Monday, July 16, 2012

Calculating Gap-Closing on a Curve

So 25% of RIDE's new school rating & classification is based on "gap closing." Unfortunately, Rhode Island schools and districts are largely segregated, so this does not actually make much sense. In most PPSD schools, there aren't enough white middle class kids to form a comparison group to establish the "gap." To a lesser extent the opposite can be true for poor and minority students in the suburbs.

So, to deal with this, RIDE calculates a score for a hypothetical Performance Reference Group of non-poor, non-special ed, English speaking whites. This is adjusted according to district types as shown:

So built into the model is the assumption that, for example, the default cadre urban middle class white high school student inherently underperforms their suburban cousins' proficiency by 27 points in reading and math? Huh? What, what, why?

Let's say we have a hypothetical urban high school with 100% poor minority students, 44% proficiency in reading and 5% proficiency in math.

In terms of absolute proficiency, they get 1 point in reading and 1 in math:

In terms of gap-closing, they get four points in reading and four in math. In fact, it seems to be impossible for an urban high school to get less than four points in math since the white man's score is 19 and you get 4 points for a gap less than 20:

As a result you get examples like Central Falls High School, which in absolute proficiency is 6/35, but in "gap closing" gets 28.50/30. In fact, that is 28.5 out of only 50 points the school gets total in the rating system. A suburban high school with poor and minority students getting the same reading and math scores as the urban school above would have a gap-closing score of 2 (compared to 8).

So... who cares? It is horseshit layered on top of horseshit. The overall system is biased against high poverty schools from stem to stern, so I don't really care. What RIDE gets out of this is a system that looks a little fairer to urban schools. One suspects that in the end no suburban schools will be harmed in the production of this film.

Nonetheless, that's how it seems to work, if you were wondering. If I've parsed this wrong, let me know.

No comments: