No. 14 September 2000
The Education Freedom Index
Jay P. Greene
Senior Fellow, The Manhattan Institute for Policy Research
Chester E. Finn, Jr.
John M. Olin Fellow, Manhattan Institute
President, Thomas B. Fordham Foundation
About the Author
Jay P. Greene is a senior fellow at the Manhattan Institute for Policy Research. He has conducted evaluations of school choice programs in Milwaukee, Cleveland, Charlotte, and San Antonio. He has also investigated the effects of school choice on civic values and integration. His publications include the chapters, “Civic Values in Public and Private Schools,” and “School Choice in Milwaukee: A Randomized Experiment,” in the book, Learning from School Choice, published by the Brookings Institution in 1998; “The Effect of Private Education on Political Participation, Social Capital, and Tolerance,” in the Fall 1999 issue of The Georgetown Public Policy Review; and “The Texas School Miracle is for Real,” in the Summer 2000 issue of City Journal. Dr. Greene is also a research associate at the Harvard Program on Education Policy and Governance and the Tomas Rivera Policy Institute. He has been a professor of government at the University of Texas at Austin and the University of Houston. He received his Ph.D. from the Government Department at Harvard University in 1995.
The author would like to thank Patrick Wolf of Georgetown University and William Howell of Stanford University for reviewing this research. Patrick Wolf was especially helpful in reading multiple drafts and making detailed comments. The author would also like to thank Richard Komer of the Institute of Justice, George Mitchell of the Institute for the Transformation of Learning, Nina Shokraii Rees of the Heritage Foundation, and Joe McTighe of the Council for American Private Education for providing useful data and encouragement. Finally, the author would like to thank Henry Olsen and Chester E. Finn, Jr. of the Manhattan Institute for their guidance throughout this project.
This path-breaking and immensely creative study of education freedom in the fifty states could not be more timely or more revealing.
Ever since Freedom House, in the darkest days of the Cold War, began its invaluable survey of political freedom around the planet, it’s been evident that such comparisons are illuminating and helpful—and that they can contribute to needed change. That celebrated, annual map of “freedom around the world” made it possible both to compare countries along a vital dimension of the human experience and to track individual nations’ evolution over time. It inspired the residents of “more free” lands to hang onto the liberty they had and invigorated those living in “less free” places to strive for more of it.
More recently, the Wall Street Journal and The Heritage Foundation have teamed up to produce an international index of economic freedom and the State Policy Network has developed a new state-by-state gauge of economic freedom. Both are hugely beneficial for monitoring how countries and states are doing in securing their citizens the right to earn a living. Both are apt to inspire change.
But what of education, today’s great domestic policy frontier? Thanks to Jay Greene and the Manhattan Institute, we can now see how the fifty states are doing at vouchsafing their residents the freedom to obtain the kind of education that they want for their children. With such knowledge comes empowerment—and the possibility of change
This is the first such report. I hope it will be sustained. Future editions can no doubt be improved as more data become available and other analytic methods are tried. But what a grand beginning, and how very interesting its results turn out to be. And, for the most part, how hopeful.
Dr. Greene’s crucial insight is that education freedom (like political and economic freedom) is primarily the result of state action. The reason people in Arizona and Minnesota enjoy far greater education freedom today than do citizens of Hawaii and West Virginia is because their states have embraced different policies and programs. But those policies are not immutable. A state could change. It could become freer with respect to education. It could also become less free. These things are determined by policymakers, by election returns, by legislative decisions, referenda and citizen action.
The education freedom differences among the 50 states are wide. People need to know this. Ours is an increasingly mobile society in which people make conscious decisions about where to live, where to locate their businesses, where to raise their children. Many factors enter into those decisions. A family might opt for Hawaii rather than Minnesota because of meteorological or demographic considerations. Hawaii, I need scarcely point out, is warmer, sunnier and ethnically more diverse than Minnesota. But that family might also want to know something about the education arrangements, in particular about the right of parents to make the education decisions they think best for their children. If it values education freedom, today it would be wiser to forget Waikiki and opt for the land of ten thousand lakes.
Education freedom is a value in its own right, but Dr. Greene takes one more crucial step. He looks to see whether such freedom correlates with education achievement. Of course, it does. After controlling for demographics, spending and other “input” variables, Greene finds that a state’s higher ranking on the Education Freedom Index is associated with stronger performance on both the National Assessment of Educational Progress (NAEP) and the Scholastic Assessment Test (SAT). The degree of education freedom in a state predicts that state’s academic outcomes, another reason why parents—and voters, taxpayers and policymakers—will want to know how their state is doing on this index.
