Introduction

The most influential people in many sectors of American society—politics, the military, business, science, academia, arts, and the media—often wield considerable power, and their decision-making and gatekeeping activities have bearing on the broader culture. The military and science sectors are respected institutions. Politicians and business leaders have clear influence and power. Both the New York Times and the Wall Street Journal, whose editors and writers have enormous gatekeeping power regarding public attention, found themselves in disputes over what views were acceptable to publish. In politics, the words and actions of the President, as well as Senators, members of the House of Representatives, and other prominent politicians from the two major political parties influenced consequential decisions on health, the economy, and education. These policy debates are often dominated by individuals who have university degrees from a small set of “elite” institutions—out of the roughly 4000 degree granting postsecondary institutions one might choose to attend (NCES 2020)—and who expect their own children to obtain these degrees as well (Ponnuru 2020), illustrating just one consequence of social stratification due to education and its implications for society (Chetty et al. 2023; Hunt 2013; Sandel 2020; Scheffer et al. 2017). This runs counter to the widespread belief that extraordinary achievers attain their positions mainly because of chance, circumstance, or personal effort regardless of their educational background (Gladwell 2011).

The outsized influence these “elites” have on society raises questions about the critical importance of the individuals who end up assuming these gatekeeping functions and influential roles and what factors in American society contribute to their ascension to these positions (Bell et al. 2019; Buckley 1951; Mills 2000; Young 1994). In higher education, is it better or worse for political, business, or media “elites” to be educated at a comparative handful of extremely selective schools with others holding similar academic ambitions and credentials? Or should these leaders be drawn from across a broader set of institutions—the roughly 4000 one could choose from in the U.S. (NCES 2020)—that are less selective and also have fewer financial barriers to entry? These questions have been critical to understanding how students are selected into “elite” schools (Karabel 2006), how faculty hiring networks operate in relation to the prestige of the doctoral institution (Clauset et al. 2015) and also inform the discussion of the purpose of higher education, particularly at selective institutions (Deresiewicz, 2015; Sandel 2020). Additionally, inequality in higher education access may be linked to inequality in markets and society (Wai 2013). For example, some scholars note that, historically, a small “elite” class has dominated most societies, and that chance alone “will drive 1% or less of the community to dominate 50% of all resources” (Scheffer et al. 2017, p. 13154). Though the term resources typically refers to wealth, it also applies to sources of human and social capital such as “elite” educational opportunities (Klitgaard 1985) and to representation among societal leaders (Khan 2012).

Political historians (Eatwell and Goodwin 2018) have observed that in John F. Kennedy’s era, 71% of the House and 76% of the Senate held bachelor’s degrees, but by Barack Obama’s term these figures were 93% and 99%. Those scholars contended that “many people are now scrutinizing the corridors of power and see fewer people who look and sound like them” (p. 109). Thus “elites,” including political and other leaders, have become more powerful yet more distinct from the general population over time, and education may be part of that divide. Domhoff (1967) documented and argued persuasively that there is danger when wealth and power became concentrated among the upper class. Competition for selective college entry remains fierce and may be increasing (Brint and Yoshikawa 2017).

Dozens of studies have shown that attending college and obtaining undergraduate and graduate degrees are associated with greater income, occupational attainment, and professional success (Sewell and Hauser 1975; Tamborini et al. 2015). Many of these studies (e.g., the Wisconsin Longitudinal Study, the National Longitudinal Surveys of Youth) employ large, fairly representative, unselected cohorts of individuals born in the same year and followed throughout their lives to assess connections between educational attainment and adult incomes, job performance, and prestige. The results are robust and have been replicated many times, and form the basis for advice that assumes causality—specifically that attending college will lead to greater lifetime income and is therefore worth the cost of tuition, interest on loans, years spent, and other factors (Sewell and Hauser 1975; Tamborini et al. 2015).

However, these studies cannot illuminate the relationship between college attendance and truly outstanding achievements, such as becoming a U.S. president or senator, a billionaire, a Pulitzer prize-winner, a celebrated entrepreneur, or a member of the National Academies of Sciences, Engineering, and Medicine. This information gap arises because such outcomes are so rare that almost no participants in even large cohort datasets achieve them, and even when they do, such achievements are unlikely to be recorded in the data. These studies also do not speak to the relative value of attending different colleges and universities, for the same reason: even in large datasets, there are few participants who go to Ivy League or other top-ranked schools. Some evidence suggests that in certain sectors, graduates of top schools are dramatically overrepresented among the most successful achievers (Brint et al. 2020; Wai 2013; Wai and Rindermann 2015); however, these studies had relatively smaller samples, focused mainly on testing field-specific theories (Brint et al. 2020), and were not designed to be able to detect robust patterns across many different areas of outstanding achievements.

If education—perhaps especially at “elite” colleges and universities—may lie at the heart of the divide between the general public and these “elites,” what about the perception of the general public? Does the general public have an accurate empirical estimate of the educational backgrounds of powerful people?

In this investigation, we sought to answer two questions about the representation of graduates of “elite” educational institutions among extraordinary achievers in U.S. society. In Study 1, we build a new data set examining the American “elite” and use it to ask what proportions of over 26,000 top achievers across 30 diverse sectors—spanning politics, the military, business, law, science, academia, and the arts—had attended one of 34 top-ranked educational institutions we designated as “elite.” Within this group, we also examined the proportions who had attended an Ivy League institution, as well as Harvard University in particular. In Study 2, we asked whether members of the public are aware of the general pattern of facts established in Study 1—or whether they over- or under-estimate the proportions of high achievers who attended the “elite” schools.

Study 1: A new database of top achievers and their educational backgrounds

Creating a standard set of “elite” institutions is challenging due to the changing nature of criteria and rankings used to select which schools are “elite” (Labaree 2017). However, top institutions around the world do tend, with surprising regularity, to be the same from year to year and across different ranking systems, even with quite varying selection criteria (Wai et al. 2018), and the same schools ranked highly in the early decades of the 1900s are quite similar to those similarly ranked today (Healey 2014). For example, the “Elite” 34 schools used in this study were in the top 34 national universities or liberal arts colleges ranked by U.S. News, and they included 29 of the 34 schools that were in the top 34 of the Times Higher Education ranking of U.S. institutions. The “Elite” 34 schools in our study were selected as the schools with the highest average standardized test scores, although it is important to note that selecting based on other rankings or at different times would not substantially change the set of schools, nor would expanding the list with the next-highest-ranked 10 or 20 schools substantially change our conclusions. Table 1 lists the “Elite” 34 schools in alphabetical order (see supplement for more detail).

Table 1 List of “Elite” 34 Schools in Alphabetical Order.
Full size table

Additionally, because of the longstanding reputation of the Ivy League as a whole, and Harvard University in particular, for educating American leaders, intellectuals, and influencers, we examined the proportion who had specifically attended an Ivy League school and Harvard. Due to its influence, Harvard is often in the news in high profile discussions (e.g., Hooven 2023; Rios and Stein 2023; Zimmerman 2022), and not always in a positive light. However, because of its longstanding influence and historical significance, criticisms are to be expected (Bok 2024; Douthat 2006), and thus rightly or wrongly Harvard remains an institution worthy of study in its own right.

We created a unique new database of 26,198 high achievers in 30 different groups of achievement or recognition that included every member of the group for whom sufficient information about educational attainment was publicly available. For some groups this was easy (e.g., all U.S. Federal judges have official biographies with education recorded), whereas for others it was harder (e.g., biographies of startup company CEOs are not systematically created or organized). We also sought to ensure that our samples were outstanding or “elite” Americans. Specifically, Table 2 shows general descriptions about the groups that we did include (more details, including cutoff dates for inclusion, are in the supplement).

Table 2 Descriptions of Groups of American Extraordinary Achievers in this Study.
Full size table

Our primary goal is descriptive: to carefully document the representation of “elite” school graduates among U.S. extraordinary achievers across politics, the military, business, science, academia, and the arts. Our groups were selected in part by ascertaining what data is publicly available and that represented each of the major areas of politics, the military, business, science, academia, and the arts. We most certainly have left out prominent groups in these domains that may be worthwhile to investigate in future research, but believe we have compiled a reasonable representation of prominent groups in these core domains. Many scholars from political science, psychology, zoology, sociology, and education have noted that even when causal inference is difficult or impossible, it is vital to start by describing the social facts (Gerring 2012; Grimaldi and Engel 2009; Singer 2019; Tukey 1980). This is the primary purpose of Study 1.

Study 2: Do people know how many top achievers attended “Elite” educational institutions?

