Thursday, February 08, 2007

The Bomb or just Bombing?

Denver Post: Science-tech charter school a beacon to Texas reformers

One of Denver's highest-performing high schools will become a model for 35 new schools in Texas as they undertake a massive education reform project, administrators announced Wednesday.

The Denver School for Science and Technology, a charter school nestled in an affluent section of the Stapleton neighborhood, this summer will begin training teachers, principals and administrators from Texas about how the school pulls off such high test scores.

I can sum up the training in a few words: "Our secret to improving test scores is to get rid of the deadwood"

Yep... that's right. I decided to check out DSST so I went over to their website and then checked out their disaggregated CSAP data.

I downloaded their data in reading and math for 2005 and 2006 and compared the numbers between their 2005 9th graders and 2006 10th graders to see how much "real" improvement there was.

The first thing I noticed, was that the number or returning black students was dismal. In 2005 there were 51 black students in 9th grade. One year later, the number of black students in 10th grade was only 36. This is almost a 33% drop out rate. The only American Indian dropped out, there were 4 less white kids, though Hispanics gained 3 students. Number for Asians stayed the same. The net change was from 121 to 104 students. In total the class was 14% smaller from one year to the next. Now let's see if we can figure out just which students dropped out.

In 9th grade, 16 black kids scored unsatisfactory on the CSAP in math. 20 scored partially proficient, 14 proficient, and only 1 scored advanced. One year later, 6 scored unsatisfactory, 15 scored partially proficient, 15 scored proficient, and not a single black student scored advanced. It appears that the loss of most, if not all of the black students was from the unsatisfactory and partially proficient categories. Now I supposed its theoretically possible that the kids who dropped out were all proficient and advanced, and that the unsat and partially proficient kids improved, but... would you bet against it?

Among Hispanic and white kids, the numbers show us that the distribution of students scoring in various categories in math didn't change much. Though I did note that in every single category that they had data for, there were less kids scoring advanced than the year before.

The school did do a tiny bit better in reading, but the kids also started out a little better. In both 9th and 19th grade there was only one black kid in the whole class that scored unsatisfactory. It appears that the big drop of in numbers happened among the partially proficient and proficient students readers. While there were 3 more advanced readers among blacks, there was 7 and 11 less partially proficient and proficient readers. Once again, we can see that the school had a small improvement in reading ability, but that many of their lower performing students likely dropped out. There were minor improvements among whites and Hispanics as far as score distribution goes.

Overall it appears that the school does a decent job of improving the reading ability of students, though at the cost of getting rid of lower performing kids.

Once again we learn that percentages and stats can hide the truth. There are things to admire at this school. They put a premium on values, they provided a safe enviroment, but academically I suspect they are no better than the average middle class school at accelerating the performance of their students. But like other public schools, it appears they fail at teaching and retaining low performing minority students.

So Texans, enjoy your trip to Denver. You might not learn anything that you don't know already, but at least you will escape the Texas heat for a while.

Hat Tip: Joanne Jacobs

Update: I checked out High Tech High in San Diego. California's results are harder to parse, especially since the school switched from Integrated Math to a standard Algebra I, Geometery, Algebra II schedule, but from 9th grade 2004 to 11 grade 2005, the number of black students in each sequential class dropped from 22 students to 13.

In 2006, not a single black student scored proficient or advanced in math. Only 29% of white students did. Silly me for thinking that "High Tech" would mean math proficiency. It probably explains why they switched their math curriculum.

Both High Tech High and DSST emphasize "inquiry" and "project based" learning. It must be so nice to have a digital portfolio of your students work to remember them by after they drop out.