COMPARE TEST TO PREDICTED SCORE BASED ON INCOME/LEP \doc\web\98\07\testlu.txt Date sent: 12 Oct 1998 16:07:20 -0700 From: "Mike McKeown" Subject: Re: More of proof of WA reforms "improvements" To: arthurhu@halcyon.com, "Lynn Stuter/WA" RE>More of proof of WA reforms* 10/12/98 The difficulty with these things is that they have done no pretesting. As we know from California, the first year tells you where you are, but can't tell how much you improved (or declined) from the year before. Paul Clopton and I have been working with a way to get some useful information from even single year data. Our strategy is essentially to set up a way to compare similar schools and then determine which are really doing better or worse. This takes advantage of the sad fact that socioeconomic status corrrelates shockingly well with academic success. For example, NYC contains 32 elementary districts, each of about 15-30K students. We have plotted % subsidized lunch vs percent passing the grade 3 reading Similarly, San Diego City Schools have 118 elementary schools of varying poverty levels and with varying numbers of limited English students. We conceptually plotted scores as a function of % subsidized lunch and % LEP. Again, there is a very high correlation (accounting for up to about 75% of the variance in scores) based on just these two parameters. Since SD is a large, heterogeneous urban district, we can then use the SD equations to predict how any other urban or near urban suburban school would do based just on SES and LEP. This approach shows that many of our so called best or better schools are really only just going up to the level predicted from the background of their students, i.e. they are not really doing better in the classroom, they just do an ordinary job with a set of students likely to score well. On the other hand, a similar analysis points out just how well a school like Kelso in Inglewood is doing. Just about every score in every grade exceeds the prediction by a stastistically significant amount. No, the scores are not as good as in the rich suburbs, but they are close. That is a school that is really working. Please note, none of the above excuses schools from doing less well with low SES students, but it at least allows us to get a shot at knowing what schools are really doing well from a single data set. If you really want to make a point with the data, challenge the state ed people to release all the data, including SES data, in an easily usable format. In CA we get the scores for each school pretty easily, but we needed to go school by school from a different data source to get the official lunch and LEP data. This means that we can't do our analysis for all districts or all schools in the state. OF course, this is not what the educrats really want, but full release, without any hoops, is the fastest Long term we need both information about success relative to similar schools and success relative to what happened in the past. The best way to do the data in the long run is to follow each and every student through time (NB, this need not be linked to his name). This means that we can now go ask how individual schools do for all their students by asking about the range of progress among students. Mike -------------------------------------- Date: 10/12/98 10:59 AM To: Mike McKeown From: arthurhu@halcyon.com \doc\web\98\07\washtrans.txt To: arthurhu@halcyon.com, MBSOLOMON@aol.com From: "Lynn M Stuter" Subject: Re: Is there data to show washington reform gains? In-Reply-To: <19981005191802.00a33784.in@mail.icehouse.net> References: Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Status: O Marty -- Washington's transformation to education reform (OBE) began in 1988 with the Schools for the 21st Century project -- 33 projects, 111 schools, 52,106 children. The project sunset in 1994. Washington passed its education reform law, ESHB 1209 in 1993, so the 21st Century Schools continued their transformation under ESHB 1209. In 1987 Marc Tucker of NCEE spoke before our House Education Committee, undoubtedly in support of the Schools for the 21st Century. At that time, he was a consultant to our Governor, Booth Gardner. We have no doubt that the 21st Century School legislation was authored by Tucker. In 1997, Washington state