Given the extent of overrepresentation of “elite” school graduates in our database in Study 1, we next sought to determine whether the general public was aware of this pattern. Perhaps education lies at the heart of a divide between “elite” leaders and the general population (Eatwell and Goodwin 2018), and thus a refined understanding of the distribution of education and the public perceptions of that distribution should be documented. Therefore, as has been done for income, wealth (Norton 2014; Norton and Ariely 2011), and social mobility (Kraus and Tan 2015), we investigated public perceptions of how “elite” higher education is distributed among these highly influential groups. Even if the general public understands that access to higher education—especially “elite” higher education—is profoundly limited (e.g., low acceptance rates of top schools are commonly advertised; U.S. News and World Report 2021), do they have accurate perceptions of the extent to which influential leaders that span various sectors of American society have attended these institutions?

To measure public awareness of the distribution of “elite” school graduates in major high-achiever groups, in Study 2 we surveyed 1,810 individuals in an online panel and, under various conditions, asked them what proportions of the members of 15 of our 30 influential groups attended one of the “Elite” 34 schools, an Ivy League school, and Harvard. We provided base rate information (e.g., the estimated percentage of all U.S. individuals who attended these institutions) to some participants, and we also asked participants about the representation among extraordinary achievers of students from a comparison group of 34 similar-size, high-quality, but somewhat lower-ranked institutions.

Method

Sample

Study 1 included 30 different groups of U.S. extraordinary achievers with a total sample of 26,198 individuals. This included seven groups with data in prior published work and two groups with partial data in prior published work (see supplement). The remaining 21 groups were new to this analysis. Data collection on higher educational backgrounds for all samples occurred between 2012 and 2018 and was largely performed by separately searching for information on each individual in the sample. Many original samples included non-U.S. individuals; thus, for this study, each sample was restricted to individuals who listed at least one U.S. higher education institution in their backgrounds. Each of the databases were compiled with the intention of ensuring that a reasonably large sample size was reached for a stable analysis to represent that domain. Some samples, such as all the U.S. presidents and vice presidents extended throughout history, whereas other samples, such as the editors and writers of the New York Times and the Wall Street Journal, could only be collected in more recent years (see supplement).

Study 2 included 1810 participants recruited on Amazon Mechanical Turk (N = 1580 after exclusions; mean age = 36.4 years; 54.1% male). Participants who signed up to participate were taken to the online survey platform hosted by Qualtrics, where they completed the study and were compensated $1.25. We report results after excluding participants who failed an attention check, though the results are similar when including these participants and when using alternative exclusion criteria (see supplement).

Study 1 involved only compilation and analysis of publicly available data, and thus is not considered human subjects research and was exempt from IRB review. The anonymous online surveys for Study 2 were determined to be exempt by the Geisinger Institutional Review Board, and informed consent was not required.

Procedure

Study 1 included a coding protocol that determined the set of “Elite” 34 schools that participants could have attended in some capacity for undergraduate or graduate education. This set consisted of schools that typically required the highest average standardized test scores (e.g., SAT, ACT, GRE, GMAT, LSAT) for admission (see Table 1 for a list of these 34 schools). This method, which also included coding of the Ivy League, Harvard and other educational categories, was able to capture a wide range of individuals who attended schools with high average scores at either the undergraduate or graduate level and was applied to each of the 30 groups. This coding protocol drew from test scores reported to U.S. News, which are explicitly done for undergraduate institutions (SAT, ACT), business schools (GMAT) and law schools (LSAT). In order to more broadly capture individuals who likely scored highly on the GRE as attaining “elite” school status, we determined those with a graduate degree outside of law and business as “elite” if they attended one of the same schools determined as “elite” for undergraduate institutions. This does not mean that highly selective public institutions are not “elite” in their own way, only that we needed to keep the list of selective schools consistent across all categories. Additionally, for categories such as admirals and generals, schools like West Point most certainly would seem to be more influential, but we aimed at a consistent set of schools to capture general “eliteness.”

As some samples extended back through history whereas others were contemporary, one might question the use of more contemporary data from U.S. News for older samples. An analysis looking at whether proportions attending an “Elite” 34 school for older versus more contemporary samples was conducted (see supplement for more detailed explanation) and the oldest cohorts were more likely to have higher percentages attending an “Elite” 34 school relative to more recent cohorts, suggesting that the findings would be biased towards higher percentages for older cohorts rather than contemporary cohorts. General population base rates for each of the coding categories were also computed to allow comparison of the degree of representation of “elite” schools across these 30 domains. As a benchmark, there are roughly 4000 degree-granting higher-education institutions in the U.S. (NCES 2020). The 34 “elite” schools thus represent less than 1% of all degree-granting institutions of higher education in the U.S. See supplement for more detail for Study 1.

