High School Graduation Rates in the United States
REVISED APRIL 2002
Jay P. Greene, Ph. D.
Senior Fellow, The Manhattan Institute for Policy Research
with a foreword by Kaleem Caire
President and CEO
Black Alliance for Educational Options
Prepared for the Black Alliance for Educational Options
INTRODUCTION TO THE REVISED REPORT
This study revises slightly the findings of my November 2001 report, High School Graduation Rates in the United States. In that study, I used an easily replicable method to estimate the percentage of public high school students receiving a high school diploma in the nation, each state and many of the nation’s largest public school districts. The same method was also used to estimate these rates for major racial and ethnic groups in each state and each of the districts examined.
I recently discovered an error in the calculations that were used to estimate the overall national and state rates. In my methodology, I estimate the graduation rate by dividing the number of public high school diplomas awarded in 1998, which is available from the National Center for Education Statistics, by an estimate of the number of students who would have received diplomas that year if graduation rates were 100 percent. I arrive at this latter number by taking the number of students enrolled in public schools in 8th grade in 1993 (also available from the NCES) and adjusting it for the percentage change in the overall student population between 1993 and 1998. The error stemmed from the inadvertent use of the percentage change in the overall population rather than overall student population between those years.
Recalculating the national rate to correct for this error, I now find that estimated national public school graduation rate in 1998 was 71 percent, slightly lower than the 74 percent originally reported. Since the overall thrust of my report was that public schools graduation rates are much lower than is commonly reported, this recalculation does not change the original report’s conclusion.
Estimated graduation rates for each state were also recalculated. These changes may be found in Table 1, at the back of this report.
The mistaken calculation occurred only for the overall state, and hence the overall national, graduation rates. The calculation was done correctly for each of the local school districts, the state-level racial and ethnic results, and the district-level racial and ethnic results.
In general, the differences between the new and previously reported numbers are modest. In fact, the two sets of numbers are correlated at .94. If the two sets were identical the correlation would be 1.0. The changes tend to be small because in most states the total population and the total student population grew at similar rates. In those states where the student population grew at a rate very different from the total state population, however, the changes could be larger.
As long the report was being revised I took the opportunity to correct a previously reported data entry error for Jefferson County, Kentucky. I also re-examined the entire data set for any other data entry errors and added information from Arizona that arrived too late to be included in the original report. No data entry errors were found in the state results but a few errors were found in the district numbers. None of the corrections change reported graduation rates by more than one or two percentage points except for a larger error for Virginia Beach, Virginia, where the overall graduation rate was lowered by 11% because of a data entry error, and Saint Paul, where graduation rates were previously understated. All district-level results stated in this report reflect these corrections. All tables in the current appendix refelct these changes.
Of the hundreds of numbers entered and the scores of calculations made I am pleased to have found relatively modest errors, but am chagrined to have found any errors at all. I, the Manhattan Institute, and the Black Alliance for Educational Options will continue to strive to provide the highest quality research.
Jay P. Greene
Senior Fellow, Manhattan Institute
The report's main findings are the following:
- The national graduation rate for the class of 1998 was 71%. For white students the rate was 78%, while it was 56% for African-American students and 54% for Latino students.
- Georgia had the lowest overall graduation rate in the nation with 54% of students graduating, followed by Nevada, Florida, and Washington, D.C.
- Iowa had the highest overall graduation rate with 93%, followed by North Dakota, Wisconsin, and Nebraska.
- Wisconsin had the lowest graduation rate among African-American students with 40%, followed by Minnesota, Georgia, and Tennessee. Georgia had the lowest graduation rate among Latino students with 32%, followed by Alabama, Tennessee, and North Carolina. Less than 50% of African-American students graduated in seven states and less than 50% of Latino students graduated in eight states for which data were available.
- The highest rate of graduation among African-American students was 71% in West Virginia, followed by Massachusetts, Arkansas, and New Jersey. The highest rate of graduation among Latino students was 82% in Montana, followed by Louisiana, Maryland, and Hawaii.
- Among the fifty largest school districts in the country, Cleveland City had the lowest overall graduation rate with 28%, followed by Memphis, Milwaukee, and Columbus.
- Fairfax County, VA had the highest overall graduation rate among the districts with 87%, followed by Montgomery County, MD, Albuquerque and Boston.
- Cleveland City had the lowest graduation rate among African-American students with 29%, followed by Milwaukee, Memphis, and Gwinett County, Georgia. Cleveland City also had the lowest graduation rate among Latino students, followed by Georgia’s Dekalb, Gwinnett, and Cobb counties. Less than 50% of African-American students graduated in fifteen of forty-five districts for which there was sufficient data, and less than 50% of Latino students graduated in twenty-one of thirty-six districts for which there was sufficient data.
- The National Center for Education Statistics (NCES) finds a national high school completion rate of 86% for the class of 1998. The discrepancy between the NCES’ finding and this report’s finding of a 71% rate is largely caused by NCES’ counting of General Educational Development (GED) graduates and others with alternative credentials as high school graduates, and by its reliance on a methodology that is likely to undercount dropouts.
ABOUT THE AUTHOR
Jay P. Greene is a Senior Fellow at the Manhattan Institute for Policy Research where he conducts research and writes about education policy. He has conducted evaluations of school choice and accountability programs in Florida, Charlotte, Milwaukee, Cleveland, and San Antonio. He has also investigated the effects of school choice on civic values and integration. His publications include “An Evaluation of the Florida A-Plus Choice and Accountability Program” forthcoming in Education Next; “The Surprising Consensus on School Choice” in the Summer 2001 issue of The Public Interest; “Vouchers in Charlotte” in the Summer 2001 issue of Education Matters; the chapters, “Civic Values in Public and Private Schools” and “School Choice in Milwaukee: A Randomized Experiment” in the book, Learning from School Choice, published by the Brookings Institution in 1998; and “The Effect of Private Education on Political Participation, Social Capital, and Tolerance” in the Fall 1999 issue of The Georgetown Public Policy Review. Dr. Greene has been a professor of government at the University of Texas at Austin and the University of Houston. He received his Ph.D. from the Government Department at Harvard University in 1995. He lives with his wife and three children in Weston, Florida.
