scholarly journals for education

• All subject areas
• Agricultural and Biological Sciences
• Arts and Humanities
• Biochemistry, Genetics and Molecular Biology
• Business, Management and Accounting
• Chemical Engineering
• Computer Science
• Decision Sciences
• Earth and Planetary Sciences
• Economics, Econometrics and Finance
• Engineering
• Environmental Science
• Health Professions
• Immunology and Microbiology
• Materials Science
• Mathematics
• Multidisciplinary
• Neuroscience
• Pharmacology, Toxicology and Pharmaceutics
• Physics and Astronomy
• Social Sciences
• All subject categories
• Acoustics and Ultrasonics
• Advanced and Specialized Nursing
• Aerospace Engineering
• Agricultural and Biological Sciences (miscellaneous)
• Agronomy and Crop Science
• Algebra and Number Theory
• Analytical Chemistry
• Anesthesiology and Pain Medicine
• Animal Science and Zoology
• Anthropology
• Applied Mathematics
• Applied Microbiology and Biotechnology
• Applied Psychology
• Aquatic Science
• Archeology (arts and humanities)
• Architecture
• Artificial Intelligence
• Arts and Humanities (miscellaneous)
• Assessment and Diagnosis
• Astronomy and Astrophysics
• Atmospheric Science
• Atomic and Molecular Physics, and Optics
• Automotive Engineering
• Behavioral Neuroscience
• Biochemistry
• Biochemistry, Genetics and Molecular Biology (miscellaneous)
• Biochemistry (medical)
• Bioengineering
• Biological Psychiatry
• Biomaterials
• Biomedical Engineering
• Biotechnology
• Building and Construction
• Business and International Management
• Business, Management and Accounting (miscellaneous)
• Cancer Research
• Cardiology and Cardiovascular Medicine
• Care Planning
• Cell Biology
• Cellular and Molecular Neuroscience
• Ceramics and Composites
• Chemical Engineering (miscellaneous)
• Chemical Health and Safety
• Chemistry (miscellaneous)
• Chiropractics
• Civil and Structural Engineering
• Clinical Biochemistry
• Clinical Psychology
• Cognitive Neuroscience
• Colloid and Surface Chemistry
• Communication
• Community and Home Care
• Complementary and Alternative Medicine
• Complementary and Manual Therapy
• Computational Mathematics
• Computational Mechanics
• Computational Theory and Mathematics
• Computer Graphics and Computer-Aided Design
• Computer Networks and Communications
• Computer Science Applications
• Computer Science (miscellaneous)
• Computer Vision and Pattern Recognition
• Computers in Earth Sciences
• Condensed Matter Physics
• Conservation
• Control and Optimization
• Control and Systems Engineering
• Critical Care and Intensive Care Medicine
• Critical Care Nursing
• Cultural Studies
• Decision Sciences (miscellaneous)
• Dental Assisting
• Dental Hygiene
• Dentistry (miscellaneous)
• Dermatology
• Development
• Developmental and Educational Psychology
• Developmental Biology
• Developmental Neuroscience
• Discrete Mathematics and Combinatorics
• Drug Discovery
• Drug Guides
• Earth and Planetary Sciences (miscellaneous)
• Earth-Surface Processes
• Ecological Modeling
• Ecology, Evolution, Behavior and Systematics
• Economic Geology
• Economics and Econometrics
• Economics, Econometrics and Finance (miscellaneous)
• Electrical and Electronic Engineering
• Electrochemistry
• Electronic, Optical and Magnetic Materials
• Emergency Medical Services
• Emergency Medicine
• Emergency Nursing
• Endocrine and Autonomic Systems
• Endocrinology
• Endocrinology, Diabetes and Metabolism
• Energy Engineering and Power Technology
• Energy (miscellaneous)
• Engineering (miscellaneous)
• Environmental Chemistry
• Environmental Engineering
• Environmental Science (miscellaneous)
• Epidemiology
• Experimental and Cognitive Psychology
• Family Practice
• Filtration and Separation
• Fluid Flow and Transfer Processes
• Food Animals
• Food Science
• Fuel Technology
• Fundamentals and Skills
• Gastroenterology
• Gender Studies
• Genetics (clinical)
• Geochemistry and Petrology
• Geography, Planning and Development
• Geometry and Topology
• Geotechnical Engineering and Engineering Geology
• Geriatrics and Gerontology
• Gerontology
• Global and Planetary Change
• Hardware and Architecture
• Health Informatics
• Health Information Management
• Health Policy
• Health Professions (miscellaneous)
• Health (social science)
• Health, Toxicology and Mutagenesis
• History and Philosophy of Science
• Horticulture
• Human Factors and Ergonomics
• Human-Computer Interaction
• Immunology and Allergy
• Immunology and Microbiology (miscellaneous)
• Industrial and Manufacturing Engineering
• Industrial Relations
• Infectious Diseases
• Information Systems
• Information Systems and Management
• Inorganic Chemistry
• Insect Science
• Instrumentation
• Internal Medicine
• Issues, Ethics and Legal Aspects
• Leadership and Management
• Library and Information Sciences
• Life-span and Life-course Studies
• Linguistics and Language
• Literature and Literary Theory
• LPN and LVN
• Management Information Systems
• Management, Monitoring, Policy and Law
• Management of Technology and Innovation
• Management Science and Operations Research
• Materials Chemistry
• Materials Science (miscellaneous)
• Maternity and Midwifery
• Mathematical Physics
• Mathematics (miscellaneous)
• Mechanical Engineering
• Mechanics of Materials
• Media Technology
• Medical and Surgical Nursing
• Medical Assisting and Transcription
• Medical Laboratory Technology
• Medical Terminology
• Medicine (miscellaneous)
• Metals and Alloys
• Microbiology
• Microbiology (medical)
• Modeling and Simulation
• Molecular Biology
• Molecular Medicine
• Nanoscience and Nanotechnology
• Nature and Landscape Conservation
• Neurology (clinical)
• Neuropsychology and Physiological Psychology
• Neuroscience (miscellaneous)
• Nuclear and High Energy Physics
• Nuclear Energy and Engineering
• Numerical Analysis
• Nurse Assisting
• Nursing (miscellaneous)
• Nutrition and Dietetics
• Obstetrics and Gynecology
• Occupational Therapy
• Ocean Engineering
• Oceanography
• Oncology (nursing)
• Ophthalmology
• Oral Surgery
• Organic Chemistry
• Organizational Behavior and Human Resource Management
• Orthodontics
• Orthopedics and Sports Medicine
• Otorhinolaryngology
• Paleontology
• Parasitology
• Pathology and Forensic Medicine
• Pathophysiology
• Pediatrics, Perinatology and Child Health
• Periodontics
• Pharmaceutical Science
• Pharmacology
• Pharmacology (medical)
• Pharmacology (nursing)
• Pharmacology, Toxicology and Pharmaceutics (miscellaneous)
• Physical and Theoretical Chemistry
• Physical Therapy, Sports Therapy and Rehabilitation
• Physics and Astronomy (miscellaneous)
• Physiology (medical)
• Plant Science
• Political Science and International Relations
• Polymers and Plastics
• Process Chemistry and Technology
• Psychiatry and Mental Health
• Psychology (miscellaneous)
• Public Administration
• Public Health, Environmental and Occupational Health
• Pulmonary and Respiratory Medicine
• Radiological and Ultrasound Technology
• Radiology, Nuclear Medicine and Imaging
• Rehabilitation
• Religious Studies
• Renewable Energy, Sustainability and the Environment
• Reproductive Medicine
• Research and Theory
• Respiratory Care
• Review and Exam Preparation
• Reviews and References (medical)
• Rheumatology
• Safety Research
• Safety, Risk, Reliability and Quality
• Sensory Systems
• Signal Processing
• Small Animals
• Social Psychology
• Social Sciences (miscellaneous)
• Social Work
• Sociology and Political Science
• Soil Science
• Space and Planetary Science
• Spectroscopy
• Speech and Hearing
• Sports Science
• Statistical and Nonlinear Physics
• Statistics and Probability
• Statistics, Probability and Uncertainty
• Strategy and Management
• Stratigraphy
• Structural Biology
• Surfaces and Interfaces
• Surfaces, Coatings and Films
• Theoretical Computer Science
• Tourism, Leisure and Hospitality Management
• Transplantation
• Transportation
• Urban Studies
• Veterinary (miscellaneous)
• Visual Arts and Performing Arts
• Waste Management and Disposal
• Water Science and Technology
• All regions / countries
• Asiatic Region
• Eastern Europe
• Latin America
• Middle East
• Northern America
• Pacific Region
• Western Europe
• ARAB COUNTRIES
• IBEROAMERICA
• NORDIC COUNTRIES
• Afghanistan
• Bosnia and Herzegovina
• Brunei Darussalam
• Czech Republic
• Dominican Republic
• Netherlands
• New Caledonia
• New Zealand
• Papua New Guinea
• Philippines
• Puerto Rico
• Russian Federation
• Saudi Arabia
• South Africa
• South Korea
• Switzerland
• Syrian Arab Republic
• Trinidad and Tobago
• United Arab Emirates
• United Kingdom
• United States
• Vatican City State
• Book Series
• Conferences and Proceedings
• Trade Journals

• Citable Docs. (3years)
• Total Cites (3years)

Follow us on @ScimagoJR Scimago Lab , Copyright 2007-2024. Data Source: Scopus®

Cookie settings

Cookie Policy

Legal Notice

Privacy Policy

An official website of the United States government

Official websites use .gov A .gov website belongs to an official government organization in the United States.

Secure .gov websites use HTTPS A lock ( Lock Locked padlock icon ) or https:// means you've safely connected to the .gov website. Share sensitive information only on official, secure websites.

• Publications
• Account settings
• Advanced Search
• Journal List

A century of educational inequality in the United States

Michelle jackson, brian holzman.

• Author information
• Article notes
• Copyright and License information

To whom correspondence may be addressed. Email: [email protected] .

Edited by Eric Grodsky, University of Wisconsin-Madison, Madison, WI, and accepted by Editorial Board Member Mary C. Waters June 3, 2020 (received for review April 27, 2019)

Author contributions: M.J. and B.H. designed research; M.J. and B.H. analyzed data; and M.J. wrote the paper.

Issue date 2020 Aug 11.

This open access article is distributed under Creative Commons Attribution-NonCommercial-NoDerivatives License 4.0 (CC BY-NC-ND) .

Significance

There has been widespread concern that the takeoff in income inequality in recent decades has had harmful social consequences. We provide evidence on this concern by assembling all available nationally representative datasets on college enrollment and completion. This approach, which allows us to examine the relationship between income inequality and collegiate inequalities over the full century, reveals that the long-standing worry about income inequality is warranted. Inequalities in college enrollment and completion were low for cohorts born in the late 1950s and 1960s, when income inequality was low, and high for cohorts born in the late 1980s, when income inequality peaked. This grand U-turn means that contemporary birth cohorts are experiencing levels of collegiate inequality not seen for generations.

