Monday, November 26, 2012
Florida's charter schools perform worse than its public schools
From the Tampa Times by Stanley D. Smith
As a professor of finance, I advocate the use of business analysis in evaluating government programs. So when the most recent state test scores came out for Florida's elementary schools, I ran some numbers to look at the performance of charter schools as well as the effects of poverty and minority status. The bottom line? The numbers tell us we should question the state's increasing emphasis on charter schools because as a group they underperform traditional public schools.
Here's what I found:
• When the poverty and minority characteristics of the student population are controlled, the average charter school performs significantly lower than the average traditional public school by a little more than 5 percent. This has major public policy implications.
• Across all schools — traditional and charter — as the percentage of students qualifying for a free or reduced price lunch increases, school scores decrease. That by itself is not surprising.
• But what might be surprising is this: As the percent of minorities increases, the school scores increase — if poverty is controlled for. This also has major policy implications. These results mean the state should be focusing performance goals on socioeconomic status, not race or ethnicity.
Here's how I conducted my analysis. To capture the effects of poverty (as represented by the percent of students who qualify for a free or reduced-price lunch) and a school's minority status (as represented by the percent of students who are minorities) I ran a statistical technique called regression analysis on the 1,812 elementary schools' preliminary school test scores.
Regression analysis attempts to estimate how much of the variation of one measure — in this case, school test scores — is determined by other variables, income status and minority status. These two variables "explained" about 50 percent of the variation in the school test scores with income level being the more important of the two variables.
Lower income levels are highly correlated with minority status, and that is why it is necessary to control for income level when evaluating minorities' performance. If one does not control for income level then one might reach the conclusion that minorities performed at a lower level. In this study, after income level is controlled, the minorities did better than the nonminorities.
I then added the 147 charter schools to the regression model. When the poverty and minority characteristics of the student population are controlled, the average charter school performs significantly lower than the average traditional public school.
If poverty and minority status are not controlled, the average charter school performed just slightly worse than the noncharter schools. Noncharter schools have a higher percentage of students who live in poverty or near poverty (68 percent versus 57 percent), and charter schools have a higher percentage of minorities than noncharter schools (68 percent versus 60 percent).
These results call into question the Republican emphasis on charter schools. Not all charter schools are performing worse than noncharter schools, even with the adjustments. However, my analysis shows that a parent has to be even more careful about enrolling a child in a charter school than in a traditional school.
Although charter schools may be cheaper for the state to fund, the adjusted scores suggest that Florida is also getting a lower return on these schools. Is the lower average return on these schools worth the lower cost? With the state allocating more funding to building physical facilities for charter schools, the lower cost advantage may disappear while the lower return persists.
The lack of recognizing the importance of income level was also evident last month when the Florida Board of Education set learning goals by race and ethnicity with Asians having the highest goal, followed by whites, Hispanics and blacks, in that order. Former Gov. Jeb Bush is quoted as saying that such systems would send a "devastating message" that black and Hispanic youngsters weren't as capable as others. The bigger problem is not that it would be devastating but that the message is not correct.
The problem with the racial goals is that it means the Board of Education, which has no black member, is looking at our students through a racial or ethnic lens rather than an income-level lens. Learning goals should be set by income level rather than racial category.
If the board admitted that income — not race or ethnicity — is the primary factor, then the next question after the return on charter schools should be what other types of relevant social aid would help our most needy and vulnerable children to have a chance to succeed?
Stanley D. Smith, professor of finance at the University of Central Florida, wrote this exclusively for the Tampa Bay Times. The full analysis and the adjusted scores for all elementary schools may be found and downloaded at http://www.bus.ucf.edu/faculty/ssmith/page/The-Effect-of-Poverty-Minority-Status-and-Charters-on-Florida-Elementary-Schools.aspx.