Study 2 participants read a brief introduction describing the purpose of the study, and were then shown a list of the “Elite” 34 schools and instructed to read it over carefully. Participants were then given a numbered list of 15 groups (selected from the 30 groups in Study 1 for visibility and representativeness) and were able to click and drag each group so that “1” would be the group with the highest percentage of members, and “15” would be the group with the lowest percentage of members attending each group of schools. Following this ranking task, participants completed a demographic questionnaire. The list of schools (along with base rate information for those assigned to the condition to make judgements based on base rate data) remained displayed on the screen as participants completed their estimates. As an attention check, participants were shown an additional screen in which they were instructed to enter the code “HKEQI” in each of the boxes.

Results

Study 1

Across all 30 groups, we discovered that roughly half (54.2%) of the individuals included had attended one of the “Elite” 34 schools, ranging from 11.2% to 25.9% (Four Star Generals, Four Star Admirals, House members) up through 78.9% to 80.9% (Forbes most powerful men, Harvard Faculty, American Philosophical Society). Ivy League percentages ranged from 5.1% for Four Star Generals up through 70.0% for Fields Medalists. Harvard percentages ranged from 2.0% for Four Star Generals up through 44.5% for Harvard faculty members. Examples of groups that fell into the middle ranges for the “Elite” 34 schools included Fortune 500 CEOs (41.9%), Wall Street Journal editors and writers (50.8%), National Academy of Medicine members (60.5%), and National Academy of Sciences members (70.5%).

Figure 1 shows the educational backgrounds of successful and influential American extraordinary achievers classified by the educational selectivity method used in this study. Red bars indicate the proportion that attended Harvard in some capacity. Green bars indicate the proportion that attended an Ivy League institution (other than Harvard). Blue bars indicate the proportion that attended one of the “Elite” 34 schools (other than the Ivy League). Yellow bars indicate the proportion that attended graduate school (but never attended an “Elite” 34 school). Purple bars indicate the proportion that attended college (but neither graduate school nor an “Elite” 34 school). Gray bars indicate the proportion who did not attend college or those for whom we could not find the necessary data (only 0.7%, on average, of the total sample in Fig. 1 had missing education data). These six categories are independent of one another and sum to 100%.

Fig. 1: Educational backgrounds of American extraordinary achievers across 30 domains.
figure 1

Note. The domains are ordered from top to bottom by the proportion of individuals who attended any “Elite” 34 school (including Ivy League and Harvard). For a benchmark, there are roughly 4,000 degree granting postsecondary institutions in the U.S.

Full size image

The top half of the “Elite” 34 schools

Taking a direct non-weighted average of the percentages across all groups studied shows that about 54.2% attended one of the “Elite” 34 schools in Supplementary Table 1 and that about 44.6% attended one of the top 16 “elite” schools in Supplementary Table 2. That is, a background at one or more of just 34 schools is found in the majority of our 26,198 outstanding achievers, and just 16 of these schools account for most of this “Elite” 34 representation.

The Ivy league and Harvard University

Taking a direct non-weighted average of the percentages within just the Ivy League institutions shows that about 36.3% attended one of just these eight schools. Harvard alone accounts for 16%, meaning that a single institution accounts for almost a sixth of all American extraordinary achievers. Of the “Elite” 34 schools, 82.3% were accounted for by the top 16 schools, 67% were accounted for by the Ivy League, and 29.5% were accounted for by Harvard.

Undergraduate versus graduate degrees

Taking a direct non-weighted average of the percentages within Ivy League undergraduate (19.6%) and graduate (25.3%) institutions suggests that graduate degrees are more common. More broadly, the percentages within the “Elite” 34 schools (undergraduate = 32.2%; graduate = 39.4%) and the top 16 (undergraduate = 23.3%; graduate = 33.7%) suggests that overall graduate degrees from highly selective schools appear somewhat more common than undergraduate degrees.