Note that the Education Freedom Index isn’t just about vouchers. They comprise just part of one of five policy categories that figure (equally) into these calculations. Two of the others might be expected: the availability of charter school options and the ease with which a family can opt for home schooling. But the two final categories involve freedom within public schooling itself.
This study’s value for parents is obvious. But it has much to offer policymakers as well. At a time when education leads the list of voter concerns, when education issues loom larger than ever before in a national election, when myriad education policy decisions face state and local officials, and when the principal federal K-12 programs await the attention of the next Congress and President, this is very important information indeed.
Let me illustrate. Because the 106th Congress is apparently not going to finish its work on the reauthorization of the Elementary and Secondary Education Act (ESEA), this big cluster of federal programs will still be on the agenda for the 107th. One of the major policy options that has been proposed—indeed was adopted on a pilot basis a few months back by the full House of Representatives and the Senate education committee—is known as “Straight A’s”. This would give participating states wide latitude to use their federal education dollars as they see fit, so long as they boost academic achievement in return.
The underlying philosophy (as with charter schools) is that states should be free to do things differently, according to their own values and priorities—but with all of them accountable for student learning. We can see from the Education Freedom Index that states do, in fact, vary immensely. We can therefore anticipate that empowering them to use their federal dollars as they use their own dollars will yield different practices. We should be keen to know which practices yield the best results. I hope the EFI will embolden Congress and the White House to let this important experiment begin.
Not everyone will welcome this index. Tyrants, oligarchs and despots didn’t like the Freedom House map, either. It exposed their handiwork, judged its consequences and—by showing that people in other lands enjoyed far greater freedom—proved that things didn’t have to be the way they were. This became a powerful incentive for change and a source of encouragement to those bent on reform.
I don’t say that today’s U.S. public school establishment is despotic. But it’s none too fond of education freedom, it doesn’t welcome change, and it avoids criticism, both explicit and implied. The EFI will undoubtedly upset some of its crustier folks. So be it. That’s the price of giving Americans clear evidence that education freedom is possible, that people living in some places have far more of it than people in others, that state policies underlie these differences, and that with freedom comes more learning.
The Education Freedom Index
Improving education is one of the most important issues in America. People from all backgrounds and all political persuasions are offering suggestions for reform. Many of these ideas rest on the premise that increasing a parent’s freedom to choose how her child is educated will increase the likelihood that that child will be well educated. Despite the prominence this premise enjoys in today’s debate, no one has yet attempted to measure how much educational freedom parents currently possess. Nor has there been any comprehensive attempt to test the truth of this premise, that more educational freedom leads to better educational outcomes.
The extent of educational freedom varies greatly within the United States. In some states, parents have a wide selection of charter schools from which to choose, while in other states there are none. In some states, parents have access to private school options via vouchers or tax subsidies, while in other states they do not. In some states, parents can home school their children with relatively few restrictions, while in others this option is heavily regulated. In some states, school districts are small enough that parents can easily move from one to another, while in other states school districts are as large as counties or even the entire state, making choosing a different district very difficult. In some states, parents can transfer their children to other public school districts without having to move, while in other states that option is unavailable or restricted. The more options that parents have regarding the schooling of their children, the more education freedom there is.
In this report, we estimate the extent and nature of education freedom in each state, using a new Education Freedom Index (EFI). The purpose of collecting evidence on the range of education options in each state and assembling them into EFI is three-fold. First, we want to show the remarkable differences in educational freedom among the states. Second, like the Economic Freedom Index and the Political Freedom Index after which EFI is modeled, we seek to make educational freedom a concept that becomes the subject of policy debate. Third, we test whether states that increase the amount of education freedom are also likely to experience an improvement in academic achievement. We find that students in states that have higher scores on EFI also have higher scores on standardized tests, even after controlling for other demographic and policy factors.
Ranking States on the Education Freedom Index
The Education Freedom Index is composed of measures of five types of educational options:
- the availability of charter school options;
- the availability of government assisted private school options (e.g. vouchers);
- the ease with which one can home school one’s child;
- the ease with which one can choose a different public school district by relocating; and
- the ease with which one can send a child to a different public school district without changing residence.
The EFI is computed as the equally weighted average of measures of these five components.
These five forms of choice capture the range of education opportunities available to families. Public policy in each state offers up to five different options for families considering alternatives for their children. Families can consider a charter school, a government assisted private school, home schooling, moving to a different school district, or transferring their children to a different school district. Families may also consider additional options, such as paying private school tuition without government assistance, but the availability of this option is not under the control of public policy and is therefore not included in the EFI.
