Are Charter Schools the Answer to Education's 911 Call?

At last Friday's hearing on a bill to dramatically expand authority to approve and operate public charter schools in Kansas, one of the proponents offered a striking image: many public school students, especially from low income families and minority groups, are facing an educational emergency. The answer to that 911 call for educational help, said proponents of SB 196, is expanded school choice; specifically allowing entities other than public school boards to approve and operate publicly-funded schools.

Let's agree at the start that too many students do not reach the educational levels they will need to be successful, and that despite progress - which this blog has celebrated in the past - Kansas and the nation still have a large "achievement gap" in educational outcomes. If charter schools can be shown to help close that gap, it would be a powerful argument in favor of changing Kansas law. States have been experimenting with charter school laws for 20 years, and charter school enrollment has increased dramatically. What does the record really show?

The research is, at best, mixed on the impact of charter schools on students when comparing charter schools to traditional public schools.  I wanted to look at the impact of charter schools on the education achievement of the state as a whole; in other words, state to state comparisons. To do so, I considered KASB's research on state educational achievement, all based on readily accessible national reports.

The first chart below provides overall national ranks of achievement in four areas: pre-school tests (the National Assessment of Educational Progress), high school completion, preparation for college and adult educational attainment.  Across all four of these categories Kansas is tied for 8th in the nation. So we should start by pointing out that under our state's current charter school law, Kansas has higher performance than 80% of the other states, including many with more liberal charter school policies.




However, states vary significantly in the student populations they have to educate. The next chart shows the percent of low income students (those eligible for free or reduced lunch) for each state, as well as the percent of majority (white) students.  There is a very high correlation (0.83) between a high educational achievement rank and a lower percentage of low income students. That means the more low income students in a state, the lower that state's achievement tends to be.

If you aren't familiar with statistical correlation, remember that 1.0 is the highest possible correlation; it essentially means a "one to one" correspondence between two variables. For example, New Hampshire has the lowest percent of free/reduced eligible students in the nation and ranks second in achievement; Mississippi has the highest low income population and the lowest achievement. Note also that ethnicity, although significant, has a considerably weaker correlation with achievement rank than income (0.49). That is not really surprising: economically advantaged minority students usually perform as well as majority students from similar income levels, and low income white students struggle academically.

Kansas is just below the national average in low income students and ranks 31st, which means Kansas would be expected to rank in the lower middle of states in achievement. However, Kansas' achievement rank is 8th, which means Kansas “over-achieves” its income position by 23 positions.  Florida, often mentioned as a high achieving state despite its poverty, ranks 43rd in income and 37th in achievement - “over-achieving” by five positions.

I have also highlighted the 10 states closest to Kansas in percentage of low income students.  Only two rank in the top 25 on achievement; Illinois at 20th and Indiana at 23rd.  Note that Arizona, which has a slightly lower percent of low income students than Kansas, ranks 26th in low income students but 43rd in achievement.

The next chart provides three columns of information on school choice.  The first column shows the percentage of students in charter schools in 2011 as reported by the federal government.  The second column has a ranking of state charter school laws by the National Association of Public Charter Schools, an organization that supports expansion of charter schools and multiple "authorizors" as proposed in SB 196.  Note that Kansas’ law is ranked 38th - which means we have what is considered a weak charter school law - but Kansas still ranks high in achievement, even after adjusting for student income.  (KASB ranks all states without charter schools as 50th.)  The third column indicates whether the state has tuition tax credits or other funding for private schools.

Note there is an almost insignificant 0.2 positive correlation between the percent of students in charter schools and achievement rank, and an even more insignificant negative correlation with a high charter school law ranking. In other words, charter schools make almost no statistical difference in state achievement.  Also, there is no statistical correlations between the percent of students in charter schools and the percent eligible for free lunch. This means charter enrollment is not disproportionately grouped in higher or lower income states.

The next table ranks states by an average of four National Assessment of Educational Process benchmarks: the percent of all students scoring at the basic level and at the proficient level, and the percent of low income students at basic and at proficient.  Note Kansas ranks 11th.  Of the 10 states ranking higher than Kansas, six had a lower percentage of charter school enrollment than Kansas, and five had a higher percentage.

Each of the ten states with a low income student percentage closest to Kansas (in pink) had a higher percentage of students in charter school. Yet Kansas ranked higher than every state with similar low income enrollment except Indiana - which spent more per pupil and had slightly fewer low income students.

There is an insignificant, but negative, correlation between charter school enrollment and NAEP scores, which essentially means that having more students in charter schools makes no difference in fourth and eighth grade reading and math. Note that Michigan and Arizona, two of the highest charter school enrollment states, are both similar to Kansas for low income students, but score in the bottom 10 states.

NAEP tests only provide achievement data up to grade eight. The next chart ranks states by an average of four different graduation rates from national reports.  As with NAEP scores, Kansas ranks 11th in the nation.  Of higher ranking states, six had fewer charter school students than Kansas.  Kansas had higher achievement than every other similar states based on low income students, although every other similar state had higher charter school enrollment.

The table also shows that some states with high NAEP scores fare much worse in actually graduating students. For example, Indiana ranks 7th in NAEP scores, but falls to 23rd in graduation; Colorado drops from 13th on NAEP to 29th in graduation, and Florida drops from 20th in NAEP to 42nd in graduation. Note that there is 0.29 negative correlation between the percent of students in charter school and the average completion rate.

The next chart shows the states where most students take the ACT test for college, including Kansas.  We use a formula to adjust the percent of students who meet college ready benchmarks in reading and math by the percent of students tested and the graduation rate.  Kansas ranks 9th, which would be an adjusted rank of 17th if applied to all 50 states.  Five of the eight higher-performing states had more students in charter schools, but every higher achieving state spent more per pupil.

Again, there is a modest negative correlation between charter school enrollment and the percent of students meeting college-ready benchmarks, which means states with more students in charter schools are somewhat less likely to have higher achievement. Note that Florida ranks third from the bottom, and Arizona, the state with a higher percentage of students in charter schools, ranked last. On the other hand, some high achieving states like Minnesota, Utah and Colorado have a high percentage of students in charter schools, but also have a much lower percentage of low income students.

The final chart ranks states by a combined average of the percentage of adults 25 and older who have completed high school, completed a four year degree, and completed an advance degree.  Kansas is tied for 12th.  Of the 11 higher performing states, five have lower charter school enrollment, and every higher achieving state had at lower percentage of low income students than Kansas except New York.

This data is very clear. First, Kansas is already a high achieving state, despite having more low income students than most states. Second, statistically having more students in charter schools makes no difference in achievement, and in some cases has a modest negative correlation. Third, Kansas already does better in achievement in almost every case than similar states, regardless of charter school enrollment.

What the data shows is that states with fewer low income students tend to have higher achievement, regardless of charter school enrollment. On the other hand, high charter school enrollment doesn't help states with higher poverty do better. In other words, charter schools haven't solved the 911 emergency of student achievement.

What does help boost achievement? We also provide the total revenue per pupil for each state. In every case, total revenue per pupil has a statistically stronger correlation with achievement than charter school enrollment.

Let’s be clear: we can’t be satisfied with Kansas outcomes.  We need to be truthful about the number of Kansas students who are not achieving at satisfactory levels.  But we also need to be truthful about the equal or greater achievement problems in high charter school states.  Expanding charter schools has not proven to be the answer.