The conclusion that education does not affect verbal ability is a surprising one, and we have to be careful in interpreting it. Verbal abilities, though important and fundamental, are only one aspect of education; quantitative abilities are important, as are the more subject-oriented and vocational elements of the curriculum. Our aim is to show that even though education is correlated with many different positive life outcomes, there isn’t necessarily a causal relationship between the two, and each correlation has to be studied in depth. In terms of verbal ability, we have shown, education may act as an ability-sorting mechanism — a clear way to signal and certify those who are the best and the brightest.
Our findings from this research are, we believe, pregnant with important educational policy implications. Save for a small group of labor economists, the current wisdom among social science academics, policymakers and the educated public is that education is an additive social good, making the individuals who attain it and the society as a whole wealthier, healthier, more socially and politically participant and — alas — more verbally able. To paraphrase Becker’s now famous insight, education is among the most important creators of “human capital.” Much of this understanding is based on the persistent strong correlations between education and so many life outcome variables that are found at any and all points in time.
These correlations and general beliefs have led almost every public figure of almost every political persuasion to support, in one way or another, more education for an ever-larger portion of the population. But as we have shown, at least for the case of verbal ability, correlation is not always causation. Across 70 years and in multiple studies, vast increases in education have not caused an increase in verbal abilities. Verbal ability at all times is a great predictor of differences in educational attainment, but vast increases in overall amounts of education through the years had no overall effect on levels of verbal ability, which appear constant.
Our ongoing research on education and social outcomes at the Stanford Institute for the Quantitative Study of Society finds virtually identical patterns in education’s relationship with a host of other important life outcomes. Education is strongly and positively associated with measures of health, voluntary associational activity, charitable work, voting and other acts of political participation, interest in public affairs, and so on. Yet, as is the case with verbal ability, virtually none of these positive outcomes shows the pattern of increase that would be predicted if increased education caused those outcomes. In fact, measures of these attributes indicate that as more and more education is obtained by an ever-larger number of people, each outcome declines over time for every level of education. Once again, these results suggest that there is one (or more) inherent trait(s), relatively constant in mean and distribution across generations, that lead(s) to educational attainment, as well as to social and political engagement, dominance, well-being and so on.
We do not argue that ability sorting or signaling is education’s only or even predominant effect. First, while the expansion of education in America does not appear to have increased verbal ability or causally impacted a number of other important life outcomes, it surely has had other major impacts. There are no comparable data about quantitative, spatial or other abilities and skills of American adults that can be used for similar research. In the absence of such data, studying education’s impact on quantitative skills is impossible. For all we know, education could (or could not) work as a major causative factor to increase quantitative ability.
Also, general education and training certainly impart some specific skills, along with a certification that a particular curriculum has been mastered. Clearly, one would not go to her hair stylist for brain surgery or, for that matter, to her brain surgeon for legal help. Social networks and appropriate modes of speech, behavior and decorum are also acquired in schools and play important roles in life outcomes.
Finally, the most important policy implication of this research is to underline how little we firmly understand, after so many years of empirical and theoretical research, about the fundamental mechanisms and causal structure that link education to all that it purports to produce for individuals and societies. If we are to allocate educational funds with maximum efficiency, clearly we need to know with much more certainty what those funds can — and cannot — do to better both individuals’ opportunities and society as a whole.
For more than 30 years, educational attainment and vocabulary test scores have shown a positive correlation:
Huge increases in the level of educational attainment have not caused a corresponding increase in verbal ability, which has stayed roughly constant over time. Verbal ability has gone down at all levels of educational attainment over time — exactly as one would expect if the sorting theory of education were true:
As women went from receiving less education than men to receiving more, their verbal ability scores also changed in ways that coincide with an ability-sorting theory of education:
As higher percentages of men and women received 12 or more years of education, their average vocabulary scores did not rise but fell:
(Saar Golde, a research fellow at the Stanford Institute for the Quantitative Study of Society and Ph.D. candidate in the Economics Departmment at Stanford University, contributed to this article)
Sign up for our free e-newsletter.
Are you on Facebook? Click here to become our fan.