We are focusing only on those things that are influenced by public policy at the state level. If a state wished to increase the education freedom available, it could expand the number of charter schools allowed and ease the regulatory burdens placed on starting and operating a charter school. A state could also expand access to private schools with government assistance by offering vouchers, tax credits or deductions for private school expenses, and/or offering direct subsidies to private schools for certain expenditures. A state wishing to expand education freedom by increasing access to home-school options could ease the regulatory burden placed on home schooling. A state wishing to make it easier for families to relocate to alternative school districts could create additional school districts, such as by making all school districts smaller. And a state wishing to assist families to send their children to schools in districts other than the one in which they live could adopt inter-district public school choice programs and reduce the restrictions placed on existing programs.
Education freedom is substantially under the control of state policymakers, making it reasonable to rank states on the extent to which their policies provide education options to families. In Table 1, the states are listed by their ranking on the Education Freedom Index. In Table 2, the same list is sorted alphabetically. The method by which the EFI score is computed is described in the methodological appendix and in the following sections discussing each component of the EFI.
The Highest and Lowest Ranking States
Just as the Freedom House index of political freedom labeled countries as “free,” “partly free,” and “not free,” we are labeling those states with the ten highest EFI scores as “free,” those with the lowest ten scores as “not free,” and those in the middle as “partly free.” A map of the United States contained herein shows where the states in each category are located.
The state with the highest score on the Education Freedom Index is Arizona. Arizona gets high marks for its large number of charter schools, its relatively light regulation of home schooling, a private school tax credit, and its unrestricted inter-district school choice program. Arizona is closely followed by Minnesota, which scores well on the EFI because of its generous tax credits and deductions for private school expenses, its large number of charter schools, and its unrestricted inter-district public school choice program. Wisconsin is ranked third on the EFI. It receives a high score because of its pioneering voucher program in Milwaukee, inter-district choice, and openness to home schooling and charter schools. In fourth place is New Jersey, which is particularly accommodating to home schooling, makes it easy to move to a different school district, and provides direct subsidies for certain private school expenses. Oregon is ranked fifth on EFI, largely because of its lack of restrictions on home schooling and the existence of an inter-district school choice program. Texas is ranked sixth, primarily due to its accommodation of home schooling and its large number of charter schools. Delaware earned seventh place and Colorado earned eighth because of their strength in charter schools and inter-district public school choice. Maine is ranked ninth because of its significant efforts to assist families choosing private schools and because of the ease with which families can move from district to district. And Connecticut is tenth on the Education Freedom Index because of its direct subsidies to private schools and inter-district school choice program.
At the bottom of the Education Freedom Index ranking we find Hawaii. Hawaii receives low marks for having only one school district for the entire state, making moving to another district essentially impossible and precluding any inter-district choice program. In addition, Hawaii offers no assistance for private school expenses, has few charter schools, and heavily regulates home schooling. Also bringing up the bottom of the EFI ranking is West Virginia. West Virginia is highly restrictive of home schooling, has no charter schools and no inter-district choice program. The third lowest ranking state on the EFI is Nevada, which has large school districts and does not have an inter-district choice program, making moving or transferring to a new district very difficult. The fourth lowest score on the EFI belongs to Kentucky, which has no charter schools, no assistance for private school expenses, and no inter-district choice. Maryland is the fifth lowest ranking state, due largely to its large, countywide school districts, lack of charter schools, and lack of inter-district choice. Sixth from the bottom is Rhode Island, which is highly restrictive of home schooling and has no inter-district choice program. Virginia is seventh from the bottom because of its lack of assistance to private schools and absence of an inter-district public school choice program. The lack of private school assistance as well as heavy restrictions on home schooling puts South Carolina eighth from the bottom. Alaska is the ninth lowest-ranking state primarily because it has such large school districts, making moving to a different district very difficult. And the tenth lowest-scoring state is Georgia for its lack of private school assistance, large school districts, and restrictions on home schooling.
Both high and low ranking states vary. High-ranking states are found in the West (Arizona, Colorado, and Oregon), Midwest (Minnesota and Wisconsin), South (Texas), Mid-Atlantic (Delaware), and Northeast (Connecticut and New Jersey). Low-ranking states are similarly dispersed, with Alaska, Nevada, and Hawaii in the West, Rhode Island in the Northeast, Kentucky, Maryland, and West Virginia in the Border States, and Georgia, South Carolina, and Virginia in the South. Some top-ranked states on EFI have high median household incomes, such as Connecticut, New Jersey, Minnesota, and Wisconsin, while others have low average incomes, such as Arizona, Maine, and Texas. Similarly, some states that score poorly on the EFI are wealthy, such as Alaska, Hawaii and Maryland, while some are not, such as Georgia, Kentucky, and West Virginia.