Keywords: educational inequality, income inequality, long-term trend

The “income inequality hypothesis” holds that rising income inequality affects the distribution of a wide range of social and economic outcomes. Although it is often alleged that rising income inequality will increase the advantages of the well-off in the competition for college, some researchers have provided descriptive evidence at odds with the income inequality hypothesis. In this paper, we track long-term trends in family income inequalities in college enrollment and completion (“collegiate inequalities”) using all available nationally representative datasets for cohorts born between 1908 and 1995. We show that the trends in collegiate inequalities moved in lockstep with the trend in income inequality over the past century. There is one exception to this general finding: For cohorts at risk for serving in the Vietnam War, collegiate inequalities were high, while income inequality was low. During this period, inequality in college enrollment and completion was significantly higher for men than for women, suggesting a bona fide “Vietnam War” effect. Aside from this singular confounding event, a century of evidence establishes a strong association between income and collegiate inequality, providing support for the view that rising income inequality is fundamentally changing the distribution of life chances.

It has long been suspected that the takeoff in income inequality has made the good luck of an advantaged birth ever more consequential for accessing opportunities and getting ahead. The “income inequality” hypothesis proposes that intergenerational inequality—with respect to educational attainment, social mobility, and other socioeconomic outcomes—will increase as income inequality grows. Because this hypothesis shot to public attention with Krueger’s ( 1 ) discussion of the Great Gatsby curve, the proposition that high levels of income inequality have generated correspondingly high levels of intergenerational reproduction is now a staple of public and political discourse. Despite the prominence of this argument, the evidence in its favor is less overwhelming than might be assumed ( 2 ), and is largely limited to the empirical result that intergenerational income inheritance has increased in recent decades, at least in some analyses ( 3 , 4 ). Even this result has been contested and is far from widely accepted ( 5 ).

In this paper, we assess the plausibility of the income inequality hypothesis by examining changes over the past century in the income-based gaps in college enrollment and completion. This is a field in which descriptive evidence is key: Designs that would allow for convincing causal inference are in short supply, and where designs are available, the data are not. And yet most of the descriptive evidence in regard to the college level pertains only to recent decades, when both income inequality and collegiate inequalities have increased (refs. 6 – 8 ).

The trends through earlier decades of the century, within which the great U-turn in income inequality occurred, remain largely undocumented. To overcome this evidence deficit, we might be inclined to draw on evidence on other educational outcomes, such as test scores and years of schooling. Reardon’s analysis of family income test score gaps, for example, shows steadily rising gaps between cohorts born in the 1940s and those born in the present day (ref. 9 ; cf. ref. 10 ). But test scores are quite imperfectly correlated with educational attainment, and evidence from studies of inequalities in years of schooling would support different conclusions on trend. Hilger’s ( 11 ) analysis of long-term trends using Census data shows that there was a decline in the effects of parental income on child’s education between the 1940s and 1970s, while Mare ( 12 ) shows an increasing effect of family income on higher-level educational transitions for midcentury cohorts as compared to early-century cohorts. Taking these studies together, it is difficult to reach any firm conclusion about the income inequality hypothesis, as one might infer an increase, a decrease, or stability in collegiate inequalities during the midcentury, depending on which study is considered.

Extending the time series over the whole of the past century allows for a fuller assessment of the income inequality hypothesis, as the long-run historical series on income inequality exhibits a relatively complicated pattern, as opposed to the simple increase in the recent period. In much the same way as the magnitude of changes in income inequality could only be appreciated when considered in the long run, current levels of educational inequality must be evaluated and understood in full historical context ( 13 ). In a comprehensive extension of previous research on collegiate inequalities, we thus use all nationally representative data sources that we were able to locate and access. This strengthens the descriptive evidence that can be brought to bear upon the income inequality hypothesis.

In the following sections, we discuss the available data and the methods of analysis, and present our results on long-term trends in collegiate inequalities. We will focus on inequalities in completion of 4-year college, enrollment in 4-year college, and enrollment in any college (2- or 4-year). We will demonstrate an essential similarity in inequality trends across the range of collegiate outcomes. Although we will show that income inequality is strongly associated with inequalities at the college level, we will also highlight that it is not the only force at work.

College Enrollment and Completion in the Twentieth Century

The twentieth century was the first century in which education systems were widely diffused and, at least in principle, accessible to all social groups. The century witnessed substantial expansion at the college level: The college enrollment rate for 20- to 21-y-olds increased from around 15 % for the mid-1920s birth cohorts to almost 60 % for cohorts born toward the end of the century. * As Fig. 1 shows, rates of enrollment rose rapidly for cohorts born in the early century to midcentury, and flattened out and even declined for the midcentury birth cohorts, before resuming a steady increase for cohorts born in the later decades of the century.

Proportion of birth cohort enrolled in college ages 20 y to 21 y ( 14 ), and proportions completing 2- and 4-year college degrees, Current Population Survey March, Annual Social and Economic Supplement ( 15 ).

We see in Fig. 1 a stark reversal of the gender gap in college enrollment; for birth cohorts from the mid-1950s to mid-1990s, the proportion of women enrolled in college grew by around 30 percentage points, while the corresponding increase for men was just under 20 percentage points ( 16 , 17 ). The reversal occurred immediately after the rapid increase in enrollment rates observed for male birth cohorts at risk for service in the Vietnam War ( 16 ). A literature in economics has demonstrated that men born in the 1940s and 1950s were unusually likely to attend and graduate from college, although there is disagreement with respect to whether the observed increase in men’s college participation rates should be attributed to draft avoidance or to postservice GI Bill enrollments (ref. 18 ; cf. ref. 19 ).

Alongside trends in college enrollment, Fig. 1 presents rates of college completion by type of degree. While rates of completion of 2-year college are rather flat for cohorts born from the 1950s onward, rates of 4-year college completion have increased considerably. As the figure suggests, rates of 4-year college completion are highly correlated with rates of enrollment, but research shows that, over the past half-century, rates of college completion increased less sharply than rates of enrollment, because the college dropout rate increased ( 6 , 20 ).

Materials and Method

Although it is relatively straightforward to examine changes in rates of college enrollment and completion over time, it is rather less straightforward to examine income inequalities in collegiate outcomes across the span of the twentieth century, because data on parental income, college enrollment, and college completion are not routinely collected in government surveys. We must therefore piece together the trends in collegiate inequalities through the analysis of available sources of nationally representative data. We include results from the analysis of both cross-sectional surveys of adults and longitudinal surveys beginning with school-aged children, and, for a number of recent cohorts, we calculate estimates from tax data results in the public domain. Although this approach presents obvious challenges as regards comparability of data sources and measures, for much of the period that we cover, we have multiple estimates of collegiate inequalities for any given period of time. The datasets and their key characteristics are listed in Table 1 ; detailed descriptions of each dataset are included in SI Appendix .

Characteristics of the datasets included in the analysis

Add Health, National Longitudinal Study of Adolescent to Adult Health; ELS, Education Longitudinal Study; HSLS, High School Longitudinal Study.

The datasets cover cohorts born between 1908 and 1995, and it is only at the beginning and the end of the data series that our birth cohorts are represented by no more than one dataset. Although we aim to define cohorts according to year of birth, for some of the datasets we must construct quasi-cohorts based on age or grade, because year of birth was not recorded.

The biggest constraint that we face in analyzing income inequalities in collegiate attainment relates to gender. Data on the earlier birth cohorts come from the Occupational Changes in a Generation (OCG 1973) survey, which was administered in conjunction with the Current Population Survey ( 21 ). This survey was completed by men only, so we lack information on the educational attainment of women in the earliest birth cohorts. By presenting all results separately for men and women, patterns over time can be compared by gender.

The datasets were prepared to provide consistent measures of family income, college enrollment, and college completion. We produce simple binary variables that capture whether an individual completed a 4-year degree, whether an individual enrolled in (without necessarily completing) a 4-year degree program, and whether an individual enrolled in (without necessarily completing) a college program. Unfortunately, the tax data results pertain only to college enrollment per se, so we have fewer available data points for the analyses of 4-year completion and enrollment than for the analyses of enrollment in any college program. All samples are restricted to individuals who enrolled in high school, in order to maximize consistency across samples. In SI Appendix , we also include results for a smaller sample restricted to high school graduates ( SI Appendix , Fig. S6 ).

A more difficult variable to harmonize over time is family income. Although in some datasets family income is measured directly (e.g., annual net family income in dollars), in many of the available datasets family income is measured only as an ordinal variable. For these datasets, we employ the method used by Reardon ( 9 ) to calculate test score gaps from coarsened family income data; the method uses the proportions in each income category to assign an income rank to all of those in a given category, and income rank is then the explanatory variable in the analysis ( SI Appendix , SI Methods ).

We estimate logits predicting college enrollment and completion as a function of family income or income rank. Following Reardon ( 9 ), we fit squared and cubed terms to capture the nonlinear effects of income rank. Using the model, we estimate the enrollment and completion rates of those at the 90th percentile of family income and those at the 10th percentile. We choose the 90 vs. 10 comparison over other ways of defining inequality because it accords with past assessments and with the main source of trend in income inequality ( 9 ). † From these rates, we calculate log-odds ratios capturing, for example, the log-odds of completing a 4-year college degree for the 90 vs. 10 family income comparison.

We would be remiss if we did not note the difficulty in measuring family income reliably, particularly using one-shot measures, which are all that are available in almost all of the datasets that we analyze. Further worries might arise because some of the income measures are retrospective, or because the questions are asked of children, not parents. Although we would not minimize the danger of retrospection or of using children’s reports of family income, evidence suggests that child reports of parental socioeconomic characteristics are not substantially worse than parental reports of those characteristics ( 9 , 22 ). Furthermore, the types of errors that individuals make when reporting income appear to have changed very little over time ( 23 ), which is the key issue when mapping trend. To address concerns about the varying quality of the family income data, we multiply all log-odds ratios by 1 / r , where r is the estimated reliability of the family income measure (see SI Appendix , Table S5 for reliability estimates) ( 9 ).

We recognize that “researcher degrees of freedom” are of particular concern when presenting results from a large number of datasets ( 24 ). We provide additional results based on alternative specifications, in SI Appendix , and make our analysis code publicly available on Open Science Framework, https://osf.io/jxne5 .

The Great U-turn in Collegiate Inequality

We now examine collegiate inequalities for cohorts born between 1908 and 1995. Given data constraints, we are limited to examining inequalities over the whole period for men only, but we present results for women for a more limited range of birth cohorts.

In Fig. 2 we present, for the full male series, the estimated probabilities of completing 4-year college at the 90th and 10th percentiles of family income. ‡ We see in Fig. 2 that the increase in 4-year college degree attainment over the twentieth century was far from equally distributed across income groups. Men from the 90th percentile of family income were at the leading edge of the expansion; the figure shows a rapid increase in college completion rates through the 1940s birth cohorts, then a tailing off through the 1950s cohorts, followed by a further rapid increase for those cohorts born in the 1960s onward. In contrast, expansion at the bottom of the income distribution was more sluggish; 4-year college completion rates at the 10th percentile were less than 10 percentage points higher for cohorts born at the end of the century than for cohorts born at the beginning.

Probabilities of 4-year college completion at the 90th and 10th percentiles of family income, male birth cohorts, 1908–1986.

Fig. 2 shows that absolute differences in completion rates between income groups increased from the beginning to the end of the century. But this important result must be considered alongside changes over the century in the overall completion rate ( 12 ). Although the probability gap was small at the beginning of the century, the odds of college completion were around 7 times higher for the rich than for the poor, because the rich were able to secure a large proportion of the limited number of college slots. In relative terms, the poor born in the early century were more disadvantaged than their counterparts born in the 1960s, when 90 vs. 10 gaps in the probability of college completion were substantially larger. Although both probability gap and odds-ratio measures are informative, we focus from this point forward on odds-ratio measures of educational inequality, which are margin insensitive and thus feature relative—rather than absolute—advantage. But, in SI Appendix , we present probability plots for the three collegiate outcomes ( SI Appendix , Fig. S1 ), and include analyses based on probability gaps in SI Appendix , Table S3 . The key results hold for both types of analysis.

