An earlier version of this report was presented at the Making Dropouts Visible conference
sponsored by The Civil Rights Project of Harvard University and Jobs for the Future and convened at Teachers College, Columbia University on June 3, 2003.
The nonpartisan Urban Institute publishes studies, reports, and books on timely topics worthy of public consideration. The views expressed are those of the authors and should not be attributed to the Urban Institute, its trustees, or its funders.
Note: This report is available in its entirety in the Portable Document Format (PDF).
Table of Contents
Executive Summary
1. Introduction
2. No Child Left Behind, Accountability, and High School Graduation
2.1. Accountability Under NCLB
2.2. An Other Academic Indicator – High School Graduation
2.3. Flexibility for Graduation Rates
3. Measurement Considerations
4. Alternative Graduation Indicators
4.1. Longitudinal Graduation Rate (LGR)
4.2. National Center for Education Statistics (NCES) Rate
4.3. Adjusted Completion Ratio (ACR)
4.4. Cumulative Promotion Index (CPI)
5. Data and Methods
6. Results
6.1. National Graduation Rates – Indicator Quality and Estimates
6.3. State Graduation Rates – Indicator Quality and Estimates
6.4. Graduation Rates for Racial and Ethnic Subgroups
7. Discussion and Conclusion
8. References
Executive Summary
The No Child Left Behind Act of 2001 (NCLB) requires that the nation's public elementary and secondary school systems be held accountable for achieving high levels of educational proficiency for all students. Specifically, NCLB mandates that all states must establish performance-based accountability systems that include: clear standards and goals for improvement; rigorous methods of measuring progress towards established performance targets; and high-stakes consequences for both individual schools and school systems that fail to make sufficient progress in reaching the goal of universal student proficiency.
An important tenet of the law is that all students should be held to the same high standards. Toward this end, each state must establish an accountability system that has separate performance goals in reading and mathematics with all schools reaching 100 percent proficiency in each area within twelve years. Accordingly, to avoid identification as "needing improvement" school systems must meet annual academic proficiency goals established by the state that are designed to make steady progress toward the twelve year target of universal proficiency. In addition, the accountability system holds schools and districts responsible for the adequate yearly progress for students in specified subgroups, including the major racial and ethnic groups.
While achievement testing will be the central component of state accountability systems under NCLB, these systems must include graduation rates as an academic accountability indicator at the high school level. Earning a high school diploma is an important outcome in its own right and a strong predictor of future social and economic success. In addition, holding schools and districts accountable for academic achievement levels, based on both test scores and high school graduation rates is intended to help safeguard against an unintended negative incentive of high-stakes accountability systems — that pushing low-performing students out would raise test scores and help struggling schools meet test-based accountability benchmarks.
This study seeks to inform the on-going debate over accountability under these new NCLB requirements. Specifically, it examines the changing role of high school graduation rates with particular attention to the implications for racial and ethnic minority groups and the school systems that serve these students. We begin by outlining the accountability provisions of NCLB that pertain to high school graduation and then discuss their implications for developing empirical indicators to be used in high-stakes accountability systems. Several distinctive measurement strategies are examined: an approached used by the U.S. Department of Education's National Center for Education Statistics (NCES); an Adjusted Completion Ratio (ACR); and the Cumulative Promotion Index (CPI), a new indicator developed by the author that may offer particular advantages on both conceptual and operational grounds. These three methods were chosen for two main reasons. First, each adheres (at least in general terms) to the provisions of NCLB and, therefore, offers an alternative relevant to decision-makers designing state accountability systems. Second, each of these rates can be calculated using the kinds of non-longitudinal group-level data readily available in most states.
In the empirical portion of this study we calculate graduation rates based on these three distinctive methods using data from the Common Core of Data, the U.S. Department of Education's census of local educational agencies and schools. Rates are estimated for the nation as a whole and for each of the states, data and methods permitting. We then systematically compare the results generated by the three approaches, with respect to: the quality of the indicators from an informational perspective (i.e., coverage of school districts or student population) and differences in estimated graduation rates across methods.
The final set of empirical analyses examines national and state graduation rates disaggregated by major racial and ethnic categories, as is required under NCLB. Of the three alternative approaches explored, however, we find that only the CPI method is able to overcome the technical challenges involved in calculating reasonably reliable graduation rates for individual racial and ethnic categories.
The main results from this study indicate that graduation rates are likely to be much lower than the 85 to 90 percent that prevails in the conventional wisdom. In fact, as few as two-thirds of high school students nationwide may graduate with a regular diploma. Even that statistic, which would strike many observers as surprisingly and unacceptably low, overlooks truly abysmal performance that prevails in certain places. Although in a small number of states graduation rates reach as high as 80 percent, in other states the odds of graduating from high school fall below fifty-fifty for the average student. While the CPI and ACR methods typically generate very similar results, we find that the NCES indicator always produces a higher estimate of the state graduation rate than both of the alternative measures. This overestimation (about 10 percent on average) is likely to be attributable to the NCES method's heavy reliance on dropout counts, which tend to be underreported for any number of reasons. The other two indicators use information only on graduation and enrollment counts.
Findings for disaggregated graduation rates using the CPI method demonstrate well-known disparities among racial and ethnic groups. Nationwide, we find a graduation rate of close to 75 percent for White and Asian students (a level approaching the average graduation rate in the highest-performing states). By contrast, graduation rates for Native American, Hispanic, and African American students nationwide hover at around 50 percent (about same level found for the average student in the lowest performing states). Our results further suggest that, even in many generally high-performing states, students from these particularly disadvantaged minority groups often graduate from high school at rates as much as 20 to 30 percent lower than their White and Asian American peers.
In the final portion of the paper, we discuss the study's implications for the development high-stakes educational accountability systems like those required by NCLB. One clear and consistent theme that emerges from our findings is that graduation rate estimates (like the NCES indicator) that are heavily dependent on dropout counts should be viewed with considerable skepticism. One root of this problem is the largely inadequate state of our national systems for defining and collecting information about dropping out. We find that the NCES indicator cannot be calculated for the majority of states and districts nationwide due to a lack of consistent information on the number of students who drop out of high school. This problem takes on a particularly policy-relevant dimension when we realize that many of the states missing dropout data are also home to large and rapidly growing minority populations.
But beyond the issue of data availability, we also find that even in cases where values can be computed dropout-dependent measures appear to consistently and considerably overestimate graduation rates compared to other available estimates. The best evidence available clearly recommends a rapid move away from a reliance on graduation rates that employ dropout data and towards alternative strategies for more accurately measuring this key indicator of educational performance. Unfortunately, it appears that many states are proposing to use an NCES-like approach or other methods dependent on dropout data for calculating graduation rates as part of the statewide accountability systems they are required to implement under the terms of NCLB.
In conclusion, two recommendations are offered in the interest of promoting informed decision-making in this critical area of policy implementation. Ideally, we should establish scientifically-based standards for calculating graduation rate indicators consistent with the way NCLB requires that assessments be used for purposes for which they are valid and reliable, consistent with relevant, nationally recognized professional and technical standards. Therefore, as graduation rates must be included in AYP, states should ensure that this indicator is made more reliable. Second, safeguards must be established to ensure that data reported about graduates, dropouts, and enrollments are not only complete but also accurate. After all, a graduation rate will only be as good as the data used to calculate it. The U.S. Department of Education could play a central role in supporting both of these efforts. Steps such as these will help to ensure that all states are held to the same high standards. This may, in turn, instill public confidence in the accountability regimes that will become a prominent feature of public educational landscape as the new law of the land is implemented in the coming years.
Note: This report is available in its entirety in the Portable Document Format (PDF).
Latest Research
Browse by Topic:
PDF
Printer-friendly summary