The states with high and low EFI scores are also diverse in their politics. Among the ten highest-ranking states, Oregon, Minnesota and Wisconsin have voted for each of the last three Democratic Presidential candidates, while Arizona, Texas and Colorado have voted for two of the last three Republican nominees. So, too, with the ten lowest ranking states: Hawaii, West Virginia, Maryland and Rhode Island are among the most Democratic states in the Union, while Virginia, South Carolina and Alaska are among the most Republican.
The diversity of states that receive high and low rankings on the Education Freedom Index in terms of geography, wealth, and party affiliation, suggests that the obstacles to offering greater educational freedom have more to do with politics and policy than with any inherent characteristics of the states. Policymakers can choose to adopt programs that expand educational options and therefore increase educational freedom if they have the will to do so.
In the following sections, we discuss each of the five components of the Education Freedom Index and rank the states on each of those measures.
Charter School Options
The measure of availability of charter school options is itself comprised of two indicators: the percentage of all schools in a state that are charter schools and the score given to each state by the Center for Education Reform for the extent to which it has regulations favorable to the creation and operation of charter schools. The sources of the data and other details are discussed in the appendix.
As can be seen in Tables 3 and 4, Arizona has by far the highest score for making charter school options available to parents. More than a fifth of all schools in Arizona are charter schools. (Because charter schools tend to be small, the percentage of students in Arizona that attend charter schools is much lower). This concentration of charter schools is more than 4 times that found in the next highest-ranking state, Michigan. Not surprisingly, Arizona also receives the highest score from the Center for Education Reform (CER) for its charter school laws and regulations. Arizona has many charter schools in large part because its regulatory environment makes it easy to start and operate them, not simply because there is heavy demand in Arizona for alternatives to traditional public schools.
Michigan, with the second highest percentage of schools that are charter schools, also has the second highest score from the CER for its charter laws and regulations. The third highest total charter score belongs to North Carolina, which places fourth in its concentration of charter schools and seventh in its charter law and regulations.
Fifteen states tie with the lowest total charter score because they have no charter schools. These fifteen are: Alabama, Indiana, Iowa, Kentucky, Maine, Maryland, Montana, Nebraska, North Dakota, Oklahoma, South Dakota, Tennessee, Vermont, Washington, and West Virginia.
Government-Assisted Private School Options
The measure of government-assisted private school options is computed as the equally weighted average of three items: 1) the percentage of students in a state attending private schools with government-paid vouchers; 2) the maximum dollar value of tax credits and deductions for private school expenses; and 3) the extent of direct state subsidies to private schools for expenses such as transportation, textbooks, and health care. As shown in Tables 5 and 6, the state with the highest score for private school options is Minnesota. Minnesota has generous tax credits and deductions for private school expenses and offers private schools a broad set of direct subsidies for transportation, textbooks, and “auxiliary services.” Following Minnesota on the ranking of private school options is Maine. While this is little known, Maine actually has the largest “voucher” program in the United States whereby families are given government money with which they may opt to attend private, secular schools. This practice is known as “tuitioning” and occurs in rural areas that decided not to build public schools. Instead they offer families vouchers that they can use to attend a private school or nearby public school. Nearly 2% of all students in Maine attend private schools with these vouchers. A similar program exists in Vermont. In addition to hosting the largest voucher program, Maine also offers a range of direct subsidies to private schools for transportation, textbooks, and auxiliary services. Illinois ranks third on the private options measure. It offers a moderately generous tax credit and provides private schools with a full range of subsidies.
Thirteen states tie with the lowest score for offering families government-assisted private school options. They have no voucher programs, no tax credits or deductions for private school expenses, and no direct subsidies for their private schools. These states are: Alabama, Arkansas, Georgia, Hawaii, Idaho, Kentucky, North Carolina, Oklahoma, South Carolina, South Dakota, Utah, Virginia, and Wyoming.
The home-schooling measure is the average of two indicators: the percentage of students in the state that are home schooled (when this figure is available), and the absence of state restrictions and regulations on home schooling, according to information collected by the Home School Legal Defense Association. When the percentage of students home schooled was not available (which is the case in ten states), the second indicator was used as the total home-schooling score. The states are ranked by their home-schooling scores in Table 7 and alphabetically in Table 8.