We plot, in Fig. 3 , the 90 vs. 10 log-odds ratios describing inequalities in collegiate outcomes for each of the datasets in our analyses, with trends estimated from generalized additive models (GAM). The GAMs are fitted to the plotted data points, with each point weighted by the inverse of the SE for the estimate. § In the earlier period covered by OCG, we fit the model to the estimates derived from analyses of single birth cohorts, but present point estimates representing groups of birth cohorts to show the consistency across these specifications. Confidence intervals are presented in SI Appendix , Fig. S2 ; figures showing 90 vs. 50 and 50 vs. 10 inequalities are included as SI Appendix , Figs. S3 and S4 .

The 90 vs. 10 log-odds ratios expressing inequality in 4-year completion, 4-year enrollment, and any college enrollment. ( Left ) Male birth cohorts, 1908–1995; ( Right ) female birth cohorts, 1951–1995.

We focus first on describing the trends for men, for whom we have results spanning the whole century. It is clear from Fig. 3 that the over-time trends are similar across the various collegiate outcomes and, further, that there is no simple secular trend for any of the outcomes under consideration. There are three key attributes of the trends that should be emphasized.

First, Fig. 3 shows that, toward the middle of the century, there was a great U-turn in collegiate inequality. Inequalities fell rapidly for cohorts born in the early to mid-1950s, then bottomed out until the mid-1960s, before ultimately rising steeply for cohorts born from the mid-1960s onward. The U-turn appears to be more pronounced for 4-year and “any college” enrollment than for completion of a 4-year degree, but it is present for all of the collegiate outcomes under consideration.

Had we measured collegiate inequalities in but a single dataset, we might be skeptical that our observed trend was on the mark and, in particular, that there was a rapid fall in inequality for the midcentury birth cohorts. But this trend is supported across all of the datasets from the period: OCG and National Longitudinal Study (NLS) Young Men show high inequality in the early 1950s; Panel Study of Income Dynamics (PSID), NLS72, and High School and Beyond (HS&B) pick up the lower inequality of the mid-1950s to the mid-1960s; and the subsequent uptick in inequality is captured in PSID, the school cohort surveys, and the National Longitudinal Studies of Youth (NLSY79&97). Indeed, Fig. 3 demonstrates that there is great consistency across a large number of different data sources. ¶ At the trough, inequality in 4-year college completion was reduced to a log-odds ratio of around 1.5, indicating that, even in this low-inequality period, the odds of those at the 90th income percentile completing a 4-year college degree were almost 4.5 times greater than the equivalent odds for those at the 10th percentile. Inspection of SI Appendix , Fig. S3 suggests that the U-turn observed in Fig. 3 is largely driven by changes in the top half of the income distribution: the U-turn is rather more pronounced for the 90 vs. 50 comparison than for the 50 vs. 10 comparison.

Second, if skepticism about a midcentury fall in collegiate inequality were to be sustained, suspicion would also have to fall upon all currently accepted results on over-time trends, which demonstrate a substantial increase in inequalities in college enrollment and completion between cohorts born in the midcentury and late century. If we were to impose a simple linear smooth on the century-long data series, this would indicate relatively modest increases in collegiate inequalities over the period taken as a whole (see dashed lines, Fig. 3 ). # Again, because the trends are mapped using multiple datasets, we are confident that the pattern of a U-turn in collegiate inequality is supported.

Third, any evidence of a U-turn must bring to mind the pattern of income inequality over the past century. As Piketty and Saez ( 27 ) described, toward the middle of the twentieth century, the share of income going to the top 10% rapidly declined, before rising again over the later decades of the century. The U-turn in collegiate inequality mimics this trend, although it is notable that, insofar as we see similarity in patterns of income inequality and collegiate inequalities, it is income inequality around year of birth that appears to matter most. But, despite the obvious similarities, there is at least one clear divergence in the pattern of collegiate inequality and income inequality: The U-turn in collegiate inequality comes very late. Income inequality begins to fall in the early 1940s, but inequalities in enrollment and completion begin to decline only for cohorts born in the mid-1950s. Men born in the mid-1940s onward were not just born into a period of low inequality, but they spent most of their formative years in a low-inequality society. Despite this, the evidence shows that collegiate inequality increased substantially for the cohorts born in the 1940s and early 1950s; the log-odds ratios describing inequality are increased by around a third over this short period.

Some of the same key features are visible in the results for women, shown in Fig. 3 , Right , although we only have access to data for women born after 1950. We see a basic similarity with the men’s analyses from the mid-1950s birth cohorts onward: Collegiate inequalities are relatively flat for the 1950s to 1960s birth cohorts, and increase for women born in the 1970s and onward. Just as with men, toward the end of the period we see flat and even declining inequalities in enrollment and completion. There are perhaps some subtle differences in the pattern by gender—the upturn in collegiate inequality begins, for example, several years later for women than for men—but we have little evidence here to support a conclusion of substantial difference in inequality for men and women over this period.

There is one notable difference between the men’s and women’s results, relating to the period when trends in male collegiate inequality substantially diverged from trends in income inequality. This exceptional period appears to be exceptional for men, but not for women. Although we cannot track collegiate inequalities for women across the whole midcentury period, the first data points in the female data series (NLS Young Women: 1951–1953 birth cohorts) are lower than the nearby estimates for men (NLS Young Men: 1949–1951 birth cohorts). ** This period of divergence between collegiate inequality and income inequality coincides with the period that we identified above as holding special consequences for men’s educational attainment: Men born in the 1940s and early 1950s were subject to the threat of military service in the Vietnam War.

There are no cohort studies of women that would allow us to compare male and female inequalities in college enrollment and completion throughout this period. We do, however, have access to data on men who fathered children who were at risk for service during the Vietnam War: The NLS Older Men survey can be used to track collegiate inequalities for the children of men who were aged 45 y to 59 y in 1966. The structure of this dataset is somewhat different from the datasets underlying our time series, but we nevertheless find confirmation, in Fig. 4 , that male and female inequalities diverged in the Vietnam years.

The 90 vs.10 log-odds ratios expressing inequality in 4-year college completion, 4-year enrollment, and any college enrollment, men and women born 1935–1943 and 1944–1952, NLS-Older Men data.

In the pre-Vietnam period, male and female collegiate inequalities were of similar magnitude. The log-odds ratio for 4-year enrollment, for example, was 2.3 for men (95% CI: 1.5, 3.1), as compared to 2.4 for women (1.7, 3.2). But, for the birth cohorts at risk for serving in Vietnam, the male log-odds ratio increased slightly, to 2.5 (1.8, 3.2), while inequality fell substantially for women, to 1.4 (0.8, 2.0) (see SI Appendix , Fig. S8 for a figure with CIs). These results provide support for the claim that men’s collegiate inequality was substantially and artificially raised relative to expected levels during this period because of the Vietnam War. Unfortunately, our data are not well-suited to evaluating why male and female collegiate inequality differed in the Vietnam period. But some evidence can be brought to bear on this question by comparing preservice and postservice inequalities in college participation for the men in OCG ( SI Appendix , Fig. S9 ). These data are more consistent with a draft-induced increase in male collegiate inequality than with a GI Bill-induced increase. ††

Bringing the results in Fig. 4 together with what is known about college enrollment and completion patterns during the Vietnam War period, it seems likely that the disproportionate increase in men’s college participation rates observed in Fig. 1 was achieved, at least in part, through a gender-specific change in the effect of family income on college enrollment and completion.

The Association between Income Inequality and Collegiate Inequality.

We now present a formal statistical test of the strength of the association between income inequality and collegiate inequality. We regress the log-odds for collegiate inequalities on income inequality, as measured through the share of wages going to the top 10% ( 27 ). ‡‡ In addition to the income inequality variable, for the full male series (1908–1995), we fit a “Vietnam effect,” with a dummy variable that isolates the cohorts at risk from the draft lotteries (i.e., 1944–1952 birth cohorts). We fit models to the full male series (1908–1995 birth cohorts), a compressed male series (1952–1995 birth cohorts), and the female series (1951–1995 birth cohorts). A full regression table with coefficients and standard errors is included as SI Appendix , Table S4 . §§ In Fig. 5 , we present estimates of the predicted increase in the log-odds ratios for an eight percentage point increase in the share of wages going to the top 10%; this increase is equivalent to the “takeoff” in income inequality that occurred between the midcentury and the 1990s. ¶¶

Predicted increase in collegiate inequality log-odds ratios associated with the top 10%’s share of wages increasing by 0.08 (equivalent to the takeoff in income inequality); 90 vs. 50 (dark gray), 50 vs. 10 (light gray), and 90 vs. 10 (total) comparisons.

The regression coefficients describing the associations between income inequality and 90 vs. 10 collegiate inequalities can be straightforwardly decomposed into two parts: an association between income inequality and the 90 vs. 50 log-odds ratio, and an association between income inequality and the 50 vs. 10 log-odds ratio. In Fig. 5 , the total height of each bar represents the predicted increase in the 90 vs. 10 log-odds ratio for an eight percentage point increase in income inequality, while the dark and light gray bars show the predicted increases in the 90 vs. 50 and 50 vs. 10 log-odds ratios, respectively.

Examining first the results for the 90 vs. 10 comparison, we see confirmation of a relatively strong association between income inequality and collegiate inequality over the full sweep of the twentieth century. For women, for example, the model predicts that an increase in income inequality equivalent to that observed in the takeoff period would increase the 90 vs. 10 log-odds ratio by around 1 for 4-year enrollment and completion, and by around 1.3 for enrollment in any college. Although there is variation in the strength of the association for the different outcome measures, the income inequality effects are large and positive in all of the analyses, indicating substantial support for the income inequality hypothesis.

Given that the takeoff in income inequality was largely characterized by the top of the income distribution moving away from the middle and bottom of the distribution, the income inequality hypothesis would predict larger effect sizes for the 90 vs. 50 comparison than for the 50 vs. 10 comparison. When we decompose the 90 vs. 10 results into 90 vs. 50 and 50 vs. 10 components, we see precisely this result. The income inequality effects for the 90 vs. 50 comparisons in all cases outweigh those for the 50 vs. 10 comparisons, particularly in the analyses of 4-year college enrollment and completion.

But the results also provide grounds for exercising caution when interpreting differences in effect sizes across the models, as the effect sizes in the full and compressed male series are more similar for the “any college” analyses than for the 4-year analyses, where the sample sizes are smaller. Even when analyzing all available datasets and exploiting the full range of variation in income inequality over the century, our statistical power is limited. This is even more clear when we extend the models summarized in Fig. 5 to include additional macro-level regressors that social scientists have previously used to predict inequalities at the college level. These additional variables include the economic returns to schooling, which are assumed to influence individual decisions about whether or not to invest in college education ( 33 ), and the high school graduation rate, which has been shown to influence educational expansion at the college level ( 34 ). As shown in SI Appendix , Table S1 , estimates from these models are more volatile, particularly for women.

The volatility arises because some of our analyses are, like past analyses, limited to more recent cohorts in which the takeoff assumes a monotonically increasing form. This makes it difficult to adjudicate between the large number of monotonically increasing potential causes. An important advantage of our full-century approach is that it reaches back to a time in which these competing causes did not always move together. In Fig. 6 , we present the results of a simulation exercise, in which we run 1,000 regressions for a range of different model specifications on the full and compressed male series, with each regression including a new variable containing random numbers drawn from a normal distribution ( μ = 0; σ = 1). We examine the stability of the income inequality effects with respect to inequality in college enrollment, for which we have the largest number of data points. We add to the basic model in Fig. 5 controls for time, either in the form of 1) a linear effect of year or 2) dummies for decades, and measures of the returns to schooling ( 33 , 35 , 36 ) and the high-school graduation rate ( 34 , 37 ).

Predicted income inequality effects (coefficients × 0.08) from 1,000 regressions of 90 vs. 10 inequality in “any college” enrollment on income inequality and random number variables, for various model specifications, for full and compressed series, men only. Models: 1, Inequality; 2, Inequality+year; 3, Inequality+controls; 4, Inequality+controls+year; and 5, Inequality+controls+decade.

As Fig. 6 shows, the income inequality effects estimated for the full male series are robust to the inclusion of other potential confounding variables. But Fig. 6 also highlights the extent to which a proper evaluation of the income inequality hypothesis requires researchers to exploit all of the available data. Although the bivariate analysis shows a similar effect of the income inequality variable in both the full and compressed series, the effects are a good deal more volatile in the more highly parameterized models in the compressed relative to the full series. *** The substantive implication of this analysis is clear: It is only with the full data series that we obtain relatively precise and reliable estimates of the association between inequality in collegiate outcomes and income inequality.

We have examined descriptive evidence on the association between inequality in collegiate attainment and income inequality over the past century. Although there has been much recent interest in the income inequality hypothesis, it has been difficult to make headway because commonly used datasets pertain only to recent decades, when income inequality was increasing. We have thus proceeded by reaching back to the very beginning of the twentieth century, assembling all of the available datasets, and harmonizing the variables in these datasets.

The results show that collegiate inequalities and income inequality are, in fact, rather strongly associated over the twentieth century. Just as with income inequality, we see evidence of a U-turn in 90 vs. 10 collegiate inequality, and evidence of a substantial takeoff in collegiate inequalities in recent decades. When we examine trends in 90 vs. 50 and 50 vs.10 inequalities, we find that the 90 vs. 50 trends mirror the 90 vs. 10 results. Taken together, our results offer solid descriptive support for the income inequality hypothesis.

Inequalities in collegiate attainment increased hand in hand with the expansion of college education in the United States. Rates of college enrollment and completion were higher at the end of the century than they had been at any time in the preceding hundred years, and yet, for these birth cohorts, we see substantial inequalities, as captured in both percentage point gap and odds ratio measures. In point of fact, the only time during the twentieth century for which we observe a reduction in educational inequality is during the period when expansion at the college level had paused. Although the counterfactual is obviously not observable, these results emphasize the importance of attending to the distribution of college opportunities in addition to overall levels of attainment. These distributional questions will take on even greater significance in the context of the economic and social crisis engendered by coronavirus disease 2019, a crisis that is likely to have enduring effects on both the distribution of income and access to the higher education sector.

Our analyses are not well suited to evaluating the mechanisms generating the association between income inequality and collegiate inequalities. However, given the pattern of collegiate inequality across the century, we suspect that a mechanical effect is likely to be responsible. If money matters, as we know it does, and growing income inequality delivers more money to the top, then, all else being equal, these additional dollars would in themselves produce growing inequality in college enrollment and completion. The mechanical effect is therefore a parsimonious account of the trend that we see here ( 8 ). That the over-time associations are substantially stronger for the 90 vs. 50 comparison as compared to the 50 vs. 10 comparison provides further suggestive evidence in this regard. Nevertheless, there is a period for which we undoubtedly hypothesize an increase in the relational effect of income: the Vietnam War. For the war to lead to increased collegiate inequality, the effect of income on educational attainment would have to increase, particularly given that income inequality was low and stable for these birth cohorts.

Whatever the mechanisms may be, the key descriptive result is that, over the course of the twentieth century, a grand U-turn in collegiate inequality occurred. Cohorts born in the middle of the century witnessed the lowest levels of inequality in college enrollment and completion seen over the past hundred years. Contemporary birth cohorts, in contrast, are experiencing levels of collegiate inequality not seen for generations.

Supplementary Material

Acknowledgments.

We thank David Cox, David Grusky, and Florencia Torche for their detailed comments on earlier versions of this paper, and also Raj Chetty, Maximilian Hell, Robb Willer, the Cornell Mobility Conference, the Stanford Inequality Workshop, the Stanford Sociology Colloquium Series, and University of California, Los Angeles’s California Center for Population Research seminar for useful suggestions. Additionally, we thank Stanford’s Center for Poverty and Inequality, Russell Sage Foundation and Stanford’s United Parcel Service (UPS) Fund for research funding, Stanford’s Institute for Research in the Social Sciences for secure data room access, and the American Institutes for Research for data access. We are grateful to the editor and reviewers for their helpful and productive suggestions.

The authors declare no competing interest.

This article is a PNAS Direct Submission. E.G. is a guest editor invited by the Editorial Board.

Data deposition: Code for data analysis is archived on Open Science Framework ( https://osf.io/jxne5 ).

*Throughout this paper, we use the term “college” as a shorthand for “2- or 4-year college.”

† We also include results based on comparing income quartiles in SI Appendix , Fig. S5 .

‡ The probabilities are estimated from the logit model, and we fit a GAM to establish trend. See SI Appendix , SI Methods for more details.

§ We determine the appropriate number of degrees of freedom for the trend lines by fitting a series of GAMs and comparing model fit (using the Akaike Information Criterion). For the analysis of college enrollment for male birth cohorts, we use the stepwise model builder in R’s gam package to find the best-fitting model ( 25 , 26 ). As we have fewer point estimates in the other analyses, the stepwise approach is less reliable, and we therefore choose smoothing parameters that provide a reasonable (and conservative) summary of the trend.

¶ It is also clear that some datasets are outliers from the trend. It is not surprising to see variation across samples, and we highlight this variation only because it illustrates a potential danger of using but one or two datasets to establish a trend. The estimates for National Education Longitudinal Study (NELS) (1974), for example, are substantially higher than the surrounding estimates based on one-shot income measures, and there is a surprising degree of cross-cohort volatility in the PSID estimates.

# The linear trend is strongest for 4-year completion, and weakest for enrollment in 4-year college. For all collegiate outcomes, the GAM offers a significant improvement in fit over the simple linear model.

**It would be possible to track male and female educational inequality with respect to parental education or socioeconomic index scores (SEI) ( 28 ), but the sample sizes are, unfortunately, too small for a detailed analysis of gender differences in educational attainment by birth cohort. This approach is also unattractive given that parental education, parental income, and SEI were only weakly correlated in this period ( 29 ).

†† Note that, while previous research has suggested that high-socioeconomic status (SES) individuals might have taken advantage of the GI Bill to a greater extent than low-SES individuals ( 30 ), SI Appendix , Fig. S9 provides little evidence that collegiate inequality was substantially affected. See SI Appendix for further discussion of this point.

‡‡ We choose the wages measure because, for the bottom of the income distribution, wages are a more important component of income than the types of income included in the alternative measures (e.g., capital gains). We measure wage inequality in year of birth. Surprisingly, given the prominence of the income inequality hypothesis, there is not yet adequate guidance in the literature as to the age at which income inequality most influences outcomes, although in the “money matters” literature there has been particular emphasis on the prenatal period, the postnatal period, and early childhood as the lifecourse moments when money matters most ( 31 , 32 ).

§§ In the 4-year analyses, we weight the data by the inverse of the standard errors underlying the estimates. In the analysis of any college enrollment, we do not weight the data, as this data series includes the tax data estimates. Given the size of the samples underlying these estimates, weighting would allow the relationship that pertains in the tax data for cohorts born in the 1980s and 1990s to have a disproportionate influence on the estimated century-long relationship between income inequality and inequality in college enrollment.

¶¶ The estimates in Fig. 5 are obtained by multiplying the income inequality coefficients in SI Appendix , Table S4 by 0.08.

***See SI Appendix , Fig. S10 for similar figures for 4-year enrollment and completion.

This article contains supporting information online at https://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1907258117/-/DCSupplemental .

Data Availability.

The analysis code and auxiliary data required to produce the figures and tables in this paper can be accessed at https://osf.io/jxne5 . Code to produce estimates for each of the individual datasets (see Table 1 ) is also provided. Details on how to access these datasets are provided in SI Appendix (most datasets are available for download upon registration with the data provider, while others are accessible only with a restricted use license from the National Center for Education Statistics).

