Friday, March 11, 2011

Content, Standards, and Value-Added

In sorting through the ongoing discussion of "content" and curriculum vs. the Common Core standards, it is important to think back about what the Common Core standards are for. The reform agenda of the Gates Foundation, other business-model reformers, and the federal government as expressed in Race to the Top, is based on a belief in the ultimate efficacy of value-added analysis in education. The theory of change here is that once we've got accurate measurement of how not only teachers, but every other variable, including curriculum design, influences student growth, we'll have an efficient market for educational improvement and a path to glory. But first we need better value-added data.

  • To get better value added data, we need a new test which will be given to all students and will specifically produce data amenable to value added analysis.
  • To make this test, we need a standards document which specifies what students should be able to know and do at each level, and will produce usable data for value added analysis.

What you want from value-added analysis is to be able to say, "this fourth grade child started the year in English at grade level 3.5 and ended at level 5.1." And then you want to be able to extend that analysis to the class, school, district, state, etc. The problem with this is that it only makes sense when you can define learning as a linear progression.

The design of the Common Core ELA standards is such that you should be able to produce interesting, finely grained data. For example, you should be able to say, "On an informational text of complexity 3.5, Johnnie scores a '4' on standard #4. However, on a complexity 4.5 text, he scores a '2' on the same task if the text complexity is raised to 4.3. On a science text of complexity 4.5, he scores a '3' on standard #4, but in a history text of the same complexity, he gets a '4.' Thus..." Obviously there will be a lot of graphs involved in real life.

But that only makes sense if the learning path is linear, with each step dependent on the previous. That is, you can't read at a fourth grade level without at least conceptually passing through reading at the third grade level at some point. Math doesn't fit in as neatly, but it is on the whole the most linear of all core subjects (anyone who has tried to design a progressive high school knows this).

On the other hand, learning about 20th century American literature does not depend on knowing 19th century literature. If you walk into a post-Civil War US History class having skipped pre-Civil War US History, you can still pass without learning the pre-Civil War history. But what's the value-add then? For a year's worth of content, the maximum value-add is 1 year. The best 19th century literature class isn't the one that rushes into the middle of the 20th century.

This is not to say that you can't measure the quality of learning and/or instruction in a content-based classs. It just doesn't provide the kind of data that is wanted or needed here. There isn't a wide enough spread, and the comparisons are less consistent, and the whole thing makes less sense.

So that's why content is permanently a second class citizen in the world of Common Core and Race to the Top, despite whatever other assurances, side comments or manifestos are produced.

No comments: