The graduation statistics reported in Diplomas Count come from an extensive EPE Research Center database that captures detailed information for every school system in the United States. This collection of indicators for graduation rates, demographic composition, structural features, and other district characteristics offers an unparalleled opportunity to investigate patterns of high school completion across the country.
Leveraging the data offers a way to answer such basic questions as: What district conditions are most strongly related to lower graduation rates, or higher ones? More importantly, deep engagement with data can also lend insight into the more subtle dimensions of high school performance: What would we expect a particular district鈥檚 graduation rate to be, given its specific profile of demographic and structural characteristics? Which school systems are beating the odds when it comes to graduation?
And, can we find examples of overachievers among the nation鈥檚 most at-risk communities? The short answer: Yes. Read on for details.
An original nationwide analysis exploring a wide range of factors potentially linked to high school completion identified a core set of 10 district characteristics that demonstrated consistent relationships to graduation rates: district enrollment, average high school size, student-to-teacher ratio, urbanized location, racial composition, poverty level, race- and poverty-based segregation, per-pupil spending, and the share of expenditures devoted to instruction.
Specifically, the research center found that larger district sizes (measured by student enrollment), higher student-to-teacher ratios, an urban location, and higher spending levels on a per-pupil basis are systematically associated with slightly to moderately lower graduation rates. Much stronger negative impacts are linked to high concentrations of poor or minority students and severe segregation along racial or socioeconomic lines. On the other hand, more students earn diplomas in districts with larger secondary schools and those devoting higher proportions of their budgets to instructional expenses.
Moving beyond basic correlations, it is possible to build on this framework of key predictors to construct a statistical model that determines an anticipated graduation rate for any given school district, based on its distinctive profile, as defined by size, location,poverty level, demographic composition, and other core characteristics.
Urban Overachievers
An EPE Research Center analysis identified a pool of school districts matching the profile of the nation鈥檚 largest urban systems and then singled out those demonstrating the highest graduation rates, relative to expectations based on district size, poverty level, and other characteristics. Among the 151 districts in this big-city peer group, 21 school systems posted graduation rates for the class of 2007 at least 10 percentage points higher than anticipated.
SOURCE: EPE Research Center, 2010
This type of statistical modeling鈥攁kin to, but more sophisticated than, the correlational analysis described earlier鈥攅ssentially determines the strength 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. For instance, the statistical modeling shows that poverty exerts a strong negative influence on graduation rates, over and above other factors also related to high school completion. 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.
Defying Expectations
A substantial body of research, including results presented elsewhere in Diplomas Count 2010 (see Page 25), has consistently found that the brunt of the nation鈥檚 graduation crisis is borne by certain school systems, most notably those serving large urban centers. As a result, a relatively small number of districts generate an outsize share of the nation鈥檚 total dropouts.
To focus the EPE Research Center鈥檚 investigation on the hardest-hit communities, the center developed an algorithm to further narrow results of its statistical modeling to a set of districts that closely fit the structural and demographic features of the largest urban school systems. Matched against the set of 10 core characteristics linked to graduation, the center identified a group of 151 urban peer districts with highly similar profiles that are also likely to signal common underlying challenges.
Within this group of big-city peer districts, the final analysis highlights 21 鈥渙verachieving鈥� school systems鈥攖hose exceeding expected graduation rates for the class of 2007 by at least 10 percentage points. Twenty-seven peer districts fall short of expectations by a similar 10-point margin, with the balance of this big-city group (103 school systems) performing close to the levels predicted.
At the top of the overachievers list is the Newport-Mesa Unified School District. Located in Newport Beach, Calif., the district posted a graduation rate of 86 percent, 29 percentage points higher than would be anticipated given its large size, urban environment, high degree of racial and socioeconomic segregation, and spending patterns. Five other school systems exceeded expectations by a margin of around 20 points: David Douglas, in Portland, Ore.; Jonesboro, Ark.; Memphis, Tenn.; the Texarkana Independent district in Texas; and Visalia Unified, in California.
