Schooling for students in kindergarten through grade 12 is the heart of what most people think of as “education.” Public K-12 education is a large and complex system, and it is the focus of many expectations, from producing responsible and productive citizens to boosting the nation’s standing in science and technology and its position with respect to its economic competitors.14 The presenters represented an array of research experience: all the suggested indicators are listed in Table 3-1.
One way to think about the providers of public K-12 education, Mark Dynarski observed, is as an industry, and doing so highlights several aspects that are important for thinking about which indicators of quality and effectiveness would be most valuable. Public education is, by and large, provided by governmental structures that are highly decentralized, with each state having separate authority to supply education, and more than 15,000 school districts operating within those states. Thus, the supply of education, in an economic sense, takes the form of face-to-face instruction in classrooms, supported by physical capital, as economists term it, in the form of land, buildings, and equipment. Schools also are responsible for providing transportation and meals to students. Public education is supported by federal, state, and local property taxes. The federal share is less than ten percent of the cost, though the federal government exerts significant authority over education through legislation and regulation.
As with other publicly provided goods, the “output” and efficiency of public education are challenging to define and measure. National assessments of achievement, in place for decades, and, more recently, state assessments developed in compliance with the No Child Left Behind Act, provide proxy measures. But, in Dynarski’s view, a set of indicators should provide a broader picture of the state of the complex system that educates nearly 50 million students in more than 98,000 schools and spends more than $500 billion per year.15
Other presenters had somewhat different perspectives—the group offered indicators for both schools and teachers.
14The indicators system would include data on private schools and their students but it is public education that is the primary focus of policy makers.
TABLE 3-1 Indicators Suggested for K-12 Education
|CHARACTERISTICS OF INSTITUTIONS, SERVICE PROVIDERS, AND RESOURCES|
• Surveys of safety and orderliness of the school climate
• School culture related to college and career aspirations, including :
o percentage of students who go on to two- and four-year colleges and full-time employment
o surveys of students’ expectations and their schools’ approach to preparation for college and career
• Collaborative school community focused on student learning, using surveys of teachers, parents, and students
• Surveys of parent satisfaction
• High school completion rates
• Grade retention rates through 8th grade
• Teacher-student ratio
• Proportion of teachers whose evaluations distinguish them from a basic standard, using measures of their contributions to student achievement and their professional practice
• Teachers with mastery-level and current knowledge of content they are teaching
• Teachers with mastery-level and contemporary knowledge of child and adolescent development
• Teacher-student interactions that demonstrate high levels and qualities of involvement, stimulation, and expansion of thinking and cognition, and sensitivity to students’ perspectives, individual experiences, and backgrounds
• Teacher-student interactions that foster relationships with and among students
• Teachers providing challenging opportunities to learn in the classroom
• School attendance by age
• College readiness levels by age and grade
• Voter registration rate of 18- to 21-year-olds
• Command of core content, using NAEP scores
• K-12 education spending as a share of gross domestic product (GDP)
• K-12 spending per student
• Percentage of K-12 education funding spent on research and development
• Opportunity to learn
Safety and Orderliness of the School Climate
If students do not feel physically and emotionally safe at school, Elaine Allensworth argued, they will be reluctant to go to school and will have a hard time focusing on learning when they are there. Schools that are safe have higher attendance and better teacher retention; schools with safety problems tend to be those that struggle to improve learning, and have low attendance and graduation rates, high teacher mobility, and poor emotional outcomes for students. Poor discipline strategies in unsafe schools disproportionately affect minority and low-income students, she added.
Surveys, in her view, are the best way to measure the school climate. Other measures that may seem to be more objective are often biased by variation in discipline and record-keeping practices. Although she sees many possible positive outcomes if school climate is treated as an important measure, she cautioned that it is also possible that schools worried about accountability in this area might use counter-productive strategies, such as increasing suspension rates or arrests at school or increasing teacher-led instruction to increase orderliness in the classroom.
School Culture Related to College and Career Aspirations
A primary goal for schools is to prepare students for college and careers, and there could be several components to an indicator of how well they are doing this, in Allensworth’s view. An indicator of the percentage of students who go on to two- and four-year colleges and full-time employment would provide a basic indication of school effectiveness. For more nuanced information, she suggested an indicator of students’ expectations and their schools’ approach to preparation for college and career that is based on surveys. A benefit of focusing on this indicator could be to make schools take on a greater sense of responsibility for explicitly preparing their students for the future. Doing so might also encourage collaboration across school levels, as systems consider the elements that foster preparedness even in elementary and middle school.
Collaborative School Community Focused on Student Learning
The degree to which school leaders, teachers, and parents collaborate has consistently been found to be among the strongest predictors of school improvement, school safety, and student learning gains (see, e.g., Literacy Collaborative, 2009). This important aspect of schools can be defined in different ways, Allensworth noted, but whether the focus is a sense of professional community among teachers; parent-teacher collaboration; inclusive leadership; or trust between and among teachers, parents, and principals, the research consistently shows positive effects on important outcomes. She suggested an indicator of the nature of school communities that is measured through surveys of teachers, parents, and student. Focusing on this measure may help schools improve coordination related to instruction, school climate, and staff stability.
Dynarski suggested an indicator of parents’ satisfaction. The extent to which parents are satisfied with the education their children receive is an important indicator, in Dynarski’s view, because it is they who are responsible for the demand for it. Their satisfaction could be viewed as a judgment about whether the funds are well spent. Their dissatisfaction could suggest either that the costs are out of proportion with the results, that their children are expressing unhappiness with school for reasons that parents judge to be valid, or that parents have incomplete or erroneous information about their schools.
The National Center for Education Statistics has measured parent satisfaction at regular intervals using the “National Household Education Survey.” It asks parents about their level of satisfaction with schools, teachers, academic standards, and order and discipline in schools, and recently a question about staff interaction with parents was added. The survey results show that, since 1993, levels of parent satisfaction have been stable—a result Dynarski finds surprising considering how tumultuous the period covered has been.
High School Completion Rates
High school completion rates “can’t be overlooked,” noted Rob Warren. There are fewer and fewer jobs for people without high school diplomas, he noted, and failing to complete high school is a robust predictor of many economic and social difficulties, including poverty, delinquency and crime, and civic disengagement. Much recent attention has focused on the accuracy of different ways of measuring school completion and dropout rates (see National Research Council, 2011a). Measures of a few states’ 4-and 6-year completion rates are available in publicly reported data (longitudinal student tracking systems), and it is possible to accurately approximate these data for past years, he observed, using counts of enrollment by grade and public school graduates (Greene, 2002; Heckman and LaFontaine, 2008).
Retention Rates Through 8th Grade
An equally important but less well-documented indicator, suggested Warren, is grade retention. Data on school completion have been well publicized, but few people could estimate the average grade retention rate for the nation, he noted. Grade retention is one of the leading predictors of high school dropout rates and is also a valuable indicator of students’ progress through school, in terms of both cognitive and noncognitive outcomes. Interpretations of the significance of grade retention rates vary, Warren observed. To some, low rates are evidence that there is too much “social promotion,” or a push to move students to the next grade even if they are not academically ready, while to others low retention rates mean that most children are succeeding. Conversely, high retention rates may look like evidence that schools are unfairly punitive or that they are holding students to high standards, depending on the observer.
To demonstrate the possibilities for exploring retention data, Warren compared data for twelve states that do report their grade retention rates with three sources of
publicly available data: census-based estimates of the number of people in each jurisdiction who are of a certain age (e.g., six year-olds in Kansas in a particular year); data from the Private School Survey on the numbers of K-8 students who attend private schools; and public school enrollment counts from the Common Core of Data (CCD). Using the three data sources, he constructed estimates of grade retention rates for the 12 states and found them to be highly correlated with the rates the state had calculated. For 2004-2005, he observed a range for the 1st grade among the 12 states from less than 4 percent for Wisconsin to between 7 and 8 percent for Delaware, for an estimated rate for the United States of 4.2 percent (Louisiana was an outlier at nearly 12 percent because of Hurricane Katrina).
This method works for grades 1 through 8, he noted, but the high school grades present additional challenges.
For Dynarksi, the “fundamental technology” inside schools is the relationship between teachers and students inside the classroom. One possible indicator would be pupil-teacher ratio, a measure of the contact between teachers and students, and it is reported by NCES.
