Gauging Graduation, Pinpointing Progress
Districts with higher rates offer bright spots.
Some observers have argued that graduation rates—for the nation as a whole, most demographic groups, and the majority of states—are failing to reach a level necessary to put the United States on a solid footing in a competitive global economy. Despite that arguably worrisome state of affairs, the longer-term trajectory of change for the country’s graduation rate does offer some reason to be cautiously optimistic.
From 1996 to 2006, the most recent year for which data are available, the national graduation rate for U.S. public high schools rose by 2.8 percentage points. That gain, averaging about three-tenths of a point annually, signals slow but steady progress over the past decade. In fact, for each of the past six years, the nation’s graduation rate has stayed consistently above the 1996 benchmark level of 66 percent.
The latest original analysis of high school completion conducted by the Editorial Projects in Education Research Center places the national graduation rate at 69.2 percent for the class of 2006. The center calculates graduation rates for the nation, states, and every school district in the country using the Cumulative Promotion Index (CPI) method and data from the U.S. Department of Education’s Common Core of Data (CCD). More information on the CPI methodology can be found on Page 38 of this report.
Long-term improvements can be found for all major demographic groups, although gains have been considerably stronger among non-Hispanic whites than for racial and ethnic minorities. But in other respects, progress has been more rapid in areas where graduation rates have historically languished. High school completion rates have risen much faster in high-poverty school systems as opposed to more-affluent ones, and in urban vs. suburban communities. Stronger improvements have also been found in larger districts and, to a lesser extent, in those serving majority-minority student populations.
During the past decade, graduation rates improved, at least marginally, in a solid majority of states (34), with several—Arizona, South Carolina, and Tennessee—experiencing double-digit gains. However, rates fell noticeably (by at least 1 percentage point) in 10 states. The largest drop occurred in Nevada, which has undergone extraordinary growth and demographic change in recent years.
Overall, improvements in graduation rates during the past decade have been widespread. Nearly every type of school district has posted gains. But the pace of change has been most rapid in places with historically low rates of completion, particularly large, high-poverty, and big-city school systems.
Despite signs of progress, three out of every 10 students in America’s public schools still fail to finish high school with a diploma. That amounts to 1.3 million students lost from the high school pipeline every year, 7,150 students lost every school day. Many of those nongraduates, who disproportionately come from the nation’s largest urban school systems, are living in poverty, and most are members of historically disadvantaged minority groups.
The results of the analysis for the class of 2006 show a great deal of variation—both across student groups and from state to state—around the nation’s average graduation rate. While more than three-quarters of white and Asian students earn a high school diploma, outcomes are much more troubling for other demographic groups, only about half of whom graduate. Among Latinos, 55 percent successfully finish high school, while just 51 percent of African-American and 50 percent of Native American students do.
For the nation’s students who are most at risk, the odds of earning a diploma amount to a toss of the coin.
An even larger gulf, however, separates the highest- and lowest-performing states. Among the nation’s leaders, Iowa, New Jersey, and Wisconsin each graduates more than 80 percent of all high school students. At the opposite end of the spectrum, just under half of students finish high school in the District of Columbia and Nevada. A gap of 35 percentage points divides the top and bottom states. Overall, half the states have graduation rates in the 65 percent to 75 percent range for the class of 2006.
In another series of analyses, the EPE Research Center also examined the pattern of year-by-year changes in national and state graduation rates, relative to baseline 1996 levels. Those results reveal considerable diversity in trajectories of change, even among states with similar overall increases (or decreases) over the past 10 years.
During most of the past decade, graduation rates slowly and steadily rose for public high school students as a whole and for all major racial and ethnic subgroups. In 2006, however, graduation rates dropped across all groups, eroding much of the progress made in recent years.
New York and Texas, for example, both saw gains of about 7 percentage points from 1996 to 2006. But it was only in the second half of that period, after rebounding from an earlier decline, that New York rose above its baseline 1996 graduation rate. Texas, by contrast, surpassed its baseline rate by 1998 and has maintained itself in positive territory ever since.
In Nevada, which has experienced the largest long-term declines, the graduation rate dropped between 1996 and 1997 and has remained below the 1996 baseline level for a decade, a pattern fueled by sizable annual declines in four of the past 10 years.
While long-term trends have generally been encouraging, the epe Research Center found that the nation’s graduation rate dropped markedly—by almost a point and a half—between 2005 and 2006. That is the first significant annual decline found in more than a decade.
That downturn, also observed by U.S. Department of Education researchers, appears to be quite far-reaching. Graduation rates fell nationally for all major racial and ethnic groups. Nearly half of states experienced a comparable decline of at least 1 percentage point, with the remainder about evenly split between those holding steady and those seeing improvements from 2005 to 2006.
The causes of those declines remain unclear. And it is too soon to know whether the phenomenon represents a temporary fluctuation or will mark a more lasting departure from the general upward trajectory the nation has been traveling over the past decade. The breadth of the decline does raise some reason for concern, however, and at the very least suggests a need for closely monitoring trends in the coming years.
Consisting of graduation, demographic, and structural information for every school district in the country, the extensive database created by the EPE Research Center for Diplomas Count 2009 offers an opportunity to investigate patterns of high school completion across the country.
An EPE Research Center analysis identified a pool of school districts matching the profile of the nation’s largest urban systems and then singled out those demonstrating the highest graduation rates and strongest growth, relative to expectations based on district characteristics. Thirty-three of these big-city peer districts posted 2006 graduation rates at least 10 percentage points higher than anticipated, while 27 districts from across the country surpassed expected graduation-rate improvements between 1996 and 2006 by 10 points or more.
Using that district-level database, we performed an in-depth analysis to identify the leading factors associated with both high graduation rates in 2006 and strong improvements over time. Those correlation analyses identified a set of 10 school district characteristics that are systematically associated with high school completion.
Such factors as larger district size (measured in terms of student enrollment), higher student-teacher ratios, and urban location are linked to modestly lower graduation rates, while high concentrations of poor and minority students exhibit a much stronger negative influence. High school completion levels, on average, are slightly higher in districts that have larger secondary schools and devote higher proportions of their spending to instructional expenses.
It should be noted that those analyses focus on the overall statistical association between graduation rates and the respective district features. In this instance, we did not attempt to quantify the unique, independent effects of those factors by, for example, accounting for the fact that a district may be simultaneously large and impoverished, and that these characteristics may both be related to high school completion patterns.
The analyses of 2006 graduation rates revealed a strong and consistent set of predictors quite similar to those cited as influencing a variety of other educational outcomes, including tested achievement.
In a companion analysis, the research center examined factors associated with changes in graduation rates from 1996 to 2006. As is often the case, patterns of change proved more difficult to explain. Statistical relationships in change analyses tend to be weaker and occasionally run in a direction counter to those found when taking a snapshot of conditions at a single point in time.
Districts that enjoyed greater improvements during the past decade tend to have larger secondary schools, higher student-teacher ratios, and more instructional spending. School systems characterized by concentrated poverty and those with higher overall per-pupil spending experienced smaller gains, or declines.
We found no measurable statistical connections between changes in the graduation rate and district size, urban location, minority enrollment, or racial or socioeconomic segregation levels.
The results above provide useful, albeit broad, insights about the factors related to high school graduation. But we can also build on that analytic framework to identify particular districts that are exceeding expectations. That is, we can determine which school systems have higher graduation rates than we would anticipate given their distinctive profiles, as defined by 10 characteristics, including size, location, poverty levels, and demographics.
The first step in this exercise involves calculating the expected graduation rate for each district in the country.
This type of statistical modeling, akin to but more sophisticated than the correlational analysis described earlier, essentially determines the strength (strong or weak) and direction (positive or negative) of the independent relationships between a series of specified district characteristics and the graduation rate. Those model results are, in turn, used to generate a predicted graduation-rate value for each district, based on its individual profile of demographic and structural features.
The EPE Research Center projected the number of graduates and nongraduates for the class of 2009 by multiplying the 2005-06 graduation rate by the number of 9th graders enrolled that year. The areas of the circles are proportional to the number of students failing to graduate nationwide and for each state. Nationally, 1.3 million members of this year’s graduating class will not earn diplomas, with American public schools losing nearly 7,200 students each school day.
For instance, results of that statistical modeling show that poverty exerts a strong negative influence on graduation rates. As a result, a high-poverty district will tend to have a lower expected graduation rate than one serving a more affluent population, all else being equal.
Our findings show that most school districts are performing at roughly the level we would expect given their size, poverty rate, concentrations of minority students, per-pupil spending levels, and so forth. Yet we also found that a substantial number of districts are exceeding expectations, with graduation rates substantially higher than those of other school systems that fit a similar profile.
Nationwide, nearly 2,200 school districts exceed expectations for class of 2006 graduation rates by a margin of at least 10 percentage points. In a parallel analysis of changes in those rates between 1996 and 2006, we find a similar number of districts with higher-than-expected levels of improvement over the past decade.
Research has consistently shown that large urban districts disproportionately contribute to the nation’s graduation crisis. The April 2009 “Cities in Crisis” report from the EPE Research Center, for example, estimated that the school districts serving the nation’s 50 largest cities alone account for more than one-quarter of all students in the entire nation who fail to graduate.
To focus our investigation on the hardest-hit communities, we developed a matching algorithm that allowed us to further narrow our results to a set of districts that closely fit the structural and demographic features of the nation’s largest urban school systems. Because this set of peer districts shares a wide array of underlying characteristics and challenges, they offer highly relevant points of comparison for one another. So, despite conditions and histories that are distinctive to some extent, lessons applicable to one of these districts should be highly relevant to others in the larger group.
Within this group of big-city peer districts, our final analysis identifies a total of 50 school systems that exceed expectations on either current graduation rates or improvements. Ten of those districts appear in both categories.
These overachieving districts are spread throughout the country, with some of the strongest performers found in communities as far-flung as Euclid City, Ohio; Leon County, Fla. (Tallahassee); Phenix City, Ala.; Stockton, Calif.; Texarkana, Texas; and Warren Township, Ind. (Indianapolis).
The Stockton Unified School District, to take just one example, posted a graduation rate of 77 percent for the class of 2006, nearly 17 percentage points higher than would be expected given its large size, urban environment, high rates of poverty, high degree of racial and socioeconomic segregation, and low levels of per-pupil spending.
Stockton Unified also ranks among the top tier of big-city peer districts that surpass expectations for graduation-rate improvements between 1996 and 2006. Although the school system’s characteristics would have predicted a small decline over the past decade, the graduation rate in Stockton actually rose by 29 percentage points during that period.
What is it, exactly, that separates the likes of Stockton, Texarkana, and Leon County from the rest of the pack and explains their success? The answer to that critical question, unfortunately, is beyond the scope of our study.
We can say with some confidence, though, that the answer probably has little to do with such traditional explanatory factors as size and poverty, because by national standards all those districts are large and serve high proportions of poor students. The explanation must lie elsewhere, likely in the districts’ strategies for high school reform, programmatic innovation, and other efforts to promote the success of all students.
Further research and case-study analysis will be necessary to effectively derive actionable lessons from the school systems that have emerged as potential leaders and exemplars.
Even if the EPE Research Center’s current analysis cannot provide definitive answers, we hope it offers a place to start for policymakers, education leaders, researchers, and others grappling with the graduation crisis in their own states and hometowns.
There are success stories out there, and they can be found even in the nation’s most at-risk communities. And while each of those stories may be unique, such districts’ experiences may offer important lessons and insights that serve as the basis for transformation in other school systems across the country.
Vol. 28, Issue 34, Pages 24,30-31
Get more stories and free e-newsletters!
- Richland School District One, Columbia, SC
- Fort Worth Independent School District, Forth Worth, TX
- Science Teacher
- Cle Elum-Roslyn School District, Cle Elum, WA
- Director of Math and Science
- Reading School District, Reading, PA
- Director of Special Education
- Reading School District, Reading, PA