The highest scoring state on the home-schooling measure is Oregon, where, according to conservative numbers collected by the U.S. Department of Education, nearly 2% of all students are home schooled. In addition, Oregon placed no restrictions on home schooling in six of the seven categories of possible regulations monitored by the Home School Legal Defense Association (HSLDA). Three states tied for second place on the home-schooling options score: New Jersey, Oklahoma, and Texas. In none of these jurisdictions were data available on the percentage of students who are home schooled, but all three placed no restrictions on home schooling in six of the seven categories tracked by the HSLDA. Trailing these three states was Alaska, which places no restrictions on home schooling and has more than 1.5% of its students home schooled according to the conservative estimates collected by the U.S. Department of Education.
The lowest scoring state on the home-schooling options measure is West Virginia, which places restrictions in all seven areas tracked by the HSLDA and has only a tiny fraction of its students home schooled. Rhode Island and Tennessee tie for second lowest. Both states restrict six of seven areas tracked by HSLDA. (Data on the extent of home schooling are not available for either state.) Just above them is New York, which restricts six of seven categories tracked by HSLDA and has a low percentage of home schoolers.
Inter-District Transfer Options
This is the least satisfying of the five measures that comprise the Education Freedom Index. We sought measures of intra-district choice programs, such as magnet schools, as well as inter-district choice. Unfortunately, reliable and consistent data on intra-district choice are simply not available at the state level. For inter-district choice, we sought a measure of the percentage of students who actually take advantage of the programs and/or a detailed measure scoring the state law for how easily families could choose to transfer their children to another district. Unfortunately, these data are not available, either. Instead, we were obliged to use the simple measure of the existence of inter-district choice programs as reported by Education Week in its annual Quality Counts.
According to that report (and listed in Tables 9 and 10), 18 states have full inter-district choice programs and therefore receive the highest score on this measure. Those 18 are: Arizona, Arkansas, Colorado, Connecticut, Delaware, Idaho, Iowa, Minnesota, Nebraska, New Mexico, North Dakota, Ohio, Oregon, South Dakota, Tennessee, Utah, Washington, and Wisconsin. Another 12 states have limited inter-district choice plans, giving them a middle score on this measure. The twenty remaining states have no inter-district choice program. These states, which receive the lowest score on this measure, are: Alaska, Florida, Georgia, Hawaii, Illinois, Kansas, Kentucky, Maryland, Mississippi, Montana, Nevada, North Carolina, Oklahoma, Pennsylvania, Rhode Island, South Carolina, Vermont, Virginia, West Virginia, and Wyoming.
While this measure does not capture the full range of options for transferring within the public school system, it does provide some information that helps rank the amount of education freedom available in each state.
Moving to a Different District
One of the most common ways by which people exercise choice in education is by moving to a different school district if they are unsatisfied with their current prospects. The flight to the suburbs and the prominence of school quality in housing selection illustrate this common, but little recognized, form of school choice. Relocating to a new district, however, has costs. In addition to the barriers some people face affording homes in areas with high quality public schools, relocating can force parents to leave attractive jobs and can disrupt their networks of family and friends. State policy can help alleviate these costs to some degree by altering the size and population of districts. If switching to another public school district does not involve having to move a great distance, then families can relocate without the parents having to change jobs or leave networks of friends and family.
Our measure of relocation options is an average of two items: the average number of square miles per district in each state and the average number of students per district in each state. The ease with which one can move to another school district without having to change jobs or break social ties is a function both of the distance between districts and how densely populated those districts are. In California, for example, the average school district is approximately 147 square miles, making its districts geographically smaller than those of 30 other states. But the population density is greater than all but 12 states, meaning that it is harder to commute across those miles. The number of miles from the center of the Los Angeles Unified School District to neighboring districts is not great, but given the population density, the commute would be much worse than covering the same number of miles in Montana. States can make it easier to choose alternative school districts by moving if they made their districts smaller in terms of population and geographic size.
As can be seen in Tables 11 and 12, the highest ranking state on this measure of moving options is Vermont, which has the second smallest average population per school district, with a little more than 300 students per district. Vermont also has very small districts, about 27 square miles on average, making it relatively easy to move to another district without changing jobs or friends. New Hampshire has the second highest score. It has under 900 students per school district and an average school district size of around 36 square miles. (Keep in mind that a square-shaped district that is 36 square miles is 6 miles on each end, making a commute from the middle of the district to the middle of a neighboring district a matter of minutes in thinly populated areas.) Nebraska closely follows Vermont and New Hampshire on the measure of moving options. It has a little more than 400 students per district and an average district size of around 100 square miles.
On the other end of the spectrum, Hawaii receives the lowest score by far on this measure. It has only one school district that encompasses more than 200,000 students and more than 6,400 square miles scattered over a string of islands in the middle of the Pacific Ocean. Choosing to move to a different school district would require traveling thousands of miles and leaving the state. Alaska similarly receives a very low score on the moving options measure. Though the average number of students per district in Alaska is actually lower than in 30 other states, the distance to the next district is usually very great, with more than 10,000 square miles per district. Nevada has both a relatively high number of students per school district, more than 15,000, and districts that are geographically large, averaging more than 6,000 square miles.
Does Education Freedom Help Student Achievement?
For many, freedom is a good in its own right. They contend it is desirable for policymakers to promote education freedom simply because it makes available to families a broader set of options to satisfy their diverse needs and priorities. But does freedom in education also boost student achievement? The evidence suggests that it does.
According to a simple regression that predicts the percentage of students scoring “proficiently” on the National Assessment of Educational Progress (NAEP) by each state’s EFI score, a state that has a one point higher score on the EFI can expect an additional 5.5% of its students to test as proficient. Given that the average state has 26.5% of its students scoring at the proficient level on the 1998 Reading NAEP, the 1998 Science NAEP, and the 1996 Writing NAEP, an increase of 5.5% is quite an improvement.
From other regressions that use the EFI to predict SAT scores, we would expect that a one point increase in the Education Freedom Index score for a state would lead to a 21 point increase in the average SAT verbal score and a 22 point increase in the average SAT math score. Put simply, states with more education freedom have higher average student achievement.
But does education freedom actually cause higher student achievement? Proving causation with relatively few cases (there are only 50 states) cannot be done with certainty. Even to suggest that education freedom helps contribute to student achievement requires controlling for other factors that might explain that achievement. Accordingly, we constructed a simple model that uses the Education Freedom Index scores to predict student achievement in each state, controlling for the percentage of minorities in each state, median household income in each state, per pupil spending on education, and the average class size.
As can be seen in Table 13, the results are striking. Even after controlling for these other factors, EFI is a significant predictor of student achievement. We would expect that a one point rise in EFI for a state would increase the percentage performing proficiently on NAEP by 5.5%. In contrast, a one percentage point increase in minority population would lead to a .3% decline in the percentage of students performing proficiently, while a $1,000 boost in median household income would lead to a .3% increase in the percentage of students performing proficiently on NAEP. (Per pupil spending and average class size in each state have no significant effect on the percentage of students who perform proficiently on NAEP.)
We observe similar results when we use the model to predict SAT scores. A one point increase in the Education Freedom Index leads to a 24 point increase in SAT verbal and math scores. Among the control factors, an increase in the minority population of one percentage point is associated with a 1 point decline in SAT verbal and math scores while household income, per pupil spending, and class size are not independently significantly related to SAT scores, although the income measure is close to significant in predicting SAT verbal scores.
Increasing education freedom is one of the most practical ways for policymakers to improve the quality of education. To achieve the same benefit as a one point gain in EFI (which is about the difference in education freedom between #3 ranked Wisconsin and #40 ranked Wyoming), state policymakers would have to find a way to increase average household income by $19,000, which is simply not feasible. While increasing education freedom can require politically difficult reforms, it is certainly a more attractive course of action for policymakers than the nearly impossible or undesirable alternatives required to produce similar gains in academic achievement.
A Comparison of South Carolina and Texas
The effect of education freedom on student achievement can be illustrated by comparing South Carolina and Texas, two states similar in many respects but different in the educational options they offer families. Both are southern states with a high percentage of minorities. (Around 41% of Texas students are African-American or Latino compared to 31% for South Carolina.) Both states have median household incomes below the national average, South Carolina at $34,665 and Texas at $33,072. South Carolina spends $5,827 per pupil compared to $5,815 in Texas, both below the national average. The average student/teacher ratio in both states is smaller than the national average, with 15.7 students per teacher in South Carolina and 15.5 in Texas.
While these two states are similar in characteristics that are commonly thought to influence pupil achievement, they differ significantly in the amount of education freedom they offer. South Carolina has an Education Freedom Index score of 1.64, putting it 43rd among the states. Texas has an EFI score of 2.48, putting it in 6th place among the states.