The ratio has declined steadily since 1960, when it was one teacher for every 25 students, to the current level of one teacher for every 15 students. However, Dynarksi cautioned that these figures include specialist teachers, such as reading and special education faculty who play a range of roles, so the class size likely has not declined as significantly as the ratio might suggest. For example, the average size for elementary classrooms was 21.2 students in 1999 and is currently 20.3 students. Dynarski noted that these downward trends have not yielded increases in NAEP scores in that (admittedly short) period of time. One might ask whether this finding shows that the school system has become decreasingly productive in terms of how efficiently resources are used, he added, but in considering such a question it would be critical to consider many other factors that influence the productivity of schooling—changing characteristics of student populations and changing educational goals and emphases, for example.
Brian Gill suggested an indicator of the proportion of teachers whose evaluations distinguish them from a basic standard, using measures of their contributions to student achievement and their professional practice.
“Education researchers have come to a conclusion that has long been apparent to parents and teachers,” he noted: “Teachers matter, both in the short and the long term” (see, e.g. Chetty, Friedman, and Rockoff, 2011; Hanushek and Rivkin, 2010; Kane and Staiger, 2012). This body of research has had particular policy importance as the U.S. Department of Education has focused on using its leverage to encourage states to better distinguish between high- and low-performing teachers, and to base high-stakes consequences on their evaluations. This poses both opportunities and challenges, in
Gill’s view, because the policy is actually “ahead of the research.” It is clear that teachers matter, he added, but there is a great deal more to learn about how they matter. Which characteristics and practices are most important to teacher effectiveness, and how they can be reliably measured are questions that have not been fully answered. Current evaluation systems, he added, are typically showing that more than 95 percent of teachers are satisfactory, so there is a clear need for better ways to distinguish among them.
Most attention has centered recently on “value-added” models that measure teachers’ contributions to changes in students’ standardized test results in reading and mathematics.16 In Gill’s opinion, value-added measures show promise as one component of a broader system of teacher evaluation, despite methodological challenges. The field will also need deeper and broader measures of performance that can be used with teachers of other subjects and grades not tested, and that can capture aspects of performance not measurable using standardized assessments, he added. Other measures (e.g. the Classroom Assessment Scoring System (CLASS)17 and the Framework for Teaching18) have shown promise as ways to measure teacher effectiveness (Kane and Staiger, 2012).
In Gill’s view, one concern about evaluation systems that hold teachers accountable for their students’ gains on standardized tests is that they can create incentives for teachers to “game the system,” taking steps designed to boost scores rather than to improve student learning. He believes that value-added metrics should be used with caution, that they should be used, if possible, with multiple measures of student achievement, and that they should be augmented with robust measures of teachers’ professional practice.
Robert Pianta suggested two indicators of teacher quality: (1) teachers with mastery-level and current knowledge of content they are teaching and (2) teachers with mastery-level and contemporary knowledge of child and adolescent development. These two aspects of teachers (including those who interact with students in after-school programs and other settings outside the regular school day) are basic structural indicators of their effectiveness, in his view.
“Does the person teaching algebra know algebra?” is a key question, he explained. Another is whether a teacher has up-to-date knowledge of child and adolescent development, and of learning trajectories in a variety of skill domains. These would not be difficult to measure, in Pianta’s view, but currently, “we do not assess them at all.” The proxies most often used—possession of a master’s degree, course-taking, or certification status, for example—are not associated “with much of anything,” in his view. He acknowledged that a considerable amount of research has shown the importance of content knowledge, but he argued that more precise measures than teachers having majored in the subject they are teaching are needed.
16See National Research Council (2010) for more information about value added modeling.
17For a description, see http://www.brookespublishing.com/store/books/class/index.htm.
18For a description, see http://www.danielsongroup.org/article.aspx?page=frameworkforteaching.
Pianta also suggested three indicators of classroom processes: (1) Teacher-student interactions that demonstrate high levels and qualities of involvement, stimulation, and expansion of thinking and cognition, and sensitivity to students’ perspectives, individual experiences, and backgrounds; (2) Teacher-student interactions that foster relationships with and among students; and (3) the extent to which teachers provide challenging opportunities to learn in the classroom.
Looking at teachers’ characteristics is not enough, in Pianta’s view. It is important also to look directly at the character of classroom instruction in three ways that he believes are reasonable and practicable. A variety of research has pointed to teacher engagement as a key factor in student learning (Gates Foundation, 2010). When teachers interact in an engaging way, he explained, they create cognitive demand through feedback and discourse, and they also help students feel a sense of belonging and appreciate the relevance of the content that is being taught. These are classrooms, he added, “in which children are trending higher” even on standardized assessments. These teacher behaviors, he suggested, also show an association with lower dropout rates and rates of social and behavioral problems (Gates Foundation, 2012).
