In 1948, the U.S. Supreme Court ruled that Oklahoma's public university system, which only admitted white students, violated the equal protection clause of the 14th Amendment to the Constitution.¹ In addition to setting the stage for the historic Brown v. Board decision, the ruling set in motion a decades-long process of higher education desegregation in Oklahoma and across the nation. In order to track the racial makeup of the state's universities, the Oklahoma State Regents for Higher Education built one of the nation's first comprehensive databases of college students. Today, that database shows potential to address an entirely different challenge: how to improve high school accountability under the federal No Child Left Behind Act. Current NCLB measures, it seems, aren't telling the whole story about high school achievement.

In recent years, a host of business, charitable, and governmental organizations have rallied around the need to prepare all high school graduates to succeed in the workplace and higher education. But NCLB is ill-suited to hold high schools accountable for achieving this goal. While elementary and middle schools must test students annually from the third to eighth grades, high schools are only required to test students once, in reading and math. Most do so in the 10th grade and focus on basic skills. NCLB, therefore, doesn't address the advanced skills that students need to thrive in the workplace and the college classroom.

One way to move NCLB closer to this goal would be to follow high school graduates into the workplace and college and see whether they succeed there. Individual high schools don't have the logistical capacity for such long-term tracking, which is one reason NCLB relies on tests. But a growing number of state governments, like Oklahoma's, have the resources to calculate a range of post-graduation outcomes for students from individual high schools.

The Oklahoma State Regents gathers volumes of information about every student who enrolls in a state public university, as well as students in some private institutions. In addition to demographic data like students' age and race/ethnicity, the Regents tracks college admission test scores, course enrollments, college grades, and more.² By matching this data to individual student records gathered by the Oklahoma Department of Education (which oversees K–12 schools), the Regents can calculate post-graduation performance data for every high school in the state.

The following charts compare high school English proficiency rates with several college indicators. Taken collectively, these charts suggest that NCLB doesn't paint a complete picture of high school success. Many students complete their NCLB testing in the 10th grade, with at least half of their high school career still in front of them. Some high schools are much more successful than others in moving students forward from this point, helping them prepare for college, go to college, and succeed in college-level work. This information should be considered in determining whether high schools are "in need of improvement" under NCLB.

Chart One compares scores on the state's NCLB English test the 2005–06 school year to the average English score on the ACT, a college-admission examination, from 2001 to 2006. Each blue diamond on the chart represents an individual high school. The line in this and subsequent charts shows the overall trend in the relationship between the two variables, in this case the state test and the ACT.

High schools with higher scores on the NCLB English test tend to have higher scores on the English ACT. The relationship is fairly consistent, with not a lot of variation above and below the trend line. This is unsurprising given that both measures are derived from standardized tests that students take during their sophomore and junior years in high school.

Chart Two shows NCLB English test scores compared to the percent of students who went on to enroll in an Oklahoma public university. As with ACT scores, the relationship is positive—high schools with higher-scoring students send more students to college. But the correlation is not as consistent. In particular, it's noteworthy that some high schools with below-average test scores send a relatively large number of students to college.


Chart Three compares NCLB English test scores to the percent of students who, once enrolled at an Oklahoma higher education institution, were forced to take at least one "remedial"—that is, high school level—course.


Once again, the overall relationship is as expected—higher NCLB English scores tend to be correlated with a lower likelihood of needing college remediation. But the variability among individual high schools is even wider than for college-going rates. Among high schools where roughly 60 percent of students scored as "satisfactory" or "advanced" on the NCLB English test, remediation rates varied from less than 10 percent to over 80 percent.

This pattern of wide variability continues in Chart Four, which compares NCLB English test scores to the percent of students who enrolled in an Oklahoma public university and earned at least a 3.0 grade point average as freshmen.

The relationship is positive—higher test scores are associated with higher freshmen grade point averages—but not particularly strong, which can be seen by the relatively flat slope of the trend line. Among the high schools where graduates appear to be least prepared for college—those where less than 30 percent of freshmen met the 3.0 threshold—NCLB English pass rates ranged from less than 20 percent to more than 90 percent.

It's important to note that each of the high school performance measures calculated by the Oklahoma Regents has limitations. While all students are tested for NCLB purposes, not all students take the ACT. College-going rates don't include students who attend some private colleges in Oklahoma, or any out-of-state colleges. Universities often establish different standards for assigning students to remedial courses. And the academic performance of college students is a function of the quality of education provided at both the high school and college levels.

These limitations should be taken into account when holding high schools accountable for results. That said, no single measure is perfect—including the standardized tests currently in use. But, this does underscore the need to use multiple measures when evaluating schools, so the flaws of a single measure don't skew the overall picture of school performance.

As more states follow Oklahoma's lead, there will be more measures to choose from. Some states have already connected their education data systems to databases containing information about employment, welfare, and public safety. In addition to tracking students into college, the Florida Department of Education³ publishes, for each high school, statistics such as:

• The percent of high school graduates working in Florida on a full-time and/or part-time basis the autumn after graduation
• Average hourly earnings of employed graduates, as well as the percent of employed graduates with earnings meeting certain hourly wage thresholds
• The percent of graduates working for the federal government
• The percent of graduates enlisted in the military
• The percent of graduates receiving food stamps and/or aid from the Temporary Assistance to Needy Families program
• The percent of graduates who are incarcerated or under community supervision

Florida can also track the field in which graduates are working, creating new ways to evaluate the success of vocational education and other programs focused on the work force.

The number of states with the capacity to follow the lead of Oklahoma and Florida is sure to grow. A recent report from the Lumina Foundation found that 40 states now have data systems that can track individual students enrolled in public universities.⁴ While only 11 of those states have linked their higher education and K–12 data systems, an additional 16 states are planning to do so in the near future. Twenty-three states, meanwhile, have linked their education data systems to employment records.

The time will soon come when the majority of states can do more than wonder whether their high schools are preparing all students to succeed in work, higher education, and life. They'll be able to find out for sure.

Endnotes
¹ Sipuel v. Board of Regents of the University of Oklahoma.
² These data are publicly available in pdf and Excel spreadsheet formats online at http://www.okhighered.org/studies-reports/indicators/mean-act/
³ http://www.firn.edu/doe/fetpip/high.htm 
⁴ Peter Ewell and Marianne Boeke, Critical Connections: Linking States' Unit Record Systems to Track Student Progress, 2007.