Figure 2 illustrates the degree of overrepresentation of Harvard, the Ivy League, and the “Elite” 34 schools relative to the base rate in the population. Of all adults in the general U.S. population, about 32.5% have received a bachelor’s degree or higher. Within this 32.5%, about 1.9% have a degree from one of the “Elite” 34 schools in our study, about 0.6% have a degree from one of the eight Ivy League schools, and about 0.2% have a degree from Harvard University. (These estimates were computed using data from the 2015 Current Population Survey in the U.S. and the Integrated Postsecondary Data System.) These data indicate that the percentages of individuals in each of the groups of American leaders and influencers are quite high relative to population base rates. For example, with 54% of the 26,198 individuals in our sample having attended one of the “Elite” 34 schools and the base rate at about 1.9%, this suggests a factor of overrepresentation of roughly 28 times base rate expectations (calculated as 54/1.9). Across all groups, roughly 36% attended an Ivy League school, suggesting a factor of overrepresentation of roughly 60 times (36/0.6). Across all groups, roughly 16% attended Harvard University, suggesting a factor of overrepresentation of roughly 80 times (16/0.2). This overrepresentation factor is 75 times even when Harvard faculty members are omitted (15/0.2).

Fig. 2: Educational backgrounds of extraordinary achievers (left panel) compared to the general U.S. population (right panel).
figure 2

Each icon represents 1/400 (0.25%) of the relevant population.

Full size image

Study 2

In Study 2, we examined the extent to which people are aware of the distribution of education in the U.S. among influential outstanding achievers. After conducting six pilot studies (N = 692) to pretest our study design and materials, we recruited workers on Amazon Mechanical Turk to provide estimates of the percentage of successful and influential U.S. individuals who attended one of the “Elite” 34 schools. Participants provided estimates for 15 of our 30 groups, with those 15 selected to be more likely to be publicly recognizable and representing different domains of achievement (e.g., science, arts, politics). Additionally, we wanted these 15 groups to represent the range of “Elite” 34 school percentages found in Study 1, from high percentages (e.g., the American Philosophical Society) to low (e.g., Four-Star Generals).

It is possible that participants would give similar estimates for “elite” schools as they would for any list of schools. To address this possibility, we assigned one-third of participants to provide estimates for an alternative list of 34 schools as a control. We selected a list of 34 non-“Elite” 34 schools that were similar to the “Elite” 34 schools in terms of public recognition and average enrollment. This subset of participants in Study 2 also provided estimates for the list of Control 34 schools, the eight most highly ranked of these schools, and the highest-ranked of these schools (the University of Rochester). Across all groups and categories, participants who were shown the list of “elite” schools provided higher estimates of how many extraordinary achievers attended them than participants who were shown the list of control schools, indicating that participants saw the list of “elite” schools as more “elite” than the list of “non-elite” schools (see supplement for full results). This finding suggests that people are at least somewhat aware of the relatively high influence of “elite” institutions’ graduates.

Within each of the groups, we experimentally manipulated whether participants were given population base rate information. We also manipulated the order in which participants were asked to estimate the three different categories of schools (see supplement for results). Half of participants (randomly assigned) were given concrete general population base rate estimates of the percentage of U.S. individuals who have earned a bachelor’s degree, as well as a degree from one of the “elite” (or control) 34 schools, one of the 8 Ivy-League (or control top-8) schools, and Harvard University (or control school University of Rochester).

We found that participants largely underestimated the percentage of influential individuals who attended “elite” schools (see Fig. 3). For 10 out of the 15 groups, participants gave lower estimates than the actual percentages from Study 1. Further, the greatest underestimates occurred for groups with the highest percentage of members from one of the “Elite” 34 schools: Participants underestimated by 35.2% for the American Philosophical Society (45.7% estimated vs. 80.9% actual), 15.4% for Harvard Faculty members (65.1% vs. 80.5%), 30.2% for Nobel Prize Winners (45.9% vs. 76.1%), 20.1% for National Academy of Science members (50.4% vs. 70.5%), and 17.4% for MacArthur fellows (41.5% vs. 58.9%).

Fig. 3
figure 3

Participants’ Estimates of the Percentage of Individuals who attended one of the “Elite” 34 schools and the Actual Percentages From Study 1.

Full size image

In addition, four out of the five cases in which participants made overestimates were for groups that had the lowest number of members who had attended one of the “Elite” 34 schools. Participants overestimated by 16.2% for U.S. Four-Star Generals (27.4% estimated vs. 11.2% actual), 17.9% for members of the House (43.8% vs. 25.9%), 11.6% for Federal judges (53.4% vs. 41.8%), and 5.2% for Fortune 500 CEOs (47.1% vs. 41.9%). We found similar results for Ivy League schools (11/15 groups underestimated) and Harvard (8/15 groups underestimated). In addition, at the end of the study, we asked participants to rank each group by the percentage of members they thought had attended one of the “Elite” 34 schools, one of the Ivy League, or Harvard. Participants’ rankings on this task closely matched the ranked results of their percentage estimates, suggesting that our results are robust to different methods (see supplement for full ranking results).

Base rate information

For all of the 15 groups across the “Elite” 34, Ivy League, and Harvard, participants provided with base rate information gave lower estimates than participants not provided with base rates (full results in supplement). Thus, when people were made aware of the actual percentage of individuals who attended “elite” schools in the general population, they appear to have logically, yet incorrectly, assumed that a smaller number of influential people (perhaps than they originally thought) attended these schools, and adjusted their estimates downward accordingly. But this adjustment may have led to even greater underestimation because the percentage of influential people with “elite” educational backgrounds was typically higher than they thought even without base rate information.

Discussion

The impact of “elites” in American society is substantial and may be increasing as society becomes more unequal. In Study 1, we utilized attainment of rare, low-base-rate long-term outcomes as a way to describe the presence and variation of “elite” higher education among influential Americans. These findings document the importance of “elite” institutions (including the Ivy League and in particular Harvard) as one gateway in the life course of people who end up as leaders and influencers, among whom they are overrepresented by a factor of 28 to 80 times. In Study 2, we documented public perceptions of the Study 1 empirical patterns. We found that participants largely underestimated the degree to which influential individuals attended “elite” schools, similar to findings on wealth inequality (Norton and Ariely 2011). In the cases where participants made overestimates, these were for groups who had the lowest attendance at one of the “Elite” 34 schools, or groups that may be seen as more representative of the people (e.g., Four Star Generals, House of Representatives members, and to a lesser extent federal judges and Fortune 500 CEOs), even if not selected by them directly.

Potential mechanisms explaining our findings

Three main classes of explanatory factors for the findings in this study are 1. Student characteristics, 2. Educational and social (e.g., peer) aspects, and 3. Reputational or network effects. Econometric studies (Dale and Kreuger 2002) have shown that once unobservable factors are accounted for, there is little to no impact of attending an “elite” institution on conventional outcomes such as long-run income. This finding suggests that student characteristics, such as developed cognitive abilities; knowledge and skills acquired in secondary school; personality traits, such as conscientiousness and achievement motivation; and other aspects may be important (Schmidt and Hunter 1998; Wolf 2003). These results do not, however, vitiate the possible importance for extraordinary achievement of factors such as social networks attained through schools, including alumni in positions of power, or the potential impact of going through college.

Other inputs, such as family wealth or social background, could affect one’s odds of getting into an “elite” institution and of getting a job or attaining other outcomes that are socially influenced where one would need to be elected or picked (Chetty et al. 2023). The alumni bias effect may be operating, for example, if a company has many “elite” school graduates who tend to refer and hire other such graduates. Much hiring is done in this way, especially in business and finance (Frank 2000; Rivera 2016). The findings for the Ivy League are noteworthy in this regard. Moreover, the incredible influence of Harvard University—overrepresented among remarkable achievers by up to 80 times in a pattern one might call the Harvard Effect—is documented quantitatively here for the first time. Perhaps attending Harvard is not only a reflection of inputs, a gateway to critical educational and social experiences, or access to invaluable alumni networks, but also a more subtle treatment effect that increases one’s ambitions or expectations. Of course, it is unclear whether it is the quality of the incoming student body, the education they receive, the people they rub shoulders with, or network or other effects, such as luck (Pluchino et al. 2018) that matter, and in what measure.

Potential implications

The elected “elites” tended to be more similar to the general population in education selectivity, whereas the anointed “elites” tended to be quite different (Hacker 1961). Presidents and vice presidents, Senators, House members, federal judges, and four-star admirals and generals were all in the bottom half of “elite” school representation. And people were more likely to overestimate attendance for groups more familiar to them (e.g., House members). Critics of selective higher education have noted such institutions may be contributing to inequality (Chetty et al. 2014); thus, our findings may have implications for education policies directed towards injecting socioeconomic and demographic diversity into U.S. leadership through “elite” college admissions and debates about how such spots are allocated (Hoxby and Avery 2013; Klitgaard 1985; Sandel 2020).