I would like to thank Rob Fusco for his tireless research assistance on this project. I would also like to thank the reviewers of this report, Duncan Chaplin of The Urban Institute, Adam Gamoran of the University of Wisconsin at Madison, and Diane Ravitch of New York University, for their valuable suggestions and criticisms. Kaleem Caire, Chester E. Finn, Jr., Howard Fuller, George Mitchell, Henry Olsen, and Paul Peterson kindly took time to read drafts and make constructive suggestions. Finally, I would like to thank the staff of the National Center for Education Statistics and education officials in states and school districts around the country who made the effort to provide the data used to calculate graduation rates in this report.
“Until many more…minority students…are very successful educationally, it will be virtually impossible to integrate our society’s institutions completely, especially at leadership levels. Without such progress, the United States also will continue to be unable to draw on the full range of talents in our population during an era when the value of an educated citizenry has never been greater.”
—“Reaching the Top,” The College Board (1999)
At a March 2001 education conference in Washington D.C., an audience member posed two questions to a representative of President George Bush:
- Why is so little attention paid to the high dropout rate among the nation’s African-American children?
- Why does the U.S. Department of Education (DOE) annually report incomplete and sometimes inaccurate dropout statistics to the general public?
The President’s aide responded: “The truth hurts, and few people want to share the truth about under-performing students these days.”
Six months earlier, I had asked authors of a DOE dropout study issued during the Clinton Administration why it overstated the number of African-American children receiving high school diplomas. They explained that, in addition to students who actually graduated from high school, their data included recipients of so-called high school “equivalency” diplomas. Then, referring to the controversial “wall chart” once displayed at DOE, they said the federal government stopped reporting on the number of ninth graders that completed high school in four years because it painted “too bad a picture of productivity of the nation’s public schools.”
Such anecdotes explain why the Black Alliance for Educational Options (BAEO) commissioned High School Graduation Rates in the United States.
Parents and other taxpayers must have accurate information about the educational status of our nation’s children. As the only national African-American organization trying to expand educational options for America’s children, BAEO is determined to examine honestly the effectiveness of our nation’s schools and the educational achievement of our children. BAEO knows that a high quality education is our children’s primary passport to achieving their life’s goals as adults.
This pioneering study by Jay P. Greene, Ph.D., sheds new light on an issue that adversely affects far too many American children. In particular, low graduation rates among students of color have devastating effects on their communities and thus on the nation as a whole. Children who do not graduate with a high school diploma stand little chance of sustaining themselves or a family in today’s economy.
BAEO wants all American children to complete K–12 education successfully. They will then be prepared for higher education and they will have the skills necessary to function effectively in today’s labor market.
Moreover, it is unacceptable to BAEO that Black America’s long-held goal of racial and ethnic diversity among our nation’s economic and political leadership is undermined by the massive failure of our young people to graduate from high school.
Reviewing the findings of this report—including the horrific graduation rates in such cities as Cleveland and Milwaukee—it is no wonder why parents there have led the fight for education vouchers and other new educational options for their children.
BAEO is determined that Dr. Greene’s previously unreported data will receive widespread attention. We hope that those who read this report will re-commit themselves to meeting the challenge of ensuring that all of our children truly receive a high quality education.
President and CEO, Black Alliance for Educational Options
HIGH SCHOOL GRADUATION RATES IN THE UNITED STATES
Students who fail to graduate from high school face a very bleak future. Because the basic skills conveyed in high school and higher education are essential for success in today’s economy, students who do not receive these skills are likely to suffer with significantly reduced earnings and employment prospects. Among those over 25 years old who failed to complete high school or receive a GED, 55% report no earnings in the 1999 Current Population Survey of the U.S. Census compared to 25% of those with at least a high school degree or GED. For people reporting any earnings the median income for those who left school without a high school diploma or GED is $15,334 compared to $29,294 for people with at least a high school degree or GED (see Figure 1).1 Students who fail to graduate high school are also significantly more likely to become single parents and have children at young ages. And students who do not graduate high school are significantly more likely to rely upon public assistance or be in prison.2 In short, high school graduation is a very important predictor of young people’s life prospects.
High school graduation rates are therefore also an important measure of the performance of our public school system. The better able schools are to provide students with the skills necessary to complete high school, the more successful the school system is.3 Given the strength of the relationship between high school graduation and students’ life prospects, graduation rates are at least as important as test scores in assessing the performance of our school system. Yet graduation rates have not received nearly as much attention as national test scores.
The relative inattention devoted to graduation rates is at least partly explained by the confusing, inconsistent, and sometimes misleading way in which the rate of high school completion is measured. Local and state public school officials report dropout and completion statistics that are difficult to grasp and often implausibly positive. The way in which those statistics are calculated and how they should be interpreted is often opaque to the trained researcher, let alone the general public. Even the normally very helpful National Center for Education Statistics (NCES) of the U.S. Department of Education has done little to improve the quality of statistics on high school completion. While the national government spends over $40 million for the National Assessment of Educational Progress, which the NCES uses to track performance on achievement tests, less than $1 million is spent by the national government on dropout/high school completion statistics.4
The purpose of this report is to calculate and report reliable and straightforward public high school graduation rates. Rates are reported for all students as well as broken out for African-American, Latino, and white sub-groups.5 Rates are also reported for each state, for each of the 50 largest school districts, and for a few other districts of interest. The state and district numbers are also reported separately for African-American, Latino, and white sub-groups. By reporting reliable and straightforward graduation rates we will have better information about how well school systems are performing overall as well as for each ethnic/racial group.
