GothamSchools — daily independent reporting on NYC public schools

I already heard someone argue that because the SAT scores were flat it was evidence of Klein’s “failure” without even acknowledging the fact that at least Klein has gotten more minority students to even take the test, which could also be a reason that scores are flat in NYC. Sheesh!

At the Manhattan School Governance Assembly hearing the head of School Construction, Greenberger (?) explained how the DOE employs demographers to predict space needs … based on census data … in my former district a staff member did careful analysis of census data by bloc tracts to predict space needs, the problem was that we could not predict immigration … census data does a poor job of counting immigrants … and, of course, kids have been siphoned off into charters … anecdotally Manhattan schools have had an influx of white kids … districts in distant Brooklyn (D 18/22) have decreasing number of white kids … an aging population, then again, will the bursting of the housing bubble result in more families buying in the outer boroughs … complicated but fascinating … it also impact elections, how has the voting population changed by race? Have u spoken w/ the Center for NYC Affairs (Milano) folk? They are heavily invested in these issues.

I’m interested in hearing more about this but can offer some anecdotal information. At my school (a new-ish, small, non-selective school in Brooklyn with almost 100% black population) everyone took the SAT–a really wide range of students, including special education students. I was shocked at the low scores of some of my best students–students who worked hard, got excellent grades, and practiced. The valedictorian got terrible scores and he’s now doing well at an Ivy-league institution. So I’d want to make two points: 1. I do think that with more students taking the test, scores are going to drop (so the fact that they are flat is actually good), and 2. low scores are not indicative, in my experience, of students’ inability to do well in rigorous academic situations.

Analyzing short-term trends in SAT or ACT scores, and attempting to draw inferences about the quality of secondary schools, is often misleading. There are a number of measurement reasons for this, and why caution is warranted when drawing such inferences. First, SAT test-takers are a highly self-selective group—and we know from years of research that drawing inferences from self-selected samples can be quite misleading (lots of political polling data provide evidence of this, and this summer’s news reports on various political “town hall meetings” give us a real flavor of this phenomenon). The second point is that the SAT and ACT are norm-referenced tests, i.e., they are designed to reproduce a Gaussian distribution (the familiar bell-curve), thus stablizing the mean of distribution, and maintaining a highly stable average amount of variation in those scores over time. The short story, the take-away message, is that the SAT and ACT scores are designed to be stable over time. These two reasons–highly self-selected samples and psychometric design principles–lead to trends in SAT and ACT scores that remain relatively flat over time.

In response to Jeff S.’s comment, the reasoning I referred to with respect to the SAT and ACT does not extend to the NY State accountability tests. The State’s accountability tests are (1) designed to measure a set of grade-level specific content and performance standards, and test items are selected for inclusion to reflect that purpose; and (2) the accountability tests are not administered to self-selected samples—they are designed to be administered to the population of students across a state at specific grade levels, e.g., all fourth grade mathematics students. Thus, the set of inferences that can be drawn from accountability tests are different, by design, from the kinds of inferences one can draw validly from, say, a college admissions test. Educational measurement experts, typically, are quick to point out that the initial purpose(s) of a testing program should dictate the design of the tests, and should also circumscribe the kinds of inferences one can draw from the information provided by the testing program.

all the excuses are sickening as to why black & hispanic kids don’t do as well please if you instill dicipline focus and hard work and a NO excuses and NO I am a victim mentality these kids are just like everyone else the slackers don’t do well and the hardworkers do well please stop with all the BS of anylising everyone needs to crack the whip and study act like civilized people in school and learn at any cost what you are being taught and if the techer is not good or is not able to teach then remove the teacher.everyone needs to take responsablity for their job including the students,their gardians and the teachers.all these tests are all dumb down already to make everyone feel good well what about all the Asian students that bearly speak english they score high on the test becasue they bust their butts studying their families demand sucess.