These behaviors can be quantified and assessed, he added, using some of the observation techniques discussed in the context of early childhood education. He argued that number of large-scale studies have demonstrated that both observations and student surveys can capture the nature of classroom interactions quite well. Pianta cautioned that the rigor and reliability of such measures depend on careful design, and also that if such measures are used, it will be important to watch for unintended consequences. For example, if an indicator is adopted that “simply counts the number of open-ended questions a teacher asks, you are going to get a lot of open-ended questions,” without, perhaps, much insight into whether those questions are embedded in an effective and engaging interaction.
School Attendance by Age
A basic indicator of opportunity to learn, in Elaine Allensworth’s view, is student attendance. “If kids are not in school, they are not learning,” she observed, and proposed an indicator of school attendance by age. She emphasized that a large body of research demonstrates the importance of attendance. There is wide variation in attendance rates, even across schools serving very similar populations. Attendance is often viewed as a trivial or low-level predictor, she added, but it is highly predictive of eventual educational attainment—it influences learning, grades, and graduation rates. Attendance changes as students age, with problems typically beginning in the middle grades and rates “bottoming out” in high school. She also noted that it is relatively easy and not expensive for schools to improve attendance, and they can achieve significant benefits quickly.
Data on average daily attendance are already collected by all schools (in terms of the percentage of enrolled days that are days attended), but attendance also can be
calculated to include nonenrolled students and thus be combined with information about students who drop out. Another possibility, Allensworth noted, would be to measure chronic absence, though doing so without also measuring daily attendance for all students would mean focusing on a very small subgroup of students and missing the opportunity to flag problems before they become chronic.
College Readiness Levels by Age and Grade
Allensworth proposed an indicator of college readiness. Although it may seem difficult to measure, colleges have long used systems based on high school grades and test scores, and grades have been found highly predictive of both college performance and later earnings. Tests scores are less predictive, she added, particularly when other factors are taken into account. Many people view high school grades as too subjective to be used as a reliable indicator, she added, but they do reflect students’ engagement, motivation, and other noncognitive factors associated with academic success. Since standardized tests provide measures of certain skills, the two together provide a better picture of student progress than either does alone, in her view.
There are a variety of tests already in use, as well as databases containing information about students’ coursework and high school grades, and linking those to test scores would provide the potential for developing a composite measure of readiness, Allensworth believes, and it should be possible to adapt and improve such measures to make it possible to assess students progress toward readiness in earlier grades. Though data are available at the national, state, and local levels, she added, they are not always comparable.
There is also a significant caution related to this measure, she added, because high-stakes accountability incentives based on test scores and grades can lead to poor educational practices. Emphasis on testing can lead to narrowing of the curriculum and cause teachers to devote excessive time to test-taking skills. Similarly, using grades for accountability purposes can lead to grade inflation, and, particularly in schools with very poor student engagement, grading practices may reflect effort at the expense of actual performance. These problems, she added, are most likely in schools serving the most disadvantaged students.
Voter Registration Rate of 18- to 21-Year-Olds
One of the original goals for public education, noted Gill, was to develop the educated citizenry that is essential to the functioning of democracy, and opinion polls indicate that people still believe that is an important goal. He proposed an indicator of the voter registration rate of 18- to 21-year-olds. This issue is even more important in the 21st century than it was in earlier in the nation’s history, he added, because government plays a larger role in people’s lives than in the past, and the “stream of information bombarding us” puts a high premium on the ability to distinguish fact from fiction and from opinion. New approaches to school funding and governance (e.g., vouchers and charter schools) have challenged the expectation that education will be delivered in a “common school” under the direct authority of democratically elected officials, he added. For all these reasons it is important to focus public attention on
preparation for citizenship and measure it not only at the national level but also for each sector (traditional public schools, charters, private schools, and home schooling), and for individual schools.
The rate at which young people register to vote and actually vote is one possible indicator of how well schools are succeeding in this mission, Gill explained, and the rate at which young people (including those too young to vote) participate in community service is another. The National Assessment of Educational Progress (NAEP) includes an assessment in civics that addresses civic knowledge, attitudes, and skills, though, Gill noted, that assessment has only been administered three times in the last 15 years.