• 1. Krueger A. B., The Rise and Consequences of Inequality in the United States. Speech at the Center for American Progress (Washington, DC, 12 January 2012). https://cdn.americanprogress.org/wp-content/uploads/events/2012/01/pdf/krueger.pdf . Accessed 8 July 2020.
• 2. Neckerman K. M., Torche F., Inequality: Causes and consequences. Annu. Rev. Sociol. 33, 335–357 (2007). [ Google Scholar ]
• 3. Davis J., Mazumder B., The decline in intergenerational mobility after 1980. FRB of Chicago Working Paper, WP-2017-5 (2017). https://www.chicagofed.org/publications/working-papers/2017/wp2017-05 . Accessed 8 July 2020.
• 4. Mitnik P. A., Cumberworth E., Grusky D. B., Social mobility in a high-inequality regime. Ann. Am. Acad. Polit. Soc. Sci. 663, 140–184 (2016). [ Google Scholar ]
• 5. Chetty R., Hendren N., Kline P., Saez E., Turner N., Is the United States still a land of opportunity? Recent trends in intergenerational mobility. Am. Econ. Rev. 104, 141–147 (2014). [ Google Scholar ]
• 6. Bailey M. J., Dynarski S. M., “Inequality in postsecondary education” in Whither Opportunity, Duncan G. J., Murnane R. J., Eds. (Russell Sage Foundation, 2011), pp. 117–132. [ Google Scholar ]
• 7. Ziol-Guest K. M., Lee K. T. H.. Parent income–based gaps in schooling: Cross-cohort trends in the NLSYs and the PSID. AERA Open 2, 1–10 (2016).26942210 [ Google Scholar ]
• 8. Duncan G. J., Kalil A., Ziol-Guest K. M., Increasing inequality in parent incomes and children’s schooling. Demography 54, 1603–1626 (2017). [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ]
• 9. Reardon S. F., “The widening academic achievement gap between the rich and the poor: New evidence and possible explanations” in Whither Opportunity, Duncan G. J., Murnane R. J., Eds. (Russell Sage Foundation, 2011), pp. 91–116. [ Google Scholar ]
• 10. Hanushek E. A., Peterson P. E., Talpey L. M., Woessmann L., The unwavering SES achievement gap: Trends in US student performance (Rep. W25648, National Bureau of Economic Research, 2019).
• 11. Hilger N. G., The great escape: Intergenerational mobility in the United States since 1940 (Rep. W21217, National Bureau of Economic Research, 2015).
• 12. Mare R. D., Change and stability in educational stratification. Am. Socio. Rev. 2, 72–87 (1981). [ Google Scholar ]
• 13. Hirschman D., Rediscovering the 1 % : Economic expertise and inequality knowledge. SocArXiv:10.31235/osf.io/4dws8 (13 July 2016).
• 14. Census Bureau , CPS historical time series tables on school enrollment (2018). https://www.census.gov/content/census/en/data/tables/time-series/demo/school-enrollment/cps-historical-time-series.html . Accessed 8 July 2020.
• 15. Flood S., King M., Rodgers R., Ruggles S., Warren J. R., Integrated Public Use Microdata Series (IPUMS), Current Population Survey: Version 6.0 . IPUMS, 2018. https://cps.ipums.org/cps/ . Accessed 8 July 2020.
• 16. Goldin C., Katz L. F., Kuziemko I., The homecoming of American college women: The reversal of the college gender gap. J. Econ. Perspect. 20, 133–156 (2006). [ Google Scholar ]
• 17. DiPrete T. A., Buchmann C., The Rise of Women: The Growing Gender Gap in Education and What It Means for American Schools (Russell Sage Foundation, 2013). [ Google Scholar ]
• 18. Card D., Lemieux T., Going to college to avoid the draft: The unintended legacy of the Vietnam War. Am. Econ. Rev. 91, 97–102 (2001). [ Google Scholar ]
• 19. D Angrist J., Chen S. H., Schooling and the Vietnam-era GI Bill: Evidence from the draft lottery. Am. Econ. J. Appl. Econ. 3, 96–118 (2011). [ Google Scholar ]
• 20. Bound J., Lovenheim M. F., Turner S., Why have college completion rates declined? An analysis of changing student preparation and collegiate resources. Am. Econ. J. Appl. Econ. 2, 129–57 (2010). [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ]
• 21. Featherman D. L., Hauser R. M., Opportunity and Change (Academic, 1978). [ Google Scholar ]
• 22. Hauser R. M., Andrew M., Reliability of student and parent reports of socioeconomic status in NELS-88. University of Wisconsin-Madison working paper, (2007). https://pdfs.semanticscholar.org/34da/152939d0497e7809c50b7d99d117fb427960.pdf . Accessed 8 July 2020.
• 23. Moore J. C., Stinson L. L., Welniak E. J., Income measurement error in surveys: A review. J. Off. Stat. 16, 331–362 (2000). [ Google Scholar ]
• 24. Simmons J. P., Nelson L. D., Simonsohn U., False-positive psychology: Undisclosed flexibility in data collection and analysis allows presenting anything as significant. Psychol. Sci. 22, 1359–1366 (2011). [ DOI ] [ PubMed ] [ Google Scholar ]
• 25. Hastie T. J., “Generalized additive models” in Statistical Models in S, Chambers J. M., Hastie T. J., Eds. (Taylor and Francis Group, 2017), pp. 249–307. [ Google Scholar ]
• 26. Hastie T. J., GAM: Generalized additive models, R package version 1.16, https://cran.r-project.org/web/packages/gam/index.html . Accessed 8 July 2020. [ DOI ] [ PubMed ]
• 27. Piketty T., Saez E., Income inequality in the United States, 1913–1998. Q. J. Econ. 118, 1–39 (2003). [ Google Scholar ]
• 28. Hout M., Janus A., “Educational mobility in the United States since the 1930s” in Whither Opportunity, Duncan G. J., Murnane R. J., Eds. (Russell Sage Foundation, 2011), pp. 165–186. [ Google Scholar ]
• 29. Duncan O. D., Featherman D. L., Duncan B., Socioeconomic Background and Achievement (Seminar, 1972). [ Google Scholar ]
• 30. Stanley M., College education and the midcentury GI Bills. Q. J. Econ. 118, 671–708 (2003). [ Google Scholar ]
• 31. Duncan G. J., Magnuson K., Votruba-Drzal E., Boosting family income to promote child development. Future Child. 24, 99–120 (2014). [ DOI ] [ PubMed ] [ Google Scholar ]
• 32. Hoynes H., Whitmore Schanzenbach D., Almond D., Long-run impacts of childhood access to the safety net. Am. Econ. Rev. 106, 903–934 (2016). [ Google Scholar ]
• 33. Goldin C., Katz L. F., The Race between Education and Technology: The Evolution of US Educational Wage Differentials, 1890 to 2005 (Harvard University Press, 2009). [ Google Scholar ]
• 34. Heckman J. J., LaFontaine P. A., The American high school graduation rate: Trends and levels. Rev. Econ. Stat. 92, 244–262 (2010). [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ]
• 35. Acemoglu D., Autor D., “Skills, tasks and technologies: Implications for employment and earnings” in Handbook of Labor Economics, Ashenfelter O. C., Ed. (Elsevier, 2011), vol. 4, pp. 1043–1171. [ Google Scholar ]
• 36. Valletta R. G., “Recent flattening in the higher education wage premium: Polarization, skill downgrading, or both?” in Education, Skills, and Technical Change: Implications for Future US GDP Growth, Hulten C. R., Ramey V. A., Eds. (University of Chicago Press, 2018), pp. 313–342. [ Google Scholar ]
• 37. Snyder T. D., de Brey C., Dillow S. A., Digest of Education Statistics 2015 (Pub. 2016-014, National Center for Education Statistics, 2016). [ Google Scholar ]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Data availability statement.

• View on publisher site
• PDF (1.1 MB)
• Collections

Similar articles

Cited by other articles, links to ncbi databases.

• Download .nbib .nbib
• Format: AMA APA MLA NLM

Add to Collections

• DOI: 10.59613/global.v2i8.249
• Corpus ID: 272519268

The Effectiveness of Blended Learning, Digital Literacy Programs, and Teacher Training on Student Outcomes in 2024

• Hasan Basri
• Published in Global International Journal… 6 August 2024
• Education, Computer Science
• Global International Journal of Innovative Research

16 References

The effectiveness of online and blended learning: a meta-analysis of the empirical literature, digital competence and digital literacy in higher education research: systematic review of concept use, a meta-analysis of blended learning and technology use in higher education: from the general to the applied, can we teach digital natives digital literacy, learning online: what research tells us about whether, when and how, what is technological pedagogical content knowledge (tpack), effective teacher professional development, the difference between emergency remote teaching and online learning, focus groups, related papers.

Showing 1 through 3 of 0 Related Papers

• Open access
• Published: 16 October 2024

Burnout level evaluation of undergraduate dental college students at middle eastern university

• Muhammad Qasim Javed 1 ,
• Zaina Ahmad 2 ,
• Muhammad Muhammad 2 ,
• AbdulAziz Binrayes 3 ,
• Iffat Niazi 4 ,
• Shazia Nawabi 5 , 7 ,
• Ayman M. Abulhamael 6 &
• Syed Rashid Habib 3  

BMC Medical Education volume  24 , Article number:  1155 ( 2024 ) Cite this article

Metrics details

Pressure faced by dental students from academic activities, clinical skills training, and patient care may lead to high stress and potential burnout, negatively impacting their well-being and patient safety.

The study aimed to explore the burnout level of dental students at Qassim University, Saudi Arabia and to identify the factors that are associated with the level of burnout.

The descriptive cross-sectional study was carried out at Dental College, Qassim University, Saudi Arabia, from January to February 2024. The study was conducted to evaluate burnout levels among first to final-year undergraduate dental students enrolled at Qassim University using a pre-validated questionnaire; Burnout Clinical Subtype Questionnaire, via Google Forms. Analysis was performed with IBM SPSS-24, utilizing descriptive statistics and non-parametric tests, Mann-Whitney-U and Kruskal-Wallis. A multiple linear regression analysis was conducted to predict Burnout from demographic and academic factors.

151 participants responded to the survey, with 49.7% male and 50.3% female. Results showed that burnout scores increased significantly (p < 0.05) with age and decreased family support. Males had lower burnout levels in comparison to their counterparts. The survey had three domains: Overload, Neglect, and Development, with ‘Overload’ having the highest mean score of 17.79 and a median score of 17.00. Age showed significant difference ( p  < .05) in the burnout scores amongst the groups across all three domains. 70.9% of the participants agreed that they invested an unhealthy amount of time towards their studies. The multiple regression model statistically significantly predicted Burnout, F (21, 129) = 2.190, p  = .004, adj. R 2  = 0.143. Gender was significant predictor, with female students reporting lower burnout scores compared to males (B=-5.633, p  = .012). Family support also emerged as significant, with students reporting merely good family support showing higher burnout scores compared to those reporting very good family support (B = 6.147, p  = .009).

Factors like age, family support and gender were found to have a significant with burnout levels and its domains. The study highlighted overload as the main contributing the most towards burnout in dental students of Qassim University, emphasizing the importance of tailored interventions to promote student well-being.

Peer Review reports

Introduction

Dental students face the challenge of balancing their rigorous academic curriculum, clinical skills development, and commitment to providing excellent patient care. However, this pursuit of excellence comes at a price. The relentless nature of these demands often leads to a state of overload, wherein the pressures of academic performance, mastering complex procedures, and staying up to date with technological advancements create a persistent sense of overwhelming stress. The dynamic nature of dentists’ careers, reflected in dental students’ experiences, acts as a catalyst for this stress [ 1 ].

The term ‘burnout’, has been described by Maslach and Leiter as a syndrome characterized by emotional exhaustion and cynicism in individuals engaged in professions involving substantial interpersonal interactions or “people-work” [ 2 ]. This state is an outcome of chronic emotional and interpersonal stress, manifesting itself in psychological and physical exhaustion [ 3 ]. Previously burnout has been measured on the MBI scale, particularly in occupational settings. Developed by Christina Maslach and Susan E. Jackson, this inventory measures three dimensions of burnout: emotional exhaustion, depersonalization, and reduced personal accomplishment [ 4 ].

The Burnout Clinical Subtype Questionnaire (BCSQ-12-SS) introduced by Montero-Marín et al., in 2011 was designed specifically to measure educational burnout among students of clinical fields including dental students [ 5 , 6 ]. The questionnaire evaluates burnout amongst students based on three domains; overload (based on feelings of exhaustion), lack of development (sensing a dearth of growth opportunities), and neglect (towards the job).

Academic burnout which stems from immense academic pressures, can result in a decline in enthusiasm and interest towards the field of study (academic disinterest) [ 7 ]. A decline in enthusiasm may also result from a lack of opportunities for personal growth, leading to potential interest in alternative occupations that could offer chances to enhance an individual’s skill set.

Within these demands, the phenomenon of overload emerges. The pursuit of professional excellence often comes at the expense of neglecting crucial aspects of personal well-being. Psychological influences of burnout include but are not limited to depression, loss of motivation, and suicide ideation [ 8 , 9 ]. The psychological influence of burnout may translate into physical impacts [ 10 ]. It has been put forward in the literature that burnout may pose serious health risks to individuals e.g., diabetes, and cardiovascular diseases. Furthermore, it is also associated with physical upset such as back pain, neck strain, and muscle fatigue [ 11 , 12 ].

Estimating the levels of burnout within this context is important, as it affects the well-being of the students as well as the dental healthcare system in the region. Unaddressed burnout leads to serious consequences, extending beyond the physical and mental health implications for students, potentially influencing the quality of patient care [ 13 , 14 ]. Therefore, recognizing and effectively managing burnout among dental students is pivotal not only for safeguarding their health but also for rendering quality health care to the patients. Considering this, the objective of this study is to evaluate the burnout level among dental students at a dental school in Saudi Arabia using the Burnout Clinical Subtype Questionnaire ((BCSQ-12-SS). To the best of our knowledge this is the first study that has used BCSQ-12SS to assess the burnout level in Saudi Dental Students.

The manuscript was prepared in accordance with the STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) checklist, which provides a framework for ensuring comprehensive and transparent reporting of cross sectional studies [ 15 ].

Ethics approval

The study adhered to ethical guidelines ensuring the anonymity and confidentiality of participants with restricted data accessibility to research team, only. The research protocol was approved by the ethical committee at Dental College, Qassim University (Approval no: 24-8-03).

Research question

What is the burnout level of students enrolled at under graduate Dental College of Qassim University?

Study design

The research has cross-sectional design.

It was conducted at Dental college of Qassim University, KSA.

Participants and sampling technique

A non-probability purposive sampling technique was utilized to recruit the participants. The target population were undergraduate dental students ( n  = 166) enrolled at Qassim University. All of the students that fit the inclusion criteria i.e. every student enrolled at Qassim University, Dental College between first year to final year, were approached. The students were informed that participation in the study is voluntary.

Sample size calculation

The sample size for the study was determined using the Qualtrics sample size calculator [ 16 ], with a desired confidence level of 95% and an error margin of 5%. Based on these parameters and a total student population of 166, a sample size of 133 was calculated as acceptable sample size.

Research instrument

The BCSQ-12SS utilized in the study was adapted from previous research conducted in Spain 5 . The questionnaire has 12-items designed specifically to measure educational burnout among clinical field students. The 12-items are subdivided into three domains 5 . The three domains are as follow: 1) “Overload” (Domain 1: Items 1;4;7;10) is associated with students who may risk their health or work-life balance in pursuit of academic success; 2) “Lack of development” (Domain 2: Items 2;5;8;11) ensues when individuals feel the dearth of opportunities for personal growth, urging them to switch to new jobs where they can boost their skills. This phenomenon is mostly seen in jobs where an individual is expected to perform repetitive tasks; 3) ‘Neglect’ (Domain 3: Items 3;6;9;12) involves ignoring problems and feeling ineffective, marked by a sense of feeling unappreciated and ignoring responsibilities. 5 Each questionnaire item was structured based on a seven-point Likert scale, with ‘1’ indicating “Totally disagree”, ‘2’ indicating “Strongly disagree”, ‘3’ indicating “Disagree”, ‘4’ indicating “Undecided”, ‘5’ indicating “Agree”, ‘6’ indicating “Strongly agree”, ‘7’ indicating “Totally agree. Additionally, 10 demographic survey questions were included at the start of the survey, which allowed the subjects to self-report this information. These included information about gender, age, perceived family support, year of education, marital status, distance from home, living situation, studying hours and whether they received scholarship or not.

Data collection

The questionnaire was administered electronically through Google Forms link to ensure efficient and standardized data collection. Before administering the questionnaire, informed consent was obtained from each participant. The participants were invited to participate in the study through their university email. The invites were sent by the students’ affairs office who also acted as a gatekeeper. The invites included a participant information sheet explaining the purpose and scope of the study. Three reminders were sent by the students’ affairs office at an interval of week.

Data analysis

The collected quantitative data from the completed questionnaires was analyzed by IBM SPSS version-26 software [ 17 ]. The defined outcome was Burnout level which was presented in scalar scores. The possible score range for each participant was from 12 to 84. The higher the score, the greater the burnout levels. The burnout scores were further divided into 3 subdomains.i.e. Overload, Neglect and Development as documented in the previous research 5 , which were individual components contributing to the overall burnout. The descriptive analysis was conducted for participants’ characteristics of entire sample. The Shapiro-Wilk test was performed to assess the normality of the data. Subsequently, non-parametric tests, specifically the Mann-Whitney U test and the Kruskall-Wallis test, were utilized for comparing the overall burnout score and subdomain-wise burnout score of participants based on different demographic characteristics. Moreover, percentage distribution of participants’ responses to all items were calculated as disagree (combined totally disagree, strongly disagree and disagree responses), neutral (Undecided) and agree (combined totally agree, strongly agree, agree responses).

A multiple linear regression analysis was conducted to predict Burnout from demographic and academic factors including gender, age, year of education, distance from home, place of living, family support, studying hours, subjects failed, marital status, and scholarship status.

The Enter method was used for variable selection, with all predictors entered into the equation simultaneously. This approach was chosen to assess the combined effect of all variables on burnout, as well as the individual contribution of each predictor while controlling for others, thus adjusting for potential confounding factors.

Prior to conducting the regression, assumptions were checked. Linearity was assessed using partial regression plots and a plot of studentized residuals against predicted values. Independence of residuals was examined using the Durbin-Watson statistic. Multicollinearity was checked using tolerance values. The presence of influential outliers was examined using studentized deleted residuals, leverage values, and Cook’s distance. Normality of residuals was assessed using a Q-Q plot.

The model’s overall fit was assessed using the F-test, R-squared, and adjusted R-squared values. The statistical significance of individual predictors was determined using t-tests for each regression coefficient. A significance level of α < 0.05 was used for all statistical tests.

The study consisted of 151 participants, with 49.7% being male and 50.3% being female. Out of all the participants, 59.6% were aged between 22 and 24 years, with the highest percentage (28.5%) being in their 5th year of education. A large majority (78.1%) of the participants lived within 75 km of their university, and almost all of them (91.4%) lived with their parents.

Table  1 displays a details of demographic features and comparison of mean burnout scores between different demographic groups. The study found a significant difference between the burnout scores and different age groups and self-perceived levels of family support. Burnout scores increased with age, while those who reported higher levels of family support had lower burnout levels on average. Although other groups did not show any significant results, the study did find that males had higher burnout levels compared to females, and burnout scores increased with the level of education. The study did not find any significant association between study time and burnout levels.

Table  2 displays the response rate of the participants according to year of education. Female students had a higher response rate than their male counterparts with a total response rate of 97.4%. In Mohebbi et al’s study, factor analysis identified three questionnaire domains. (7) Overload (4 questions), Neglect (4 questions), and Development (4 questions). Table  3 displays p values for various demographic variables in each domain.

There were significant differences in the scores of different age groups across three domains. In the neglect domain, males scored considerably higher than females. Moreover, in the development domain, the participants who received scholarships scored lower than those who didn’t, indicating that the latter group was more satisfied with their growth and progress. Interestingly, students who reported the least amount of family support scored significantly lower in the overload domain of the survey, suggesting that they had a better balance between their academic pursuits and other aspects of their lives.

Among the three domains, ‘Overload’ had the highest reported mean of 17.79 (5.18) and a median of 17.00 (Table  4 ).

As presented in Table  5 , the majority (70.9%) agreed they invested an unhealthy amount of time in studies, while a smaller proportion agreed they compromised health (29.8%) or ignored personal needs over studies (28.2%). A minority (18.5%) disagreed with the statement “ I feel that my current studies are hampering the development of my abilities” and only 19.2% disagreed with “I give up when faced with any difficulty in my tasks as a student” implying a high burnout score.

The multiple regression model statistically significantly predicted Burnout, F(21, 129) = 2.190, p  = .004, adj. R 2  = 0.143. The model explained 26.3% of the variance in Burnout (R 2  = 0.263). Assumptions for the regression analysis were met, including linearity, independence of residuals (Durbin-Watson statistic = 2.081), absence of multicollinearity (all tolerance values > 0.1, lowest = 0.176), and normality of residuals. No influential outliers were detected based on studentized deleted residuals, leverage values, and Cook’s distance.

Five variables added statistically significantly to the prediction ( p  < .05). Gender was a significant predictor, with female students reporting lower burnout scores compared to males (B = -5.633, p  = .012). Age also played a role, as students under 19 (B = -26.909, p  = .004) and those aged 19–21 (B = -11.874, p  = .033) experienced significantly lower burnout compared to those aged 25–27. Place of living was another significant factor, with living in a dormitory associated with lower burnout compared to living in a private flat/apartment (B = -23.910, p  = .027). Family support also emerged as significant, with students reporting merely good family support showing higher burnout scores compared to those reporting very good family support (B = 6.147, p  = .009).

Other variables, including year of education, distance from home, studying hours, subjects failed, marital status, and scholarship status, did not contribute significantly to the model. Regression coefficients and standard errors can be found in Table  6 (below).

These results suggest that gender, age, living situation, and perceived family support play significant roles in student burnout levels. Female students, younger students, those living in dormitories, and those who reported very good family support tend to experience lower levels of burnout compared to males, older students, those living in private flats/apartments, and those who reported only good family support, respectively. It’s important to note that while these factors show significant associations with burnout, the model explains only 26.3% of the variance in burnout scores, indicating that other factors not included in this model may also play important roles in student burnout.

In light of its increasing significance, the assessment of occupational burnout has prompted the development of various scales in recent years. To examine the levels of burnout among dental students in Saudi Arabia, this study employed the Burnout Clinical Subtype Questionnaire short version (BCSQ 12 SS), which is specifically designed to measure burnout levels in students exposed to clinical settings. By utilizing the BCSQ-12 SS, this research was able to make comparisons with previous studies that also utilized the same assessment tool.

Our study revealed a mean score of 13.93, 17.60, and 17.79 in neglect, lack of development, and overload domains respectively. A study conducted in Iran revealed lower scores across all three domains of neglect, lack of development, and overload (11.8, 9.7, and 8.1) [ 6 ]. Similarly, a study conducted in India, evaluating burnout amongst medical practitioners reported a lower mean value for overload (15.89), lack of development (11.56), and neglect (10.28) [ 18 ].

In our study, overload emerged as the predominant aspect of burnout among dental students in Saudi Arabia, exhibiting the highest average score of 17.79. A separate study conducted in Saudi Arabia using the Dental Environment Survey (DES) reported that “Workload” was identified as the primary source of stress [ 19 ]. The domain of ‘Overload’ is strongly linked to exhaustion [ 6 ]. Further supporting our findings, a study in Pakistan found emotional exhaustion to be the most prevalent domain of burnout among dentists [ 20 ].

When the burnout levels were compared across the demographic groups it was observed that an increase in age was accompanied by an increase in burnout score. The difference in scores across age groups was significant across all three domains. A study conducted in Saudi Arabia also revealed a significant correlation between age and burnout when multiple logistic regression model was constructed [ 21 ]. A different study conducted in Saudi Arabia measuring burnout levels among medical students presented similar findings where age was significantly associated with high exhaustion levels [ 22 ]. Factors like prolonged academic demands, the pressure to excel in specialized fields, balancing work and academic commitments, or feeling the weight of future career expectations may be significantly higher in older students resulting in these findings. Another study revealed no significant association between age and academic burnout [ 23 ].

The presence of family support was also found to correlate with levels of burnout. Specifically, individuals who reported inadequate family support scored significantly lower on the burnout scale and also exhibited lower scores in the overload domain. This may be attributed to the emotional understanding, practical assistance, and encouragement provided within a supportive family environment. Furthermore, other studies investigating the connection between academic burnout and family support have similarly highlighted the pivotal role of family in alleviating burnout [ 24 , 25 ]. Our study did not show any significant difference between the burnout levels of students according to their living place whereas a different study showed a significant difference [ 23 ]. This could be due to a difference in sample size, with a minimal number of students not living with parents.

Upon comparing the differences across the three domains, it was discovered that males obtained significantly higher scores than females in the neglect domain. A study conducted in Brazil revealed males experience higher burnout levels as compared to females [ 26 ]. Whereas, a separate study conducted in UAE found no significant difference between males and females in this domain [ 27 ]. Social and cultural factors may have influenced the results in these studies, leading to varying outcomes. In certain cultures, males may feel a greater sense of responsibility to perform well, which could result in higher burnout scores.