Calculating Graduation Rates
The method used here to calculate graduation rates is remarkably simple but also likely to be quite accurate.6 I identified the 8th grade public school enrollment for each jurisdiction and for each sub-group from the fall of 1993.7 I then collected information on the number of regular high school diplomas awarded in the spring of 1998 when those 8th graders should have been graduating.8 To adjust for the possibility that students moving into or out of an area would distort the graduation rate, I adjusted the 1993 8th grader counts for the student population change in that jurisdiction and for each ethnic/racial sub-group between the 1993–94 and 1997–98 school years.
The formula used to calculate the graduation rate was:
graduation rate = regular diplomas from 1998 / adjusted 8th grade enrollment from 1993
The formula used to adjust the 8th grade was:
Adjusted 8th grade enrollment = actual 8th grade enrollment + (actual 8th grade enrollment x percentage change in total or ethnic sub-group enrollment in the jurisdiction between 1993–4 and 1997–8)
The calculations can be illustrated by showing how the national graduation rate was computed. In the fall of 1993 there were 3,249,266 students enrolled in 8th grade. In 1998, when we would expect those students to be graduating, there were 2,440,081 regular diplomas awarded. Yet during these years the total student population in the United States increased by 6.1 %, so we adjust the 8th grade population upward by 6.1 % to 3,446,552 on the assumption that the 8th grade cohort received 6.1 % additional students from immigration or from the private sector.9 Of the 3,446,552 students we would expect to graduate in 1998, only 2,440,081 students actually received diplomas, producing a graduation rate of 71% (see Figure 2).
Similar calculations were made for each state and for each ethnic sub-group. For example, to calculate the graduation rate for African-American students in the state of Wisconsin I began by identifying that there were 5,604 African-American students in 8th grade in the fall of 1993. Between the 1993–4 and 1997–8 school years, however, the total African-American school population in the state increased from 76,446 to 85,977 students, an increase of 12.5%. To reflect this total African-American student population increase, the 8th grade African-American enrollment was adjusted up by 12.5% to 6,303 students. In 1998, when we would be expecting approximately 6,303 African-American students to be graduating, only 2,531 diplomas were awarded to African-American students in Wisconsin, yielding a graduation rate of 40%.
Even if we made no adjustment for the increasing African-American student population in Wisconsin, fewer than half of the African-American students enrolled in 8th grade in 1993 graduated from high school in 1998. If this is not a reasonable calculation of the graduation rate for African-American students in Wisconsin one has to be able to explain what happened to the over 3,000 African-American students who we expected to graduate but did not finish high school. One possible explanation is that students may take more than five years to go from 8th grade to graduation.10 This is true, but it must be remembered that the same must also be true for the cohort that was in the 8th grade in 1992 and so on, some of whom may be included in the 1998 graduate count. Students taking longer than normal to finish high school would only seriously distort the graduation rate if there were a large number of such students and if there were a dramatic increase or decrease in the proportion that took more time to graduate from one year to the next. Neither seems very likely, meaning that students taking more time to finish high school should not significantly distort the graduation rates calculated by the method employed here.
The Results: Ranking the States
As we have already seen, the national graduation rate for the class of 1998 was 71% (see Figure 3). For white students the graduation rate was 78%. For African-American students nationwide the graduation rate for the class of 1998 was 56%. For Latino students nationwide the graduation rate was 54%.11
At the state level there was considerable variation both in the overall graduation rate and in the rate for each sub-group. Table 1 presents the results for the states in alphabetical order. Table 2 presents the results for states ranked from the lowest overall graduation rate to the highest. Georgia has the lowest graduation rate of all of the states, with 54% of the class of 1998 graduating. Nevada has the next lowest graduation rate with 58% of its students completing high school, followed by Florida and Washington D.C. each with 59% graduation rates. Iowa has the highest overall graduation rate with 93% of its students graduating. With an 88% graduation rate North Dakota has the second best overall rate, followed by Wisconsin and Nebraska each with 85% overall graduation rates.
Some of the states with the best overall graduation rates, however, have some of the worst graduation rates for African-American students. As can be seen in Table 3, Wisconsin has the worst graduation rate for African-American students at 40% even though it had the third best overall graduation rate. Similarly, Minnesota, which has the second worst African-American graduation rate at 43%, has one of the highest overall graduation rates. These states have an enormous disparity between the graduation rates for whites and for African-Americans, with white students more than twice as likely to graduate. Some states, however, have low overall graduation rates and low African-American graduation rates. Georgia for example, has the third worst African-American graduation rate (44%) and the worst overall graduation rate. Tennessee has the fourth worst African-American graduation rate (44%) and sixth worst overall graduation rate. Three other states, for a total of seven states, have fewer than half of their African-American students in the class of 1998 graduating: Nevada (49%), Ohio (49%), and Oregon (49%).
Other states do relatively better with African-American graduation rates. West Virginia has the highest graduation rate for African-American students (71%) followed by Massachusetts with 70%. Arkansas has the third highest graduation rate for African-American students (67%) and New Jersey has the fourth highest African-American graduation rate with 66%.
