October 31, 2007
It's fairly widely agreed nowadays that schools should be judged by, and accountable for, their results, not just their intentions, services, or inputs.
NCLB version 1.0 embraced that view and determined that students' reading and math "proficiency," judged against absolute state-set standards, would be the way to hold schools (and districts) to account for federal purposes. Meanwhile, a lot of states had already developed their own systems that differed in various ways from each other and from NCLB. Some, such as Ohio's, considered more subjects. Some, such as Florida's, considered performance gains as well as absolute performance. Several set more than one standard, as NAEP had done with its well-known achievement levels labeled "basic," "proficient," and "advanced." (Strictly speaking, NCLB requires states to set three performance levels, NAEP-like. But only one of them counts.)
Much of the grumping, kvetching, and pushing-back against NCLB has to do with its simple, narrow conception of school results: just two subjects and one (absolute) level of performance in those subjects. The major complaints take four forms:
- Education is more than reading and math skills and excessive emphasis on those and only those leads to too narrow and mechanistic a school curriculum. (I've said that on a few occasions myself.)
- Schools are about adding to their pupils' learning, not just determining whether kids are already at or above some fixed level of performance; hence performance should be judged by achievement growth (i.e., "value-added"), not rigid cut-scores. Else we may falsely deem schools (and teachers) effective that happen to be blessed with fortunate, high-achieving kids, and we may falsely deem others ineffective because their pupils, despite making gains, maybe even big gains, are still below the bar.
- A single proficiency bar is a mistake, too, because (a) it will tend to be set low, (b) it does nothing to boost achievement for kids who are already above it and (c) it may also lead to neglecting students who are far below it.
- Schools do more than get kids ready for tests, and education is properly judged by more than objective test scores. What about creativity? Character development? Physical and social well-being? And lots more, plus other indicators of school performance such as attendance and completion rates.
Each of those oft-heard laments has elicited familiar rejoinders in the NCLB context, along the lines of "Washington can't micro-manage everything about K-12 education and these other matters are the proper responsibility of state and local folks"; "if kids lack basic reading and math skills, they won't be able to learn much else so let's put first things first"; "it's difficult to assess (fairly, efficiently, economically) the kinds of things that don't lend themselves to standardized testing, and if we try to do it we'll provoke even more arguments"; and "value-added is a swell concept but it's really hard to do right and the available data may or may not sustain it." (Note that "growth models" are even more suspect when a state's tests are not properly calibrated from grade to grade, as we discovered in a recent Fordham study.)
Now that Congress is finally beginning to focus seriously on NCLB 2.0, proposals are flying in from all sides to address these accountability-related issues. So far, though, there's no sign of consensus on the two big outstanding issues:
- What, besides reading and math test scores, should "count"--either as option or as requirement--when determining school accountability, and how should these other things be measured? Other subjects? (Objective tests? Classroom grades? Portfolios?) Graduation rates? (Which of a thousand ways of calculating completions and drop-outs and tracking or not tracking kids who move from school to school?) Teacher credentials? Teacher "effectiveness"? Professional development? And more.
- Within the sphere of testable reading and math skills, is there, in fact, a suitable "growth model" (or maybe several) that isn't ridiculously complicated; can be sustained with reliable data; allows for school-to-school and district-to-district comparisons; and doesn't wind up exonerating schools and systems that may well be showing gains but whose achievement gaps remain wide and whose pupils are still achieving far below what's needed for success in college and the workforce?
There could soon be blood on the floor. It's already evident, for example, that House education chairman George Miller and ranking committee member Buck McKeon (and the Bush administration) don't agree with one another about adding more subjects to the accountability cocktail. The latter two depict such moves as "watering down" what should remain a stiff drink--and they fear, I think with reason, that Chairman Miller may yield to relentless pressure from his liberal base to slide too far down the slippery slope of "portfolio" assessment, teacher judgments, etc.
I don't see any easy resolution. But in case you were wondering, here's my proposal for these parts of NCLB 2.0. It has four elements:
- Somewhere outside Washington, set national standards in reading and math, preferably based on some version of the American Diploma Project, vertically aligned from the early grades to the end of high school, and linked to college and workforce readiness upon completion of high school. Let states adopt these as their own NCLB standards if so inclined.
- Develop national tests aligned to the national standards, with at least three cut scores for each test in each grade. States may use these tests and cut scores if they like--and if they do, Uncle Sam will foot the bill. If they don't, they must show the Secretary of Education how their own standards, tests, and cut scores are equivalent.
- Tell states that at least 50 percent of a school's (and school system's) "adequate yearly progress" rating must depend on students' performance against fixed "proficiency" standards in reading and math, much like 1.0, except that there'd be a "national" option--and we'd also know how kids and schools are doing in relation to "basic" and "advanced."
- The other half of the accountability formula is up to the states. If they like, they can use what I sketched in the previous paragraph for the whole thing. Or they can hinge up to 50 percent of a school's rating on (take your choice, mix and match): achievement in other subjects, assessed however they like; student growth (including, say, how many kids moved from "basic" to "proficient"; or other school-linked or policy-linked factors that can be calculated at the school level.
You say you don't like the 50-50 split. Okay, how do you feel about 60-40? 70-30? Personally, I'd hate to see the fixed-score reading/math part go to less than half. You may disagree. I'd be happy to hear from you.