The students not availing themselves of scholarships scored lower in the lack of development domain, as they felt they were not being challenged enough. In contrast, a study conducted in UAE revealed no significant difference in the lack of development domain, however, scholarship-availing students had significantly higher scores in the neglect domain [ 27 ].

In our research, a significant proportion of 70.9% of participants acknowledged dedicating an excessive amount of time to their studies. Conversely, a study carried out in the United Arab Emirates (UAE) demonstrated that 57.9% of respondents agreed with this assertion [ 27 ]. This disparity could potentially be attributed to variations in cultural perspectives towards academic pursuits, diverse educational systems, and societal expectations concerning academic achievement. Notably, the statement that received the least agreement in our study was ‘I ignore my own needs to satisfy the requirements of my studies’ with only 22.5% of participants concurring. In contrast, the study conducted in the UAE revealed that 52.8% of participants agreed with this statement [ 27 ].

To the best of our knowledge this research is the first to utilize the BCSQ 12 SS to assess burnout among dental students in Saudi Arabia. While previous studies have employed different instruments, our approach provides insights across the three dimensions outlined by the questionnaire: Overload, lack of development, and neglect. Another notable strength lies in the exploration of correlations between demographic variables and potential factors linked to academic burnout among dentistry students, including living arrangements, family support, and distance from home.

Reducing burnout among dental students warrants measures including but not limited to; addressing workload, improving support systems and advocating for mental health and gender differences. Addressing the workload would require improvement in time management to help students manage their responsibilities better and by offering flexible schedules. Support systems can be strengthened through family workshops, peer groups, and mentorship programs providing emotional and practical assistance. On campus counseling and wellness programs can help students cope with stress and help improve the mental health of students. Incorporating resilience training and career counseling into the curriculum can aid in the development of necessary skills.

However, certain limitations must be acknowledged. Firstly, the study’s sample was exclusively drawn from a single university. To provide a more comprehensive and representative understanding of burnout levels among dental students in the country it would be better to have participants from different schools. Another limitation is that the study does not investigate the correlation between the number of clinical work hours and burnout levels, which is a significant factor in understanding the dynamics of academic burnout in the context of dentistry. Moreover, the reliance on self-reported burnout is also a potential limitation as self-reporting might not offer the most accurate representation of burnout levels because individuals perceive their stressful experiences differently. Lastly, the study instrument was not validated for Saudi Population.

Further research is warranted to validate the BCSQ-12SS in Saudi population and assess the correlation between burnout and various curriculum patterns and educational systems worldwide to establish a connection between the education system and burnout among dental students. Furthermore, investigating the relationship between the number of clinical work hours and the number of patients seen daily could provide insights into the potential correlation between clinical working hours and burnout. Subsequent studies could also concentrate on longitudinal research to determine if burnout levels fluctuate among dental students as they transition from non-clinical years to clinical years. Additionally, exploring the impact of burnout on treatment outcomes, grade point average (GPA), and potential interventions to address burnout in dental students could be the focus of other investigations. Moreover, it would be beneficial to explore the link between internal factors such as personality types, organizational skills, confidence, self-esteem levels, and burnout.

In conclusion, the study highlighted the prevalence of overload as a predominant aspect of burnout among the three domains; overload, neglect, and lack of development. It also revealed burnout is influenced by factors such as age, family support, gender, and scholarship availability. Understanding these intricate associations is crucial in formulating tailored interventions to alleviate burnout, promoting a healthier academic environment that supports the well-being of students.

Data availability

Data will be made available by the corresponding author upon reasonable request.

Bassam S, Mohsen H, Barakat Z, Abou-Abbas L. Psychometric properties of the arabic version of the maslach burnout inventory-human services survey (MBI-HSS) among Lebanese dentists. BMC Oral Health. 2023;23(1):451. https://doi.org/10.1186/s12903-023-03169-7 .

Article   Google Scholar  

Maslach C, Jackson SE. The measurement of experienced burnout. J Organizational Behav. 1981;2(2):99–113.

Maslach C, Jackson SE. Burnout in organizational setting. Appl Social Psychol Annual. 1984;5:133–53.

Google Scholar  

Maslach C, Jackson SE, Leiter MP. Maslach burnout inventory. In: Evaluating Stress: A Book of Resources, 3rd Edition, Scarecrow Education. 1984; 191–218.

Montero-Marin J, Monticelli F, Casas M, Roman A, Tomas I, Gili M, Garcia-Campayo J. Burnout syndrome among dental students: a short version of the Burnout Clinical Subtype Questionnaire adapted for students (BCSQ-12-SS). BMC Med Educ. 2011;11:103. https://doi.org/10.1186/1472-6920-11-103 .

Mohebbi SZ, Yazdani R, Talebi M, Pakdaman A, Heft MW, Bahramian H. Burn out among Iranian dental students: psychometric properties of burnout clinical subtype questionnaire (BCSQ-12-SS) and its correlates. BMC Med Educ. 2019;19(1):388. https://doi.org/10.1186/s12909-019-1808-3 .

Mohebbi SZ, Gholami M, Chegini M, Ghoreyshi Y, Gorter RC, Bahramian H. Impact of career choice motivation on academic burnout in senior dental students: a cross-sectional study. BMC Med Educ. 2021;21(1):52. https://doi.org/10.1186/s12909-020-02475-w .

Hopcraft MS, Stormon N, McGrath R, Parker G. Factors associated with suicidal ideation and suicide attempts by Australian dental practitioners. Community Dent Oral Epidemiol. 2023;51(6):1159–68. https://doi.org/10.1111/cdoe.12849 .

Giri S, West CP, Shanafelt T, Satele D, Dyrbye LN. Distress and well-being in dentists: performance of a screening tool for assessment. BDJ Open. 2024;10(1):3. https://doi.org/10.1038/s41405-024-00185-9 .

Baldwin PJ, Dodd M, Rennie JS. Young dentists–work, wealth, health and happiness. Br Dent J. 1999;186(1):30–6. https://doi.org/10.1038/sj.bdj.4800010 .

Shirom A, Melamed S. Does burnout affect physical health? A review of the evidence. In: Antoniou ASG, Cooper CL,editors. Research companion to organizational health psychology. Edward Elgar Publishing. Cheltenham, UK;2005. p.599–622.

Gorter RC, Eijkman MA, Hoogstraten J. Burnout and health among Dutch dentists. Eur J Oral Sci. 2000;108(4):261–7. https://doi.org/10.1034/j.1600-0722.2000.108004261.x .

Garcia CL, Abreu LC, Ramos JLS, Castro CFD, Smiderle FRN, Santos JAD, Bezerra IMP. Influence of Burnout on Patient Safety: systematic review and Meta-analysis. Med (Kaunas). 2019;55(9):553. https://doi.org/10.3390/medicina55090553 .

Hopcraft MS, McGrath R, Stormon N, Tavella G, Parker G. Australian dental practitioners experience of burnout. J Public Health Dent. 2023;83(4):397–407. https://doi.org/10.1111/jphd.12594 .

von Elm E, Altman DG, Egger M, Pocock SJ, Gøtzsche PC, Vandenbroucke JP, STROBE Initiative. The strengthening the reporting of Observational studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. Bull World Health Organ. 2007;85(11):867–72. https://doi.org/10.2471/blt.07.045120 . PMID: 18038077; PMCID: PMC2636253.

Qualtrics Sample Size Calculator. https://www.qualtrics.com/blog/calculating-sample-size/

IBM Corp. Released 2019. IBM SPSS statistics for Windows, Version 26.0. Armonk, NY: IBM Corp.

Langade D, Modi PD, Sidhwa YF, Hishikar NA, Gharpure AS, Wankhade K, Langade J, Joshi K. Burnout syndrome among Medical practitioners across India: a questionnaire-based survey. Cureus. 2016;8(9):e771. https://doi.org/10.7759/cureus.771 .

Halboub E, Alhajj MN, AlKhairat AM, Sahaqi AM, Quadri MFA. Perceived stress among undergraduate Dental students in relation to gender, clinical training and academic performance. Acta Stomatol Croat. 2018;52(1):37–45. https://doi.org/10.15644/asc52/1/6 . PMID: 30034003; PMCID: PMC6050751.

Ahmad Z, Zaidi RZ, Fatima Z, Muhammad M, Zafar MS, Kolarkodi SH, Javed MQ. Burnout level in Pakistani dentists during COVID-19 pandemic: cross-sectional national study. Heliyon. 2023;9(12):e23061. https://doi.org/10.1016/j.heliyon.2023.e23061 .

Bahlaq MA, Ramadan IK, Abalkhail B, Mirza AA, Ahmed MK, Alraddadi KS, Kadi M. Burnout, stress, and stimulant abuse among Medical and Dental students in the Western Region of Saudi Arabia: an Analytical Study. Saudi J Med Med Sci. 2023 Jan-Mar;11(1):44–53. https://doi.org/10.4103/sjmms.sjmms_98_22 . Epub 2023 Jan 14. PMID: 36909001; PMCID: PMC9997854.

Bahathig AM, Aldhowaihy FA, Alsmari KA, Alluhidan TA, Alturki AA, Alghuraybi AS, Alotaibi AT. Prevalence of burnout among KSU medical students. Majmaah J Health Sci. 2022;10(3):101–11.

Shokrpour N, Bazrafcan L, Ardani AR, Nasiraei S. The factors affecting academic burnout in medical students of Mashahd University of Medical Sciences in 2013–2015. J Educ Health Promot. 2020;9:232. https://doi.org/10.4103/jehp.jehp_83_20 .

Yu J, Wang Y, Tang X, Wu Y, Tang X, Huang J. Impact of Family Cohesion and adaptability on Academic Burnout of Chinese College students: serial mediation of peer support and positive psychological capital. Front Psychol. 2021;12:767616. https://doi.org/10.3389/fpsyg.2021.767616 .

Schrodt P, Ledbetter AM, Ohrt JK. Parental confirmation and affection as mediators of family communication patterns and children’s mental well-being. J Family Communication. 2007;7(1):23–46.

Pacheco JP, Giacomin HT, Tam WW, Ribeiro TB, Arab C, Bezerra IM, Pinasco GC. Mental health problems among medical students in Brazil: a systematic review and meta-analysis. Braz J Psychiatry. 2017;39(4):369–78. https://doi.org/10.1590/1516-4446-2017-2223 .

Al-Rawi NH, Yacoub A, Zaouali A, Salloum L, Afash N, Shazli OA, Elyan Z. Prevalence of Burnout among Dental students during COVID-19 lockdown in UAE. J Contemp Dent Pract. 2021;22(5):538–44.

Download references

Acknowledgements

The authors appreciate the support from Researchers Supporting Project number (RSPD2024R950), King Saud University, Riyadh, Saudi Arabia.

The research was funded by Researchers Supporting Project number (RSPD2024R950), King Saud University, Riyadh, Saudi Arabia.

Author information

Authors and affiliations.

