The No Child Left Behind Act, the new school accountability measure that became law in 2001, requires that all states develop and administer standards-based tests to students in grades 3 through 8 and report results by student subpopulation (race/ethnicity, gender, socioeconomic status and disability) for districts and campuses. But some ways of reporting student data are more revealing than others.
What Is Value-Added Assessment?
The idea behind value-added measurement is that schools should be judged based on what they do, not on the abilities of the students who happen to be in their classrooms. Under school performance measures in Texas, for example, school ratings are based on how many students have passed the state exam. As a result, some schools located in neighborhoods with high-performing students automatically draw praise, while others in neighborhoods with low-performing children unfairly draw scorn. The relevant question is: What does each school do to improve the scores of its students from one year to the next?
The difference between evaluating schools based on raw test scores and evaluating them based on the schools' value-added contributions to those scores is important. For example, a comparison of the top 100 schools in Texas based on raw scores and the top 100 based on value-added finds only 12 schools on both lists.
How Is Value-Added Measured?
Value-added analysis is conducted by removing factors that affect performance such as a student's background and family characteristics through state-of-the-art statistical regression so researchers can isolate and measure the school's effect on a student's performance from one year to the next. Our Report Card on Texas Schools examines student-level data and performance on the math and reading sections of the Texas Assessment of Academic Skills for the 1999-2000, 2000-2001 and 2001-2002 school years.
Value-Added Differences among Schools
The following results are based on an evaluation of sixth-grade classes in Dallas. Similar results were obtained statewide. Consider two schools, Harrell Budd Elementary School in the Dallas Independent School District and Kennedy-Curry Elementary School in Wilmer-Hutchins Independent School District:
- The students of Harrell Budd are predominantly low income (92 percent) and predominantly minority (99 percent).
- The students of Kennedy-Curry also are mostly low income (64 percent) and predominantly minority (95 percent).
These schools are similar and are only six miles apart. However, Budd is one of the best elementary schools for value-added in Dallas, while Kennedy-Curry is at the bottom. Moreover, these two schools illustrate that school performance cannot be explained by school resources alone. Kennedy-Curry spends $500 more per pupil ($4,620 vs. $4,157) and has smaller classes (22 versus 25 pupils).
Value-Added Differences within Schools
Clearly, differences exist among schools with similar students. Differences also exist within schools with respect to their ability to educate different types of students. For example, in the Dallas area:
- Nancy Moseley Elementary (DISD) ranks eighth best for teaching high achievers (students who scored in the top quarter the previous year), but is No. 278 out of 361 overall.
- Everett L. Degolyer Elementary (DISD) and Arapaho Elementary (Richardson ISD) rank Nos. 4 and 6, respectively, for teaching low achievers (students who scored in the bottom quarter in the previous year), but overall their value-added rankings are a mediocre No. 97 and No. 71.
- Edison Medrano Elementary (DISD) ranks No. 2 for high achievers but No. 301 for low achievers.
- By contrast, W. W. Bushman Elementary and Bowie Elementary (both in DISD) are ranked in the bottom 10 schools for teaching high achievers, yet both rank in the top quarter of all schools for teaching low achievers.
Thus, parents of children with certain characteristics may be quite satisfied with their progress at a particular school, while parents of other children in the same school may be dissatisfied.
Value-Added Racial and Ethnic Differences
Schools also exhibit differences in their ability to teach different racial and ethnic groups. Many would like to believe that Dallas schools do an equally good job regardless of the race or ethnic background of the students. However, this is not the case. For example, as the figure shows, 
some schools add more value to black students' performance than to the scores of students overall, while others add less to minority students' education. We do not know why some schools do better than others with different groups. But the differences exist, and they are significant:
- John W. Carpenter Elementary (DISD) ranks No. 1 for teaching Hispanic children but No. 89 for teaching black children.
- Commerce Middle School ranks No. 2 for teaching Hispanics but No. 117 for teaching blacks.
- Tom W. Field Elementary (DISD) ranks second best for teaching black children but No. 49 for Hispanics.
Value-Added Rankings for Charter Schools.
Charter schools are special public schools that are free from the bureaucracy that hampers innovation and flexibility in traditional public schools. Charter schools typically attract lower-performing children so they do not compare well with other public schools on the basis of passing rates, but we might expect them to do well on a value-added basis. Interestingly, charter schools are found throughout our index; for instance, two charter schools are among the 10 best schools in the Dallas area, and two charter schools are among the 10 worst. Generalities cannot be drawn about charter school performance. Like traditional public schools, each campus has its own strengths and weaknesses.
But charter schools give parents vital educational options where previously they may have had none. The presence of charter schools – and the threat that traditional public schools in the same neighborhood may lose students to them – encourages traditional public schools to improve.
The NCPA's value-added Report Card on Texas Schools is available online.
Policy Implications for School Choice
All states should incorporate value-added assessments in their school evaluations. Measuring schools based on passing rates does not reveal whether they are improving their students' scores. Thus, schools currently identified as "successful" may not be the ones worth emulating. These schools may simply be benefiting from inheriting very able students. That is fine, but it is hardly a model lower-performing schools can copy.
Additionally, our analysis demonstrates that schools have strengths and weaknesses. Perhaps schools should not try to be all things to all people and instead should specialize in the type of students they educate well. In fact, nothing in economic theory – and certainly nothing in the performance of Dallas area schools – suggests that we will succeed in making most schools equally good in meeting the needs of all students. We found only one school in the Dallas area that ranks in the top 10 in serving all its student populations. By contrast, 41 schools are in the top 10 in serving at least one student population. Concerns about racial segregation may limit the degree of specialization allowed in the education marketplace. But within broad parameters, much can be gained by encouraging competition and parental choice.
At the very least, this information should be reported to parents, who should be able to send their children to the school that is best for their child. Parents should be allowed and encouraged to move their child when they find an opening at another campus that has superior skills in teaching students with their child's characteristics.
Matt Moore is a policy analyst with the National Center for Policy Analysis.