Limitations

Our study is descriptive and cannot identify causal mechanisms explaining the patterns we documented. And though our dataset of over 26,000 U.S. extraordinary achievers spread across 30 different sectors is the largest of its kind to date, the samples we chose may not fully represent the broader population of American remarkable achievers. For example, these individual U.S. focused samples detailed in Study 1 were sometimes drawn from different years, and may confound the comparison with the baseline comparison data taken in 2015. Broadly, the temporal heterogeneity of the data may have impacted our findings in ways we cannot quite anticipate, though prior studies have illustrated these findings more globally (Wai et al. 2019) (see methods and supplement for detailed discussion; also see [Laouenan et al. 2022] for types of historical databases such as Wikipedia and related biographies that might be used for future work on this topic). Additionally, see Heisig et al. (2020) for a cross-national dataset that shows elite formation is not only an American phenomenon. As one example, we could have included other media outlets, but we chose the two newspapers with the largest audiences, longest histories, and clearest identification as serious mainstream outlets with contrasting political leanings. It is a lot of work to assemble and code these data, so some choices had to be made. A deeper investigation of the broader and evolving media landscape could be a good topic for future research, and this applies also to a variety of other extraordinary achievement domains.

Additionally, though understanding the degree of representation of African Americans, Hispanics, Native Americans, and other underrepresented groups is important, we could not reliably or comprehensively obtain race and ethnicity information in a comparable way across our samples. Our analysis is also unable, by design, to adjudicate between the multiple explanations for associations between “elite” education and achievement.

The strength of this study is the compilation of a large and completely novel dataset with a focus on groups who have achieved outcomes unlikely to be observed in large cohort studies, and its description of the empirical data patterns that exist (Gerring 2012; Grimaldi and Engel 2009; Singer 2019; Tukey 1980). However, as noted, our data sources cannot disentangle the relative importance of the inputs compared to the outputs of a college or university education (Dale and Kreuger 2002), or the role of human capital versus signaling (Huntington-Klein 2020). Perhaps whether one merits admission to an “elite” school is questionable, but the fact of having been admitted itself serves as a more or less accurate or inaccurate signal in future selection procedures. For example, deliberate or unintentional self-replication of elites in selection procedures may occur, whereby, for instance, Ivy League graduates hire primarily other Ivy League graduates, or the military gives preference in promotion to graduates of military academies. In other words, a labelling effect may be taking place. We looked back at history, but past performance is no guarantee of future results, and whether the “elite” schools will retain their influence or legitimacy is something that will need to be studied carefully in the future. Limitations of Study 2 include the possibility that participants may not have had a clear understanding of some of the extraordinarily achieving groups we studied, though we did provide brief explanations of what each of the groups were (see methods and supplement). Regarding estimates of “elite” education of various groups, the lack of complete understanding of these groups due to lack of exposure, may have led to idiosyncratic ways in which these estimates may have been biased (Landy et al. 2017; Orth 2022).

Conclusion

In this study, we documented the social relevance of “elite” education in America, specifically the surprisingly high prevalence of graduates of a small number of “elite” schools in positions of power and influence. This concentration is greatest in positions not subject to public elections or political appointments, and this cognitive and social stratification may have important implications for society (Chetty et al. 2014; Hunt 2013; Khan 2012). The leaders we studied are in decision making positions that impact multiple areas of our lives. The developed characteristics and experiences of the people who end up in these positions of influence most certainly matter, and if most of them have gained their formative educational and life experiences at a specific set of institutions, this can have consequences for many aspects of social stratification, networks, and the way that educational or other forms of inequality are distributed in online markets that may increasingly be winner take all, where the 1% possess half of all resources, including wealth or other dimensions (Scheffer et al. 2017).

Perhaps it would be better if there were a greater diversity of people being educated at one of the “Elite” 34 institutions, or perhaps it would be better if the people who ended up in these positions of influence were educated at a wider range of institutions that are more representative of populations that span all regions of the U.S. (for example, one of the other roughly 4,000 schools; NCES 2020). Whichever side one comes down on, we provide for the first time a mega-analysis of the “elite” structure of the American social world based on educational history and how people understand that structure. We show that “elite” education is highly concentrated at the top in groups of important societal influencers and that the general public tends to underestimate how much that concentration is. We are certainly not the first to notice the American system of “elite” privilege. However, we have created a new data set which highlights how the “elite” interacts. These social facts are important to document, if nothing else, to stimulate discussions around what could be versus what currently is.