Department of Conservative Dental Sciences, College of Dentistry, Qassim University, Buraidah, Qassim, 52571, Saudi Arabia

Muhammad Qasim Javed

Department of Operative Dentistry, Islamic International Dental College and Hospital, Riphah International University, Islamabad, Pakistan

Zaina Ahmad & Muhammad Muhammad

Department of Prosthetic Dental Sciences, College of Dentistry, King Saud University, Riyadh, 11545, Saudi Arabia

AbdulAziz Binrayes & Syed Rashid Habib

Department of Operative Dentistry, Frontier Medical and Dental College, Abbottabad, Pakistan

Iffat Niazi

Consultant Prosthodontist, P. O. Box. 45730, Regenix, Islamabad, Pakistan

Shazia Nawabi

Department of Endodontics, Faculty of Dentistry, King Abdulaziz University, P.O. Box 80209, Jeddah, 21589, Saudi Arabia

Ayman M. Abulhamael

Department of Prosthodontics, HBS Medical and Dental College, Islamabad, Pakistan

You can also search for this author in PubMed   Google Scholar

Contributions

Muhammad Qasim Javed: Conceived and designed the experiments; Performed the experiments; analyzed and interpreted the data; and contributed to drafting the manuscript. Zaina Ahmed: Muhammad Muhammad: Analyzed and interpreted the data; and contributed to drafting the manuscript. Ayman Moaz Abulhamael: AbdulAziz Binrayes: Syed Rashid Habib: Contributed reagents, materials, analysis tools or data; and contributed to drafting the manuscript. Shazia Nawabi: Iffat Niazi: Conceived and designed the study; contributed to drafting and reviewing the manuscript.

Corresponding author

Correspondence to Muhammad Qasim Javed .

Ethics declarations

Ethics approval and consent to participate.

The research protocol was approved by the ethical committee at Dental College, Qassim University (Approval no: 24-8-03). Informed consent was obtained from each participant through written consent forms, which clearly outlined the purpose of the study and the rights of the participants.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.

Additional information

Publisher’s note.

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/ .

Reprints and permissions

About this article

Cite this article.

Javed, M.Q., Ahmad, Z., Muhammad, M. et al. Burnout level evaluation of undergraduate dental college students at middle eastern university. BMC Med Educ 24 , 1155 (2024). https://doi.org/10.1186/s12909-024-06149-9

Download citation

Received : 26 March 2024

Accepted : 07 October 2024

Published : 16 October 2024

DOI : https://doi.org/10.1186/s12909-024-06149-9

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

• Academic burnout
• Dental education
• Dental students
• Work overload

BMC Medical Education

ISSN: 1472-6920

• General enquiries: [email protected]

• Accreditation
• Value of Accreditation
• Standards and Process
• Search Accredited Schools

• Educational Membership
• Business Membership
• Find a Member

• Learning and Events
• Conferences
• Webinars and Online Courses

• All Insights
• Quality Standards
• Business Education
• World of Work
• Societal Impact

• Leadership and Governance
• Media Center

• Accredited School Search
• Advertise, Sponsor, Exhibit
• Tips and Advice
• Is Business School Right for Me?

Enhancing Student Engagement With Custom GPTs

• GPTs tailored to specific applications can provide students with personalized, instant feedback that works to enhance their critical thinking and communication skills.
• At the Indian School of Business, custom GPTs encourage students to write more thoughtful responses to assignments, while reducing opportunities for academic dishonesty.
• AI offers educators a dynamic solution that boosts student engagement and introspection, making learning more interactive and effective.

  It has been nearly 15 years since Srikant Datar, David A. Garvin, and Patrick Cullen released their book Rethinking the MBA , in which they call for business schools to place greater emphasis on complex skills such as critical thinking and communication. But what are the best ways to cultivate those skills in future leaders?

Researchers such as Daniel Kahneman and Gary Klein in their 2009 paper have pointed to the need for timely and unambiguous feedback to master such complex skills. Such feedback is best delivered via personal tutoring. In his 1984 paper , Benjamin S. Bloom shows that personal tutoring can result in as much as a two-sigma improvement in student performance. Bloom noted that, when tutored, an average student performs at the 96th percentile of those taught by traditional methods.

But as effective as we know personal tutoring can be, it can be challenging for higher education institutions to deliver. The main issue is the lack of scalability due to its cost and a shortage of qualified faculty.

As early as almost 100 years ago, cognitive psychologist Sidney L. Pressey of Ohio State University proposed technology to solve this problem. Pressey developed a “teaching machine” that gave feedback on student responses to multiple-choice questions and even delivered a candy after a predetermined number of correct responses!

Yet, in all this time, the promise of that technology has not been realized—until now.

Designing an AI Intervention

My co-instructor, Harish Raichandani, and I teach Corporate Governance and Business Ethics, a course in an MBA program for working executives at the Indian School of Business (ISB) in Hyderabad. When we started teaching this course in August 2022, GPT technology was unknown. We gave traditional assignments that required students to submit write-ups. These took us more than month to grade, even with the help of teaching assistants.

In 2023, with the rise of ChatGPT , we soon realized that students could cut and paste responses that they generated with the AI platform, and Turnitin would fail to detect plagiarism. So, we switched to asking groups of students to record oral responses to questions given on the spot (noting that their grades would be affected if only a few students did most of the talking). But while effective, this method resulted in an even longer grading process.

It was in 2024 that I read about custom GPTs , a feature that allows users to tailor versions of ChatGPT to suit their specific purposes. I visualized the possibility of creating a tool that could provide students with personalized guidance and instant feedback on critical writing assignments.

Because ChatGPT does not record student responses, I wrote a web application for the purpose. The application communicated with the ChatGPT platform on the back end using the OpenAI application programming interface . Because I was unfamiliar with today’s tools, such as the Flask package in Python, I used large language models such as Anthropic’s Claude 2 for help creating the app. Developing, testing, and debugging the GPTs and the app took me about a week.

As we experimented with custom GPTs, Harish and I saw that we could achieve what we wanted through simple prompts. After some further tweaking and testing, we recently introduced our custom intervention to the 73 students enrolled in our course.

Keeping Our Prompts Simple

We created a custom GPT for each assignment, giving it the following instruction: “You are an expert viva examiner examining executive MBA students in a course on business ethics and corporate governance.” We then asked it to generate questions for each of our assessments.

Students accessed each assignment through the web app. After authenticating the student, the app presented the student with questions generated on the spot by the custom GPT and waited for the student’s response. Once the student answered, the app sent the response to the GPT and displayed the output.

Although we did not mention climate change or stakeholder capitalism in our instructions, the GPT came up with these topics for the questions independently.

When asking the GPT to generate the questions, we kept our instructions simple. For example, for Part 1 of an assignment on Milton Friedman’s 1970 article about the social responsibility of business, we gave the AI platform this brief instruction: “Give three multiple-choice questions that test whether the student has read the article.”

For Part 2 of the assignment, we told the custom GPT to “ask an open-ended question demonstrating a basic understanding of a core message from Friedman’s essay.” Based on this instruction, it produced questions such as “Explain the core message of Friedman ’ s argument regarding the social responsibility of businesses” and “Discuss what Friedman means by the ‘rules of the game’ in the context of corporate social responsibility.”

Finally, in Part 3, we instructed the custom GPT to ask a question that “tests students on whether they can apply the arguments to issues arising in contemporary times and carefully analyze Friedman ’ s opinion in that context.” In our prompt, we made it clear that “students should be able to articulate their opinion on the validity of Friedman’s arguments, consider counterarguments to their stated position, and draw a conclusion.”

In response, the GPT asked questions such as, “In the context of contemporary socioeconomic issues, such as climate change, how do you argue [whether] Friedman’s viewpoint holds or does not hold?” and “Given the rise of stakeholder capitalism, where businesses are increasingly expected to account for the interests of all stakeholders, not just shareholders, evaluate Friedman ’ s argument.”

Keep in mind, we did not mention climate change or stakeholder capitalism in our instructions. The GPT came up with these topics for the questions independently.

Using AI as Personal Tutor

Next, we instructed the AI to act as a tutor to help students refine their answers in parts 2 and 3 of the assignment. For this purpose, we entered the following prompt: “Once students have submitted their initial responses, you will give them an opportunity to add further to their response, hinting at specific areas they could address.”

At this stage, the AI let students know what they did well and what they could improve upon. This gave them opportunities to introspect, which is critical for learning complex skills. By inviting students to add to their responses in real time and offering suggestions for improvement, the AI made the assignment less stressful, more educational, and more engaging—indeed, a couple of students even completed the assignment twice.

We also asked the GPT to grade these assignments. To train the AI in the grading process, we wrote the following simple prompt, assigning point values for each of the three questions: “You will evaluate and award a maximum of 12 points at the end of a student's submission (Q1–3 points; Q2–4 points; Q3–5 points).” Even though we did not provide the GPT with an elaborate rubric, we found that its grading ability was good. The class average was about 10.5 points out of 12.

Students were free to approach us if they felt the grading was incorrect. A few students came to us, but it was mostly for grading errors in Part 1, where the GPT made obvious mistakes, such as saying, “The correct answer is ‘B’ and your response was ‘B.’”  However, we did not receive a single request to regrade the open-ended questions in parts 2 and 3.

Protecting Academic Integrity

The suggested time to complete the assignment was 30 minutes, with a hard stop at the 60-minute mark (most students took about 40 minutes). Since the questions were generated on the spot by the GPT, students had little opportunity to look up answers on the web or in another AI tool.

The real-time tutoring that AI platforms such as ChatGPT can provide is a game changer in higher education.

Moreover, the application did not permit students to paste any text—students had to type their responses. We could have reduced the possibility for cheating even further by requiring students to agree to be recorded through their webcams. But we did not feel it necessary to use such an intrusive method in an executive MBA program.

As we deployed this assignment, we did keep issues of privacy in mind. Even though OpenAI promises that the data sent to its platform will not be used to train its models, we advised students not to reveal any identifying information in their interactions with the GPT.

Future Directions and Final Say

We plan to continue fine-tuning the prompts to the custom GPTs. Our goal is to ensure that, when students provide shallow responses to questions, the technology challenges them to go into more depth. Additionally, we are working on a more detailed grading rubric for the next version.

Now that we can turn to AI for help, we also intend to return to oral group assignments. From our experience, we have found that students come well-prepared to these recordings, because they do not want to let other group members down. Students also told us that they found group work useful in fostering positive classroom dynamics. For these reasons, we believe that combining oral recordings with the GPT’s tutoring will be a significant improvement over the first iteration of this pedagogical approach.

In Rethinking the MBA , Datar, Garvin, and Cullen point out that MBA education is at a crossroads. They find that MBA students have been engaging less and less with course material, and the time they spend preparing has decreased substantially over time. For instance, the authors offered a stark comparison: In 1975, one institution found that its students spent 45 to 50 hours per week attending and preparing for classes; by 2003, this number had decreased to 30.

Digital distractions have grown substantially since 2010, when their book was published. Moreover, in large classes, professors do not have the time to provide detailed feedback to every student on every assignment. Given these factors, it’s not surprising that students treat assignments as necessary evils rather than learning opportunities, and take shortcuts where they can.

But we now have a viable solution to these challenges. Because students’ every keystroke can be recorded and time-stamped, the technology reduces both motive and opportunity for cheating, two of the three sides of the fraud triangle . In addition, AI can quickly generate and grade assignments, saving instructors valuable time.

And because AI can offer immediate, tailored feedback, it offers true learning opportunities for students. The real-time tutoring that AI platforms such as ChatGPT can provide is a game changer in higher education. Pressey’s “teaching machine” is finally here.

I thank Seema Chowdhry, director of ISB’s Centre for Learning and Teaching Excellence, for her help in creating the GPT tool for our course and in preparing this article.

• Artificial Intelligence
• Learning Delivery
• Student Engagement