Command of Core Content
For Rob Warren the fundamental question is whether kids are learning. He proposed an indicator of command of core content. He turned to NAEP scores as the best starting point for examining what students know about core subjects in multiple grades, in a way that can be disaggregated by state, by social-economic group, and across time. NAEP has some drawbacks, he acknowledged. The cut scores are sometimes seen as too high or as arbitrarily set,19 for example, but students who lack basic proficiency in the subjects tested by NAEP are less likely to complete postsecondary schooling and are at a disadvantage in the labor market as a result. NAEP scores are already widely publicized, and the American public already views them as indicators of the effectiveness of public schools.
Education Spending as a Share of Gross Domestic Product
The proportion of a country’s expenditures that is devoted to education is a possible indicator of the importance the country attaches to it, and is used as a basic point of international comparison, Dynarski explained, and he proposed an indicator of education spending as a share of gross domestic product. The United States does not rank very high in that comparison, but that is partly because it is so wealthy, he noted: it could be spending a great deal on education without the total constituting a large share of gross domestic product (GDP). This ratio fluctuates, because the GDP fluctuates, but the trend since 1984 has been a steady increase in education spending, in real dollars.
This ratio can be calculated using GDP data and separately collected public expenditure data.
Spending per Student
It is also important, Dynarski added, to consider that if the number of students has increased significantly, expenditures per-pupil may actually decline even while the
19The cut scores for NAEP are determined through a judgment-based process in which a panel of educators and public members reach consensus about descriptions of the sorts of skill and knowledge students should demonstrate to meet performance levels. For a description of the standarrds-setting methods used for NAEP, see http://nces.ed.gov/nationsreportcard/set-achievement-lvls.asp.
proportion of GDP increases, and he suggested an indicator of spending per student. He noted that data collected by NCES shows that per-pupil spending has trended upwards since the early 1960s, sharply up until the 1990s and more gradually since then.
Spending on Research and Development
The expectation in many fields—particularly medicine—is that a great deal of research will be conducted and practice will follow that research as it develops, noted Brian Gill. The same has not been true in education, and the result is that “the technology in education has been remarkably unchanged for a century or more.” In response to Dynarkski’s observation that schools may not have become more productive, Gill suggested that stagnant productivity is partly a result of a chronic underinvestment in research and development, and he proposed an indicator of the percentage of education funding that is spent on research and development.
Moreover, the demands of the job market and of effective citizenship have steadily grown, making the stakes of schooling higher than they have ever been. In his view, dramatic improvements in educational productivity are needed. New technologies that have been described as revolutionary, he suggested, have not changed the core of classroom instruction. Gill believes that there is reason for cautious optimism about the potential in educational technologies that are now being developed, but that dramatic increases in productivity will require both large additional investments in research and development and a willingness on the part of school systems to allow new technologies to disrupt longstanding institutional practices
This effort should include investment in research into ways to better measure student achievement and evaluate the effects of interventions, in Gill’s view. “One of the frustrations” of the new research on teachers’ value added, he added, “is that it has thus far provided little information on what highly effective teachers are actually doing that makes them highly effective.” Research into these and other questions is needed not only for its own sake but to support immediate needs.
Opportunity to Learn
Although it may be hard to measure, opportunity to learn cannot be ignored, argued Rob Warren, and he proposed it as an indicator. He likes the definition used by Grodsky and colleagues, “the resources available to students, most often in the classroom setting, that facilitate their acquisition of knowledge or skills” (2008, p. 388), and he stressed that it is important to consider this aspect of schooling separately from other indicators. Measures of achievement (e.g., NAEP scores) and measures of school completion both conflate two separate factors: (1) student abilities and effort and (2) the structures that influence students’ opportunities.
But, Warren noted, there is considerable debate about how opportunity to learn might be measured. Possibilities for which data are available include class size; pupil-teacher ratio; number of days and hours of instruction per year; teacher quality, perhaps based on value-added measures; or measures of facilities or such offerings as Advanced Placement and International Baccalaureate Programs. These are all proxies rather than direct measures, Warren observed. In his view, further work is needed to determine
which measure or combination of measures best captures this essential aspect of schooling, and he finds it “problematic that such a fundamental concept has received so little measurement attention.”
Though the panelists were all given the same charge, they approached it in somewhat different ways. Each considered the indicators they chose as a set, intended to cover at least one aspect of the status of education. Some focused only on one of the three aspects defined in the committee’s framework; see Appendix C for a list of the indicators arranged by presenter. Thus, individually, each indicator tells a story, Mark Dynarski noted, but taken as a group they should reveal more. For example, if the teacher-student ratio is changing over time, that suggests students may be getting more or less time with teachers. But if that indicator is examined in light of other indicators, of teacher quality or of other factors that may influence the ratio, for example, a more nuanced picture can emerge.
Though the workshop was not intended as a vehicle for making a final selection or recommendations of indicators, discussant Henry Braun encouraged the panelists to use the initial suggestions each had made to consider the characteristics the system as a whole should have. The indicators put forward, he suggested, could easily have been suggested 10 or even 20 years earlier, and he wondered whether further thinking would be needed to ensure that the indicators ultimately chosen will support important contemporary goals for education. The discussion addressed this question from several angles.
College Readiness? Broader Goals?
Several discussants also wondered whether the indicators suggested reflected a sufficiently ambitious vision of what public education can accomplish. One noted that the indicators chosen for a similar project covering European countries (Hoskins, Cartwright, and Schoof, 2010) are based in fundamental social goals: economic security, social cohesion, and sustainability. The United States might have somewhat different goals, the participant added, such as promoting participatory democracy or a balance between individual freedom and responsibility, but the set of indicators chosen could be conceived as measures of how well the system is meeting such goals.
Another participant suggested a different sort of educational purpose that also could provide a conceptual underpinning for the indicators chosen. Beyond the specific skills and knowledge education should impart, he explained, is the idea that education should help children develop images of “possible selves” as they gain understanding of what their options are. Prior efforts to “rethink” high school, he added, such as the Youth Act of 198020 or the School-to-Work Opportunities Act of 1994,21 have not yielded the desired changes. He hopes that current thinking focuses less on the distinctions between college and career preparation and more on helping students prepare to grow and change.
Several presenters pointed out that the skills needed to succeed in the 21st century are not necessarily different from those needed in the past. While adaptation—such as taking advantage of advances in the cognitive and behavioral sciences—is important, it may also be important to recognize that even if ideas have been under discussion for a long time, the system may not yet have succeeded in implementing them. In particular, Pianta noted, possibilities for measuring aspects of education that may have been recognized as important but were not easily captured using large-scale standardized tests have opened up considerably.
It is important to recognize, though, Allensworth noted, that “we have ratcheted up expectations for students” significantly in the last 10 to 20 years. “Almost all schools now say they want their students to leave school college ready,” she explained, and this is a significant change from the way things were a generation ago. Tracking student grades is useful, she added, precisely because they capture skills that cannot be measured on tests but are important for both postsecondary schooling and work, such as the ability to get things done, to work in groups, and to solve problems. “There’s not a whole lot of evidence out there, she added, “that what you need to be career ready is so different from what you need to be college ready.”
Nevertheless, Dynarski observed, “one could read the data and say the K-12 system has just moved too slowly [so it is putting students into] the pipeline who aren’t ready for colleges that have adapted more quickly. The empirical evidence is pointing to the shortcomings not of colleges but of the K-12 system.”
What Might Be Missing?
Discussion highlighted a few issues that were not raised in the suggested indicators. One participant noted the relative lack of emphasis on contextual factors, particularly in comparison with the suggested preschool indicators. “Presumably the family and the community are just as important for school-age students as for kids from zero to age 5,” this person remarked. Another noted that “education is actually coproduced with parents,” but that parental inputs were not addressed by any of the indicators. The important role parents play is explicitly addressed in the early childhood context, this person observed, but “over time we assign more responsibility to the education provided by schools.” It would be possible, this person suggested, to develop a composite indicator using, for example, parents’ reading to children, helping with homework, or helping children prepare for college, to gauge involvement.
It is difficult to address equity issues with national indicators, Mark Dynarski noted, because national statistics may disguise significant variation. For example, steady growth in per pupil spending, on average, may result because affluent school districts are increasing their spending a considerable amount while poorer districts are not. It might be possible to construct an indicator of the variance in the states’ spending, perhaps as a share of their own budgets. But even this indicator would not capture other kinds of inequity, such as the distribution of up-to-date buildings, experienced teachers who are teaching the subject in which they were prepared, and other factors. Participants noted that measuring spending at the state level, as well as the national level, would add important information about equity and other issues.