The lowest state Latino graduation rates are even lower than those for African-American students (see Table 4). In Georgia, which has made repeated appearances among the list of worst states, only 32% of Latinos in the class of 1998 graduated. Alabama had the second lowest graduation rate for Latino students (33%), followed by Tennessee (38%). Five additional states, making for a total of eight states, had less than half of their Latino students graduating in the class of 1998: North Carolina (38%), Nevada (40%), Oregon (43%), Colorado (47%), and Arkansas (48%).
Some states, however, had relatively high Latino graduation rates. For example, Montana has the highest Latino graduation rate, with 82% of Latino students completing high school. Maryland and Louisiana have the second and third best Latino graduation rates, each with 70% of Latino students graduating. Hawaii has the fourth highest Latino graduation rate (66%). The Latino graduation rates in Montana and Hawaii have to be taken with a grain of salt, however, because there are relatively few Latinos in those states.
The graduation rates for whites follow fairly closely the graduation rates for all students because whites are the large majority in most states (see Table 5). Georgia has a graduation rate of 61% for its white students, followed by Florida with 63%. Tennessee has the third worst white graduation rate at 64% and Nevada has the fourth worst rate for white students at 65%. All four of these states are among those with the lowest overall graduation rates. On the positive end of the scale, Iowa (95%), Wisconsin (92%), and Nebraska (90%) have the highest graduation rates for white students and are also among the highest ranking states for the overall graduation rates.
The gap between white and minority graduation rates is alarmingly large. Indeed, the lowest state graduation rates for white students are close to the highest rates for African-American and Latino students. In some of the states the disparity between white and minority graduation rates is exceptionally high. For example, Wisconsin has the largest difference between its graduation rates for white and African-American students, with 92% of whites graduating compared to 40% of African-Americans. The gap between white and Latino graduation rates in Wisconsin is also among the largest differences in rates (92% vs. 56%). Minnesota also has a very large disparity between its white and minority graduation rates, with 87% of white students graduating compared to 43% of African-American and 53% of Latino students. Nebraska and Iowa also have some of the greatest disparities between white and minority graduation rates. Interestingly, all four of these states are predominantly rural, white states with concentrated, smaller minority and urban populations. This may reveal that the problem of low graduation rates is really an urban problem. An examination of graduation rates in large, mostly urban school districts is in the following section.
The Results: Ranking the Districts
The 50 largest districts in 1993 vary widely in their graduation rates (see Table 6). The district with the lowest graduation rate is Cleveland City, where only 28% of students complete high school. The district with the highest graduation rate is Fairfax County, Virginia, with 87% of students graduating (see Table 7). Altogether, five districts among the 50 largest districts in the U.S. have overall graduation rates below 50%: Cleveland (28%), Memphis (42%), Milwaukee (43%), Columbus (45%), and Chicago (47%). On the other hand, five districts of the 50 largest districts have overall graduation rates at or above 80%: Fairfax County, Virginia (87%), Montgomery County, Maryland (85%), Albuquerque, New Mexico (83%), Boston, Massachusetts (82%), and Jordan, Utah (80%).
Focusing upon the district results for African-American students reveals a more strongly negative picture (see Table 8). Sixteen of the 50 largest school districts failed to graduate more than half of their African-American students. Cleveland has the lowest graduation rate for African-American students (29%), followed by Milwaukee (34%), Memphis (39%), Gwinnett County, Georgia (40%), Pinellas County, Florida (41%), New York City (42%), Hillsborough County, Florida (42%), Columbus, Ohio (45%), Chicago (45%), Duval County, Florida (45%), Orange County, Florida (45%), Dekalb County, Georgia (46%), Cobb County, Georgia (47%), Clark County, Nevada (49%), Jefferson County, Kentucky (49%), and Mobile, Alabama (50%). Only four districts are able to graduate 75% or more of their African-American students: Boston (85%), Fairfax County, Virginia (77%), Prince Georges County, Maryland (76%), and Montgomery County, Maryland (75%).
The picture is even bleaker for Latino graduation rates in the 50 largest school districts (see Table 9). All but 15 of the districts for which rates can be computed have Latino graduation rates below 50%. Six districts have Latino graduation rates below 40%: Cleveland (26%), Dekalb County, Georgia (29%), Gwinnett County, Georgia (33%), Cobb County, Georgia (34%), Clark County, Nevada, and Dallas, Texas (39%). Only five districts have more than two-thirds of their Latino students completing high school: Montgomery County, Maryland (73%), Albuquerque, New Mexico (70%), Prince Georges County, Maryland (70%), Boston (68%), and El Paso, Texas (67%).
Only five districts are unable to graduate more than half of their white students: Cleveland (23%), Detroit (43%), Columbus, Ohio (46%), Baltimore City, Maryland (48%), and Memphis, Tennessee (50%) (see Table 10). Yet eleven districts have white graduation rates of at least 80%: Albuquerque (99%), Fairfax (92%), Philadelphia (91%), Prince Georges County, Maryland (90%), Montgomery County, Maryland (88%), Boston (87%), El Paso, Texas (86%), Houston, Texas (84%), Baltimore County, Maryland (84%), Los Angeles (81%), and New York City (80%).
For the most part, districts with low African-American and Latino graduation rates also had relatively low white graduation rates. A few districts, however, have large disparities between their white and minority graduation rates. For example, New York City graduates 80% of its white students but only 42% of its African-American and 45% of its Latino students. Dekalb County, Georgia has a 77% graduation rate for white students but only 46% of African-Americans and 29% of Latinos complete high school. Gwinnett and Cobb Counties in Georgia have similar large disparities between white and minority graduation rates. Milwaukee has a 73% graduation rate for whites while African-American and Latino students have graduation rates of 34% and 42%, respectively.
In these districts with a large gap between the white and minority graduation rates, it is clear that there are shortcomings in the education system that are particular to minority students. In most of the districts, however, where the white and minority graduation rates are both low, the failure of the education system to produce graduates is a problem that transcends race and ethnicity.
Comparing Graduation Rates to Other Dropout/High School Completion Statistics
Given that local, state, and national governments as well as non-governmental researchers report various statistics on the rate at which students drop out of school or complete high school, it is necessary to describe how the graduation rates reported here compare with some of those other statistics and to explain the differences between them. There are generally four different types of statistics that are reported: event dropout rates, status dropout rates, high school completion rates, and promoting power rates. Let us consider what each of these statistics means and how each is calculated.
An event dropout rate is the percentage of students who drop out of school in a given year. It is not the percentage of students who will eventually become dropouts, it is simply the percentage of enrolled students who leave in a one-year period. Since students tend to drop out between 8th and 12th grade, the event dropout rate only captures one year of what is usually a five or six year span in which students leave school. It is like calculating a credit card interest rate as a monthly percentage instead of an annual percentage: The rate feels low but in truth it compounds over a longer period of time.
A status dropout rate is the percentage of young people (usually 16 through 24 years old) who are not currently enrolled in school and who have not received a high school diploma or a General Educational Development (GED) credential. The status dropout rate more closely reflects what most people imagine when they hear “dropout rate” than does the event dropout rate. The status dropout rate is supposed to capture the proportion of students who leave school and never receive a degree.
The high school completion rate is almost exactly the complement of the status dropout rate (that is, 1 minus status dropout rate). The only difference is that the completion rate is based on surveys of a slightly older population, 18 to 24, instead of the 16 to 24 year olds included in the status dropout rate. If a student dropped out of school when he was 16, he would have an impact on the status dropout rate but not the high school completion rate until he were two years older.
Promoting power resembles in some ways the graduation rate reported in this study. It is the ratio of the number of students in a certain grade to the number that graduate when those students would be expected to graduate.12 Promoting power differs from the graduation rates reported here in that it does not usually make adjustments for aggregate changes in student population to account for the inflow or outflow of students from a given jurisdiction. It also differs in that it usually compares the number of 9th or 10th graders to the number of graduates, rather than 8th graders as in the calculation of graduation rates. Since students may dropout of school in 8th or 9th grades, promoting power may be higher than graduation rates. Yet because 9th grade is a common grade in which to retain students for an extra year, creating an artificially large 9th grade population, promoting power rates that compare 9th graders to graduates may be lower than graduation rates that use 8th graders.
In broad terms, graduation rates should be roughly similar to high school completion rates, promoting power rates, or the complement of status dropout rates. According to a recent report from the National Center for Education Statistics, the national high school completion rate is 86%.13 The national graduation rate calculated in this report is 71%. What accounts for the difference? The bulk of the difference between these two numbers can be explained by the fact that only 77% of students in the NCES report completed high school by receiving a regular diploma. The other 9% counted as having completed high school received an “equivalent” credential, such as a GED.
People who received GEDs or other alternative credentials were not counted in the graduation rates calculated in this report for a number of reasons. First, the purpose of computing graduation rates here was to develop a measure of the success of high schools at graduating students. Recipients of GEDs are not, properly speaking, “graduates” of any high school. The fact that some students leave high school and later receive a credential from a community college, while in prison, or from some other organization cannot be credited to the high school. Similarly, a doctor cannot claim as “cures” patients who have transferred to other doctors for treatment.
Second, the GED is simply not equivalent to a regular high school diploma. Similar effort and knowledge are not necessary to achieve a GED as are necessary to receive a typical high school diploma. Most importantly, the future prospects for recipients of GEDs are significantly worse than the future prospects for recipients of regular high school diplomas. In fact an analysis of national data by Stephen Cameron and Nobel prize winning economist, James Heckman, concludes that: “Exam-certified high school equivalents are statistically indistinguishable from high school dropouts.”14 Other researchers find moderate benefits of receiving a GED for certain groups, but no research supports the claim that the GED is equivalent to a regular high school diploma.15 Counting GEDs in the same group as those awarded regular diplomas masks the true graduation rate.
If we exclude GEDs from the high school completion rate reported by NCES we have a number that is similar to the national graduation rate reported here. Breaking out the results by racial/ethnic groups also reveals similar number for whites and Latinos once GEDs are removed. I report a national graduation rate for white students of 78% compared to a high school completion rate reported by the NCES of 82%. For Latinos I calculate that 54% graduate from high school compared to 55% according to the NCES, once GEDs are excluded.
For African-American students, however, my graduation rates and the NCES high school completion rates remain very different even after GEDs are excluded. I find a graduation rate of 56% for African-American students compared to a 73% high school completion rate according to the NCES. What could account for the difference between these rates for African-American students? The difference may largely be explained by “coverage bias” in the Current Population Survey (CPS) from which the NCES high school completion rates are calculated. The CPS is a very well-conducted survey but like all surveys it has difficulty reaching certain groups of people, particularly low-income minorities, who are not easy to find and interview. Dropouts are disproportionately likely to be among those groups that are difficult for the CPS to find and interview. According to Phillip Kaufman (who was also the primary author of the NCES report on dropouts), if 50% of African-Americans who are not properly covered by the CPS sample are dropouts, then the true high school completion rate for African-Americans would drop by 9%.16 If 100% of those African-Americans not covered were dropouts (a figure that is an upper-bound rather than a realistic assumption), then the true high school completion rate for African-Americans should be adjusted down by 18%. The 17% gap between my African-American graduation rate and the NCES African-American high school completion rate could largely be explained by this CPS coverage bias that could distort results by as much as 18% (but more realistically around 12%).
Other factors may explain the modest differences between my graduation rates and the NCES high school completion rates after excluding GEDs and adjusting for African-American “coverage bias” in the CPS. CPS relies upon self-reported educational status for NCES to compute high school completion rates. That is, people have to describe honestly to the survey researchers whether they received a high school diploma. While most people are likely to be honest, some people deceive themselves or others to hide the embarrassment of dropping out of high school. The self-deception that people have a high school diploma when they really do not may be reinforced by the frequency with which people may falsify resumes to claim that they graduated from high school when they are in fact dropouts. This self-reporting bias may be small, but it may account for much or all of the remaining difference between the graduation rates I computed and the high school completion rates reported by NCES. Duncan Chaplin of the Urban Institute has suggested that self-reporting biases may be more severe among African-Americans “if they felt a greater need to use education as a ‘signal’ to overcome potential discrimination.”17 Because the graduation rates calculated here rely upon enrollment and diploma counts, which are unlikely to be distorted by self-reporting or other biases, they are likely to be slightly more accurate in identifying the percentage of students who complete high school with regular diplomas than a phone survey.
Advantages of Calculating Graduation Rates
If the graduation rates reported here and the national high school completion rates reported by NCES are similar (after excluding GEDs and adjusting for African-American coverage bias), why calculate graduation rates at all? These graduation rates have several advantages. First, they can be calculated with relative precision for states and districts. The CPS simply does not have large enough sub-samples to compute high school completion rates for districts or for ethnic/racial sub-groups in states, so those statistics are not reported by NCES. Having information on the graduation rate for school districts as well as for those districts broken out by race is an important benefit of calculating graduation rates. Even at the state level, high school completion rates are based on small survey populations and have very large confidence intervals around each estimate.18
Second, dropout statistics derived from the Current Population Survey are based on young people who live in an area but who may not have gone to high school in that area. This fact may create a fairly large bias in areas with fast growing populations related to higher-skilled economic development. The graduation rates reported in this study more directly measure the success of schools in each jurisdiction to produce graduates.
Third, the Current Population Survey does not include in its sample people who are incarcerated. Since dropouts are disproportionately represented among people in prison, this is likely to overstate the graduation rate. This bias is more severe for ethnic or racial groups that have a disproportionate number of young people in prison.
Fourth, the self-reporting bias in CPS is especially severe when it comes to distinguishing GED recipients from regular high school graduates. As Duncan Chaplin of the Urban Institute put it: “The major problem with the CPS data is that information on GED status appears to be very inaccurate.”19 Chaplin reports that more than 60% of people initially described as GED recipients in the first survey are later described as regular high school graduates when re-surveyed the next year. As Chaplin explains: “it appears that there is a very large amount of random misreporting of GED status in the CPS, perhaps because respondents are rushing to answer questions quickly and/or because they are not aware of the GED status of teenagers living in their households.” Chaplin also reports that the number of new GED recipients according to the CPS is less than half the actual number of GEDs awarded according to the GED Testing Service. In short, the lack of quality results on the number of GED recipients in the CPS undermines the reliability of its estimate of high school graduates.
The advantage of the calculation of graduation rates reported here is that it relies on enrollment and diploma numbers as collected by NCES. Those enrollment and diploma numbers do not suffer from sample coverage biases because there is no sampling involved. Diploma and enrollment numbers are not biased by excluding prison populations. Diploma and enrollment numbers more directly measure the performance of school systems in an area than surveys of young adults in the area who may not have attended school locally. And diploma and enrollment numbers do not suffer from confusion about who has a GED or a regular diploma or other self-reporting biases.
Dropout Statistics Reported by Districts and States
The NCES report also contains state event dropout rates that have been collected from the states rather than from the CPS. Essentially, they survey the states and ask them for event dropout rate statistics. In addition to the fact that these rates have to be compounded over several years to produce something equivalent to a status dropout rate, which is what most people have in mind when they discuss dropout rates, there are serious reporting problems with event dropout rates. Only 37 states report event dropout statistics to NCES and of those only “26 said that they adhered exactly to the standard definition and collection procedures” outlined by NCES.20 The frequency of missing and incomparable data make these event dropout rates unhelpful for trying to compare the effectiveness of different states at graduating their students.
Event dropout rates reported directly by states and districts are subject to severe self-reporting problems and are often implausibly low. Rather than relying on a survey, like CPS, districts and states calculate their own event dropout rates by asking school officials to track individual students and report the percentage of students in certain grades who drop out during the year. The self-reporting bias stems from the fact that we are depending upon school officials to track the status of individual students. Because school systems and their officials are under strong pressure not to have high dropout rates, they have incentives to assume that students moved out of town or fell into some other category that exempted them from being called dropouts. In Austin, TX the mis-reporting of dropout and other accountability statistics was so severe that the entire district was criminally indicted. As a result of an agreement to settle the case the event dropout rate was re-calculated and the district’s rate more than doubled.21
Even when event dropout rates are not willfully or negligently under-reported, school officials usually do not have the resources or skills to attempt to track individual students and compute an event dropout rate. Ironically, the attempt by school officials to compute dropout statistics by tracking individual students is supported by the claim that it is more “precise.” The truth is that it is far more precise to examine cohorts of students by comparing enrollments to graduation counts (with adjustments for population changes), as I have done with graduation rates. Computing dropout rates by trying to track individual students is like trying to measure how much rice you have eaten in a month by summing the weight of every grain that was cooked. There is measurement error when each grain is weighed and some grains are “lost” by sticking to the side of the pot. It is much more accurate and cost-efficient just to weigh the bag at the beginning and end of the month. It sounds more precise to track the individual grains but it ends up being much less precise.