Texas also has higher test scores than South Carolina. In Texas, 24.0% of students score at the proficient level on the NAEP compared to 17.7% in South Carolina. The average verbal and math SAT scores in Texas are 494 and 501, respectively, compared to 479 and 474 in South Carolina. Texas has 6.3% more students scoring proficiently on NAEP, 15 point higher scores on the SAT verbal, and 27 point higher scores on the SAT math.
Many factors undoubtedly contribute to these differences in test scores, but Texas may have significantly higher test scores than South Carolina in large part because it offers greater education freedom. The Lone Star state does not spend more on education per pupil, it does not have significantly smaller classes, it does not have wealthier families, and it does not have fewer minority students. What Texas has is more freedom. Texas offers more charter schools than South Carolina. Texas places far fewer restrictions on home schooling. Texas offers a limited inter-district school choice program while South Carolina does not offer any. And Texas has smaller school districts than South Carolina, making it easier for families to move to different districts without having to disrupt work or social connections.
By offering families greater educational options, Texas helps students learn. This higher level of student achievement produced by more educational options could be caused by families being better able to find schools that address the particular needs of their children. If one has more options from which to choose, it is easier to find the school that is right for a child and that will help the child learn more. If there are more educational options, schools also have to be more attentive to student needs to attract students. Schools cannot simply take students for granted if families have alternatives.
Freedom matters in education. Simply providing families with additional options in the education of their children has a larger independent effect on student achievement than increasing education spending or reducing class size. In addition, the magnitude of the benefit of education freedom for student learning is comparable to the benefit of significantly increasing median household income. Note, too, that it is far easier for state policymakers to expand education freedom than it is for them to increase median household income by several thousand dollars. Education freedom is not only a good in and of itself, but it also appears to help students learn.
We hope that the Education Freedom Index draws attention to the importance of freedom in education. Some states are beginning to recognize the importance of this freedom and are expanding the education options available to families. Other states have done little in this regard. The Education Freedom Index will perform its task if all states begin to debate the extent to which parents should have the freedom to choose how to educate their children.
The Education Freedom Index is an equally weighted average of five measures: charter school options, government-assisted private school options, home-schooling options, inter-district transfer options, and relocation options. In this appendix we explain each of those measures, including the sources for data and how they were computed.
All items were “standardized” so that the score each state received is the number of standard deviations above the lowest scoring state. A standard deviation is a measure of variance. By transforming all scores into units of standard deviations we are able to make scores of unlike things comparable. For example, by standardizing we are able to average scores that are percentages of charter schools with scores that are rankings of charter regulations. Having twice the percentage of charter schools as another state may not be as big of an advantage in terms of freedom as having twice the score on charter regulations because there may be more variance in the charter school percentage measure. That is, having twice as many charter schools may be easier than having twice the regulation score. Converting all measures into units of standard deviations allows us to treat all measures as comparable and therefore makes averaging across them meaningful and valid. We therefore standardized all components of the Education Freedom Index as units of standard deviations above the lowest scoring state on that item. Converting all scores into units of standard deviation does make it a little harder to interpret the scores, but it is necessary if one wants to combine several scores from unlike measures.
The charter school score is an average of two measures, the percentage of public schools in a state that are charter schools and a score for the laws and regulations governing the creation and operation of charter schools. The total number of public schools in each state was obtained from the National Center for Education Statistics’ web site at: http://nces.ed.gov/pubs99/digest98/d98t098.asp. The number of charter schools in each state was obtained from the Center for Education Reform’s web site at: http://www.edreform.com/pubs/chglance.htm. The percentage of schools that are charter schools was computed simply by dividing the number of charter schools by the total number of public schools in each state.
The extent to which each state’s laws and regulations enable the creation and operation of charter schools was scored by the Center for Education Reform (CER). This score considers numerous factors, such as whether there are multiple chartering authorities, whether charters are exempt from regulations, and whether they are free from collective bargaining requirements. The CER score can be found on-line at: http://www.edreform.com/charter_schools/laws/ranking_2000.htm.
Private School Assistance Scores
The government-assisted private-school options score was computed from three measures: the percentage of students attending a private school with a government paid voucher; the maximum benefit available from tax credits and deductions for private school expenses; and the range of direct state subsidies to private schools for certain expenses.
