Data & Accountability

From test scores to teacher salaries, from graduation rates to grade-point averages, the education world is full of data. The federal No Child Left Behind law, signed in 2002, created an unprecedented demand for detailed information about students and schools. No longer are public schools judged simply by average test scores for all students. The law requires states, school districts and campuses to break out (“disaggregate,” in education-speak) test scores by race, gender, English proficiency, socioeconomic status and more. Students in every group must meet the same academic standards.At the same time, people want to know which individual teachers have the most success working with various groups of students. 

The demand for data is growing in higher education, too. Parents, policymakers and taxpayers want to know which college freshmen need remedial education. They want to know how many students graduate on time, and whether graduates find good jobs. In both K-12 and higher education, some political leaders and others want to track academic performance with spending to see which schools and colleges seem to provide the best bang for the academic buck. This Topics section examines what this proliferation of data means for education reporters and offers links to key data resources.

This national push for accountability in education means that schools and colleges must first collect lots of useful, timely and precise data. Then, they must analyze and use that data – that is, make good decisions based on evidence — all with an eye toward improving student success. 

Here’s how Jeffrey Wayman, assistant professor of education at the University of Texas at Austin, describes the goal on his Data Use website: “If offered in a useful form, such data can help teachers, principals and other educational personnel learn more about their students, improve their teaching craft, and ultimately impact a variety of educational outcomes.”

When the Data Quality Campaign, a national coalition advocating for better data, formed in 2005, no states had extensive longitudinal databases for education. The campaign created “10 essential elements” it deemed the state data systems should have. By 2011, the group reported that 49 states (all but Montana) have eight or more of the elements. Now the group is focusing on what states should do with all of that data.

Aimee Guidera, the campaign’s executive director, says: “The need is urgent: state policymakers need to allocate scarce resources based on what works to help students, and they cannot do that without data.”

Sources of data

Federal, state and local agencies now keep a wealth of education data. At the federal level, one of the best sources is the National Center for Education Statistics, part of the U.S. Department of Education. The center publishes annual reports (namely the Condition of Education and the Digest of Education Statistics) with state and national education trends on student enrollment and demographics, school staffing, education funding, graduation and dropout rates, and much more. The center also keeps searchable and downloadable databases, such as the Common Core of Data (for K-12 education) and the Integrated Postsecondary Education Data System (for higher ed). While the information at NCES is comprehensive, it’s also often dated, lagging two years behind.

State education agencies also offer a trove of data on student achievement (test scores from state exams, SATs, ACTs, Advanced Placement), graduation and dropout rates, student demographics, school funding, teacher salaries and staffing levels, and more. Some states put many databases online so they’re easily downloaded, while other states may not readily post such information — and that means reporters will need to specifically request it.

School districts also keep data on student test scores, employee salaries, budgets and the like. Again, the type and availability of information vary by district.

Keep in mind that federal privacy laws, namely the Family Educational Rights & Privacy Act (FERPA), restricts disclosure of individual student education records. In practical terms, that means a journalist can’t ask a school district or state agency for a list of student names with their race, gender and test scores. But a journalist can ask for student-level data with all identifying information (names, social security numbers, etc.) removed. Depending on the state, education agencies may even redact information in particular for small subgroups of students (for instance, test scores for the only four white, female students in a school, under the theory someone could still deduce the individual students’ scores).

Controversies, challenges and caveats

As schools and education agencies are able to collect powerful data, and they’re able to analyze it with more sophisticated tools, controversies have risen.  Some school systems around the country rate teachers based on how much academic growth students made while in their classrooms. These “value-added” models look at which teachers are the most (and least) effective with their students by these measures. Some education leaders want to tie teacher pay or raises to their students’ academic growth, instead of compensating them solely on their education credentials and years of experience. Skeptics say the value-added models have too much statistical error to rate individual teachers accurately.

There can be other problems with education data. If students or their teachers cheat on state exams, the test scores (and everything they’re based on, such as state academic ratings) become meaningless. If schools don’t accurately report the reasons students leave school, then dropout and graduation rates aren’t accurate. Some colleges have been accused of reporting false or exaggerated data to U.S. News & World Report so they do better in the magazine’s annual college rankings. Journalists, and anyone else who uses data, need to ask questions about data when they see red flags.

With so much emphasis on data, some experts advise that journalists, policy makers and others should also examine things that can’t be easily measured — classroom observations of students and teachers, portfolios of student work, parent involvement, and school culture. And we don’t have sufficient data for all students, such as those who are exempt from state exams, or part-time college students who don’t count in the official graduation rates.

In the classroom, meanwhile, educators face the challenge of taking all of the data on students and using it well. The research isn’t yet extensive on the use of data to improve instruction or make decisions about teachers or students, but the research that is available shows that districts need to offer extensive training to principals and teachers to make the data useful. If educators don’t receive the training, they won’t use the data.

“The greatest perceived area of need among districts is for models of how to connect student data to instructional practice,” a U.S. Department of Education study found in 2010. “Districts want examples of how to identify which practices work best for which students and how to adapt instructional strategies to meet the needs of individual students.”