As I have pointed out, one of the easiest ways to improve test scores in a class or a school is to get rid of the students who generally score low, and to retain the ones who do well. I suspect that may have been the secret behind the fact that Michelle Rhee, after two years of complete and utter failure with her second-grade classes, finally managed to raise the scores to somewhere near the 50th percentile mark.

It appears that the ‘weeding-out’ was pretty drastic.

The cohort that started the first grade at Harlem Park in 1992-1993 had 84 students, probably 3 or 4 distinct classes.

When they arrived in the second grade in 1993-1994 and endured Michelle Rhee’s second failed year of teaching, they still had 83 students – probably 3 or 4 classes again.

But when this cohort arrived in the third grade in 1994-1995, Rhee’s “miracle year”, * their numbers dropped by nearly half, to only 44 students*. I doubt strongly that so many students dropped dead. I can’t prove it, but I would not be surprised if the school (and Rhee) ‘counseled out’ the ones who were doing poorly, and kept the ones who had high test scores.

You may be wondering if the same thing happened at other Edison schools and at the regular Baltimore public schools that were used for comparison purposes. Look for yourself at the graphs below – and I think you will conclude that while some of the cohorts in some of the schools had small rises in populations, some went down a little bit, and some stayed about the same. **NONE OF THEM HAD SUCH HUGE DROPS AS WE SEE IN MICHELLE RHEE’S “WONDER YEAR.”**

(By the way, this trick doesn’t always seem to work. The first grade in 1991-2 at Harlem Park had 102 students, and scored at the 63rd percentile in math in the spring. The next year, they were promoted to the second grade, and only had 77 to 84 students, depending on which count you follow, and they had Rhee as one of the second grade teachers. The average percentile rank of this cohort fell from the previous (respectable) 63rd percentile to an abysmal 18th percentile. As someone might say, echoing Michelle Rhee’s recent interview with Harry Jaffe,

**“ these kids were getting screwed because people wanted to blame their low achievement levels on the single-parent households and on the poverty in the community. In that two-year period, none of those things changed. Their parents didn’t change. What changed? What we were doing with them in school.” **

Well, if she wants to take credit for a 20-percentile rise in scores from 1994 to 1995, then she needs to take the blame for a 45-percentile-point DROP from 1992 to 1993, even with the weeding out.)

As has been the case for my past several data-driven posts, all of these tables and graphs were made by me from the tables in the appendices to the UMBC Evaluation of the Tesseract Program in Baltimore City, written by Lois C. Williams and Lawrence E. Leak in 1995. You can find it on the web at this URL.

The STATE tested kids below the 4th grade level there? That’s unusual, isn’t it?

I don’t know.

[...] This post was mentioned on Twitter by Jeff Liwag and gail simmons, Joe Bower. Joe Bower said: The Secret to Raising Test Scores? (Part 2) http://bit.ly/hqQEO9 <<Rhee the liar (part 2) #abed #edchat #edtech [...]

Michael, Yes, kids were given the CTBS test starting in first grade. I don’t if it was a mandate from the state, but it was done in Prince George’s County.

Amazing. The lower the socio-economic level of the district, the higher the probability they’ll be testing kids on math and literacy in Kindergarten, if not any public pre-school program.

Soon, testing math in the womb, coming to a district near you.