Wednesday, December 30, 2009

More on Texas Charters

A couple of days ago I pointed out the explosive growth in enrollment in charter schools in Texas shown on this Texas Policy Foundation graph.

Notice charter school enrollment exploded in Texas after 1996. And, this graph does not include another group of thousands more Texas students enrolled in special charter school districts. Overall, I estimate that charter school enrollment in Texas now amounts to around three percent of the entire school population in the Lone Star State.

Because the rise in charter enrollment was so dramatic, I was curious about what might have happened to Texas on the National Assessment of Educational Progress (NAEP) after charters were introduced. Was there any indication that charters impeded or helped? I focused on eighth grade math because this is one of Kentucky’s major educational stumbling blocks, especially for our minority students.

What I found is remarkable.

This graph shows how blacks in Texas and Kentucky (which still has no charter schools) did on the federal NAEP tests. Keep in mind that Texas’ charter school growth began after the 1996 NAEP was administered.

Back in 1990 when the first state level NAEP math test was conducted, Texas’ blacks scored 234 while our black kids scored 240. Note that Texas’ blacks then essentially caught up to our kids and stayed even with them until a couple of years after the Texas charter movement started. After the year 2000, however, Texas’ blacks left our’s in the dust. As of the new results for 2009, which were just released a short time ago, Texas’ blacks scored a dramatic 14 NAEP Scale Score points higher than our kids.

Basically, when it comes to the only racial minority present in Kentucky in significant numbers, KERA got shellacked by Texas’ education system.

Also, to reiterate, note that the difference in performance began to appear several years after Texas started its charter program.

Unfortunately, the NAEP data can’t tell us how blacks in charter schools in Texas performed over time. The NAEP samples from the charter schools were too small to provide statistically reliable information.

All we can tell from the NAEP is that the overall performance for all students in charters in Texas lagged the statewide average for all Texas students in 2005, 2007 and 2009 NAEP assessments but that Texas’ charters reduced that lag by two points in that time.

However, there is no way to know how the students in Texas charters would have scored had they remained in the regular school system. Likewise, there is no way to know if the nine point improvement in performance of Texas’ blacks who attended non-charter schools in 2005 to 2009 would be reduced if those schools had not faced competition from charter schools.

What we do know is that something dramatic is happening for blacks in Texas, something far better than Kentucky has offered its blacks. And, we do know that the timing of the increase in charter school enrollment in Texas fits nicely with the possibility that charter schools have played a role, maybe a major role, in making this happen.


Anonymous said...

Looks like the blacks in Texas are smarter than our Legislators. I'll bet you are not surprised.

Anonymous said...

It is apparent that the teachers unions in TX are either less powerful than the KY teachers unions or less radical. Either one works for me.

The reason teachers unions oppose charter schools, private schools, and home schools is related to political power which they appear to believe is more important than the quality of instruction students receive.

I realize that focusing on minority students' performance is how we get liberals to listen, but I sure would like to see less division by race, etc.

IMHO, categorizing by race is a genuine form of racism. I know liberals won't admit this, but it is true.

Richard Innes said...

RE: Anonymous, December 30, 2009 9:22 PM Part 1

It does appear that the teachers’ unions in Texas operate somewhat differently from the current situation in Kentucky.

In addition to the possibilities you mention, it could be that the unions in Texas are more in tune with the way I think a professional organization should operate. While member representation and service is important, professional organizations also work to improve the state of the art of their profession and strongly value customer service their members provide.

In the case of teachers, that would mean understanding that the needs of children should come ahead of such things as allowing the best qualified, longest tenured teachers to avoid service in the most needy and demanding schools. That isn’t consistently happening in Kentucky.

During testimony at the October Kentucky Board of Education meetings, two school principals and the superintendent from Jefferson County admitted that two of the school district’s lowest performing schools didn’t get the experienced teachers they needed because of the union contract. That contract provision is unworthy of professionals.

In fact, the National Education Association not long ago sent out a message to local union affiliates to remove such language from their contracts. So far, that has changed nothing in Louisville.

Richard Innes said...

RE: Anonymous, December 30, 2009 9:22 PM Part 2

Regarding your second set of comments about breaking out educational performance by race, we unfortunately don’t live in a world where we can do the right thing for children by ignoring data by race. The facts are that ignoring our education system’s very different performance for different races would hide serious disparities while allowing the situation to continue unchanged. Not collecting data by race just allows the true “bigotry of low expectations” to continue.

For example, one of the biggest failings in Kentucky’s old CATS and KIRIS assessments was that neither held any penalty for schools with big performance gaps. In fact, the way the CATS system was designed, a school could do a very inadequate job for a very significant sub-proportion of its students and still wind up with high enough Accountability Index Scores to totally escape sanctions all the way to 2014.

I did a paper several years ago that showed schools could avoid CATS sanctions in 2014 with performance that included math proficiency rates as low as 39 percent and writing proficiency rates of 0 percent (on line here: That provided more than enough slack for schools to largely ignore minority performance and still look stellar under CATS.

CATS was designed (perhaps inadvertently, perhaps not) to ignore serious failure for minority kids in Kentucky.

It took the federal No Child Left Behind (NCLB) Act to finally bring the performance gap issue in Kentucky to real attention. NCLB definitely required separate performance determinations by race, but that was not a racist provision – far from it. It was placed in the federal statute so that the education system could no longer hide the very different educations the different races were receiving.

There is another reason why the Bluegrass Institute regularly uses NAEP data disaggregated by race. State to state comparisons with the NAEP can be very misleading if only the overall average student scores for each state are examined. The reason is that student demographics have changed dramatically in many states and overall across the nation. Meanwhile, here in Kentucky our student demographics have remained remarkably stable since NAEP started to report state-level results in 1990.

For example, Kentucky’s NAEP samples run around 85 percent white and 10 percent black, plus or minus a few points, in all of the NAEP state testing to date.

In California, by way of contrast, back in the early 1990’s the state was about 51 percent white. Today California’s student makeup is only about 27 percent white. The change has largely come from immigrant Hispanics, many of whom don’t speak English as a first language.

The education problem in California is therefore much more severe than the one we face in Kentucky. With many severely under-educated students flooding its system, California’s overall all student NAEP scores don’t reflect what is actually happening in the schools there.

Kentuckians would be very unwise to fool ourselves into feeling good by only doing simplistic comparisons of overall all student scores for Kentucky and California.

You have to look deeper, and because the NAEP currently has no correction for state-to-state demographic differences (or differences due to different rates of exclusion of students and provision of testing accommodations), the only reasonable way to do a better analysis is to examine the disaggregated data (You can learn more about problems with interpreting the NAEP here:

In a perfect world, we wouldn’t have to do all of that analysis. Our education system would serve all children equally, and disaggregation of data would be unnecessary.

Sadly, we don’t live in such a world today.