Using a method that involves trying to track individual students the Dallas Independent School District in Texas reports an annual dropout rate of 1.3%.22 This number is implausibly low. Consider that according to my calculations Dallas has a graduation rate of only 52%. Even if 1.3% compounded over several years it does not come close to matching the picture drawn by my graduation rate. If only 1.3% of students dropout each year, how is it that Dallas had 9,914 students in 8th grade in 1993 but only 5,659 graduates in 1998 while the total student population in the district went up by 10.5%? It cannot be that several thousand students moved out of town while the whole city and school district population was increasing. It cannot be that thousands of students were held back a grade and that no students were held back a grade in the cohort from the year before. Frankly there is no reasonable explanation for what happened to those several thousand students in Dallas other than that they dropped out, making the 1.3% event dropout rate simply unbelievable.
This example illustrates another reason why the graduation rates in this study are beneficial to compute and report. They are easy to calculate, they are consistent with the common sense notion that thousands of missing students are probably dropouts, and they are a nice reality check on implausible official numbers.
The reporting of implausible dropout rates is not confined to Dallas. The state of Texas reports a 1.6% annual dropout rate while I calculate a graduation rate for the state of 68%.23 If it is true that only 1.6% of students in Texas drop out of school each year, what explains the fact that there were 274,208 8th graders in Texas in 1993 and only 197,186 graduates in 1998 while the state’s student population increased by 5.9%? The state of California reports an annual or event dropout rate of 2.8% while I calculate a graduation rate of 73%. If the 2.8% figure is correct, then how did California go from having 380,223 8th graders in 1993 to 282,897 graduates in 1998 while the state’s total student population increased by 2.1%?
New York City claims that only 19.3% of the class of 2000 dropped out of high school. I found a graduation rate for the class of 1998 of 55%. What explains the difference? The New York City report admits that only 50% of the class of 2000 actually graduated, while 31% continued to work toward a degree. 24 The truth is that very few of those 31% receive regular high school diplomas, yet the city’s method of calculating results generously excludes all of them from the dropout category.25 This would be like an accounting system that excluded from the delinquent accounts category everyone who said that they were working on paying their invoice. Not counting those who say “the check is in the mail” among the delinquent accounts presents a grossly distorted financial picture. In New York City, 31% of all high school students have the check in the mail.
Some districts, however, appear to be willing to be brutally honest in reporting their dropout/graduation situation. For example, the Charlotte/Mecklenburg district in North Carolina reports that only 47% of their African-American students in the class of 1999 graduated high school. I calculated the graduation rate as 53% for African-American students in Charlotte. The district places their total graduation rate at 54% compared to my calculation of a total graduation rate of 63%.26 The district’s numbers may be too harsh (it is not clear whether they adjusted for the population increase in the district), but at least Charlotte is willing to face its problems and discuss them openly.
The lack of candor about the rate at which public school students graduate high school is a fundamental problem in education. The rates at which students graduate high school provide us with information about the effectiveness of those schools. Unless we have reliable information about graduation rates we cannot begin to consider the severity of problems or make comparisons about the effectiveness of schools in different areas or for different groups of students. The graduation rates provided here provide simple, straightforward, and accurate information about schools nationally, in each state, and in the 50 largest school districts, as well as for racial/ethnic sub-groups, facilitating discussions about the severity of problems as well as comparisons about those problems.
The results are consistent with high school completion rates reported by the NCES (after GEDs are excluded and African-American coverage biases are adjusted), but this report expands upon the NCES report by providing graduation rates for states, districts, and ethnic/racial sub-groups that are not provided by the NCES. This report also improves upon state and district reported dropout rates, which unfortunately often implausibly understate problems.
The graduation rates reported here have to be seen as part of the beginning of a discussion and not the final word. This report does not consider why graduation rates are what they are. It does not attempt to explain why rates are lower for some areas or for some populations. And it does not attempt to compute whether these rates are lower or higher than they were in the past.27
The graduation rates reported in this study, however, convey strongly that far fewer students are graduating high school than we may have believed and far fewer than we would wish. The graduation rates are shockingly low for African-American and Latino students nationwide. We also see far too many states and school districts with remarkably low graduation rates. But there is also hope in these numbers. Some districts appear able to graduate a relatively high percentage of African-American, Latino, and white students. We should begin to examine those districts to see if there are formulas for success that can be imitated elsewhere. And where we see severe problems we should be more open to new ideas for how to revitalize our schools and improve those situations.
- See http://ferret.bls.census.gov/macro/032000/perinc/new03_001.htm
- Phillip Kaufman, Jin Y. Kwon, and Steve Klein, "Dropout Rates in the United States: 1999," National Center For Education Statistics, Statistical Analysis Report, November 2000, p. 1.
- Of course, some school systems may increase their graduation rates by having lax standards for receiving a high school diploma while other school systems may experience lower graduation rates by having more rigorous standards for receiving a diploma. This report makes the simplifying assumption that the standards for high school graduation are relatively consistent throughout the United States. It would be interesting in future research to relax this assumption and examine the potential inter-action between the rigor of school standards and graduation rates.
- Phillip Kaufman, "The National Dropout Data Collection System: Assessing Consistency," Harvard Civil Rights Project, January 13, 2001. Available on the web at: http://www.law.harvard.edu/civilrights/publications/dropout/kaufman.html
- For simplicity of language I use the terms "white" for non-Hispanic whites and "African-American" for non-Hispanic African-Americans. I use "Latino" to refer to Hispanics of any racial group.
- Most enrollment and diploma numbers were obtained from the Common Core Data (CCD) from the National Center for Education Statistics of the U.S. Department of Education. If the data were not available from CCD, the information was requested from the state or school district.