To compute the percentage of students receiving a voucher we obtained the total number of public and private students in elementary and secondary school for each state from the National Center for Education Statistics on the web at: http://nces.ed.gov/pubs99/digest98/d98t040.asp and http://nces.ed.gov/pubs99/digest98/d98t064.asp. To the best of our knowledge, publicly financed vouchers are available in five states. The number of students receiving vouchers in Wisconsin, Ohio, and Florida was obtained from George Mitchell of the Institute for the Transformation of Learning at Marquette University. He can be reached at email@example.com. The number of students receiving vouchers in Maine and Vermont was obtained from Richard Komer of the Institute for Justice in Washington, D.C.. Mr. Komer can be reached at firstname.lastname@example.org.
Information on tax credits and deductions for private school expenses in each state was obtained from the School Choices 2000 report issued by the Heritage Foundation. It can be found on the web at: http://heritage.org/schools.
The range of direct subsidies to private schools was a four-point scale computed from data collected by the U.S. Department of Education. In its June 2000 report, The Regulation of Private Schools In America: A State-by-State Analysis, Table 3 indicated whether states provided subsidies to private schools for pupil transportation, textbooks, and auxiliary services, such as school nurses or educational materials. A state received a score of 3 if it provided subsidies in all three categories, a 2 if it provided subsidies in 2 of these three categories, and so on. Table three of the report can be found on-line at: http://www.ed.gov/pubs/RegPrivSchl/chart3.html.
These three measures were then standardized in the same manner as described above. The government-assisted private school options score is an equally weighted average of these three standardized measures.
The home-schooling options score is an average of two measures: the percentage of students in each state who are being home schooled, as reported by estimates from the U.S. Department of Education; and a measure of the extent of restrictions on home schooling in each state computed from information collected by the Home School Legal Defense Association. The percentage of home-schooled children was computed by dividing the number of home-schooled children as estimated in a U.S. Department of Education report, available at http://www.ed.gov/offices/OERI/SAI/homeschool/AppendixA.html, by the total number of elementary and secondary school students, available at the web sites listed above. The home-schooling estimates, for school year 1995-6, were available for forty states.
The measure of restrictions facing home schooling was based on descriptions of each state’s laws and regulations governing home schooling found on the Home School Legal Defense Association web site at: http://www.hslda.org. The web site describes seven categories of state regulation. Each state’s score on this measure was the percentage of those seven categories in which there were no restrictions placed on home schooling.
Both of these measures were then standardized, using the same method described above. The total home-school options score was the average of these two standardized measures when both measures were available. For those ten states without information from the U.S. Department of Education on the number of home schoolers, the total home-school option score was simply the same as the home-school law measure.
Inter-District Public School Choice Score
The inter-district transfer options score is a single measure obtained from the Quality Counts 2000 report published by Education Week. This report, available on-line at http://www.edweek.org/sreports/qc00/templates/state-comp.htm, says whether each state has no inter-district school choice, a limited inter-district choice program, or a statewide inter-district choice program. From this description we calculated a 3-point scale, where a state received a 3 if it had statewide inter-district choice, a 2 if it had limited inter-district choice, and a 1 if it had not inter-district public school choice program. This measure was then standardized in the manner as described above.
The relocation option score was calculated as the average of two measures, the state’s average number of students per school district and the average number of square miles per district. The number of districts in each state was obtained from the National Center for Education Statistics at: http://nces.ed.gov/pubs99/digest98/d98t092.asp. The number of students was obtained from the sources described above. The number of square miles per state was obtained from the U.S. Census at: http://www.census.gov/population/censusdata/90den_stco.txt.
Both of these measures, square miles per school district and student population per district, were then standardized in the manner described above. The total relocation options score was then the average of these two standardized measures.
Additional measures were obtained for the regression models. The information on NAEP proficiency scores was obtained from Quality Counts at: http://www.edweek.org/sreports/qc00/templates/state-comp.htm. The percentage of students scoring “proficient” on the NAEP was the average reported as proficient on the 8th grade reading test in 1998, the 8th grade science test in 1998, and the 8th grade math test in 1996. If scores were not available for all of these tests (due to a state’s non-participation in NAEP), then the average of those available was used. The average SAT verbal and math scores for each state were obtained from the National Center for Education Statistics at: http://nces.ed.gov/pubs99/digest98/d98t134.asp.
Information on median household income for each state was also obtained from the National Center for Education Statistics at: http://nces.ed.gov/pubs99/digest98/d98t020.asp. The percentage of African-Americans and Latinos in each state was obtained from the U.S. Census at: http://quickfacts.census.gov/qfd/index.html. Per pupil spending was obtained from Quality Counts, which adjusts spending to account for differences in the cost of living by state. The student-teacher ratio was obtained from the National Center for Education Statistics at: http://nces.ed.gov/pubs99/digest98/d98t067.asp.