- I chose to use 8th grade enrollments because some students drop out of school before 9th grade. In addition, 9th grade is a common grade in which students repeat the grade, which can artificially inflate 9th grade enrollments and understate the true graduation rate.
- I decided that the results were not sufficiently reliable if there were fewer than 150 students in 8th grade because the results would be too sensitive to population inflow and outflow in ways that would be difficult to detect and adjust.
- Some of the total population change is from changes in birth rates or population inflow at younger grades, but adjusting the 8th grade enrollment by the total student population change is the most parsimonious assumption for an adjustment and it is still likely to be reasonably accurate. If the total population changes more in the younger grades, then the graduation rate will be slightly underestimated. However, the total student population change can also be influenced by a high rate of dropouts that could cause the graduation rate to be overestimated. In sum, there is little reason to expect systematic bias from this adjustment and it is likely that any errors are small.
- The reviewers of an earlier draft of this report suggested other alternative explanations for enrollment changes. For example, some of the reviewers worried that urban enrollments in particular might decline after 8th grade if families switched to private or suburban schools in large numbers. In fact, far fewer students are enrolled in private high schools than are enrolled in private elementary and middle schools. (See http://nces.ed.gov/pubs2002/digest2001/tables/dt060.asp. So, the net flow of students is into public schools after 8th grade, meaning that the most likely bias here is that I have overstated public high school graduation rates. Similarly, there is no evidence of a large shift of students from urban to suburban schools after 8th grade because suburban districts do not show an increase in high school enrollments relative to their primary grade enrollments. See for example that Montgomery and Prince Georges counties do not experience an increase in student enrollment for high schools: http://nces.ed.gov/pubs2000/100largest/table5.html
- The results broken out by ethnicity are based on 41 states for which data were available in the Common Core Data or were provided by the states. Data were requested from all states but Arizona, Idaho, Kentucky, North Dakota, New Hampshire, South Carolina, Utah, Vermont, and Washington either did not have the data available or refused to respond. Fortunately, most of these states have relatively small minority populations, making it unlikely that their exclusion distorts the national graduation figures for African-Americans, Latinos, and whites.
- For an example of research using promoting power see: Robert Balfanz and Nettie Legters, "How Many Central City High Schools Have A Severe Dropout Problem, Where Are They Located, and Who Attends Them? Initial Estimates Using the Common Core of Data," Harvard Civil Rights Project, January 13, 2001. Available on the web at http://www.law.harvard.edu/civilrights/publications/dropout/kaufman.html
- Phillip Kaufman, Jin Y. Kwon, and Steve Klein, "Dropout Rates in the United States: 1999," National Center For Education Statistics, Statistical Analysis Report, November 2000, Table 4, p. 19.
- Stephen Cameron and James Heckman, "The Nonequivalence of High School Equivalents," Journal of Labor Economics, volume 11, number 1, 1993, p. 1.
- See for example, Richard J. Murnane, John B. Willett, and Kathryn Parker Boudett "Do High School Dropouts Benefit from Obtaining a GED?" Educational Evaluation and Policy Analysis, 17(2), 1995, pp. 133-147.
- Phillip Kaufman, "The National Dropout Data Collection System: Assessing Consistency," Harvard Civil Rights Project, January 13, 2001, Table 1. Available on the web at http://www.law.harvard.edu/civilrights/publications/dropout/kaufman.html
- From Duncan Chaplin's review of an earlier draft of this report, September 20, 2001.
- Phillip Kaufman, "The National Dropout Data Collection System: Assessing Consistency," Harvard Civil Rights Project, January 13, 2001, Figure 5. Available on the web at http://www.law.harvard.edu/civilrights/publications/dropout/kaufman.html
- Duncan Chaplin, "GEDs for Teenagers: Are There Unintended Consequences?" Urban Institute, November 26, 1999. Available on the web at: http://www.urbaninstitute.org/education/ged.html
- Phillip Kaufman, Jin Y. Kwon, and Steve Klein, "Dropout Rates in the United States: 1999," National Center For Education Statistics, Statistical Analysis Report, November 2000, Table 2, p. 10.
- For a statement from the Austin Independent School District on the criminal indictment of the district see: http://www.austin.isd.tenet.edu/newsmedia/releases/oldarchive/response.html
- See: http://www.tea.state.tx.us/perfreport/aeis/index.html
- See: http://www.tea.state.tx.us/perfreport/aeis/2000/state.html
- See: http://www.nycenet.edu/daa/reports/index.html
- It is true that some of these students receive GEDs. According to a New York City report on the class of 1997 who were tracked until the year 2000, 69.7% received a degree of some kinds, but 14.9% of those degrees were equivalency degrees. If we exclude those GEDs, then New York City is reporting a 59% graduation rate for the class of 1997 compared to my graduation rate of 54% for the class of 1998. See "The Class of 1997, Final Longitudinal Report, A Three-Year Follow-up Study" Table 1, p. 5. Available at: http://www.nycenet.edu/daa/reports/index.html
- See: http://www.cms.k12.nc.us/departments/instrAccountability/schprofile/schoolprofiles.as p
- An interesting and reliable way of viewing graduation trends over time is to consider a statistic reported in the Digest of Education Statistics, 2000, Table 101. It reports the ratio of regular high school graduates (excluding GEDs) to the total 17 year-old population in the United States going back as far as 1870. This ratio is a reasonable approximation of a national graduation rate and can be consistently calculated for more than a century. The table shows that graduation rates steadily climbed to a peak of 77.1% in 1969 and have since fallen back to 70.6% in 2000, a level that was first achieved in 1963. Available on the
web at: http://nces.ed.gov/pubs2002/digest2001/tables/dt101.asp