Monday, July 26, 2010

The Perils of Alignment

Dan Willingham:

A tall order, but without such a measure, we can predict that in about five years NAEP scores will not have moved much, and the mutterings will begin. In another five years we’ll have bounced on to the next fix.

I think there is good reason to think NAEP scores will go up, because the Common Core assessments are likely to be more closely aligned with NAEP, and thus test prep for Common Core will be more directly relevant to preparation for NAEP.

The more alignment there is, the less data. If graduation is based on achievement on Common Core assessments, college admissions and the difficulty of entry-level college courses aligned to CCRS, NAEP and Common Core assessments in alignment, then test scores, grad rates, and college enrollment and success rates should converge, which will lead to fewer checkpoints on the real preparedness of students, not more.


Claus von Zastrow said...

Perhaps. Champions of NAEP see its value as an audit test, so lack of alignment can uphold that goal. But NAEP also is too broad to be easily gamed (or so the story goes), so it could retain much of its value as an audit tool.

Also, if NAEP scores go up, state test scores go up AND college success rates go up, might that not validate NAEP and the state tests? My big worry would be that college success rates would go nowhere even after state test scores climb.

Tom Hoffman said...

Then the question becomes, who's wrong, all the tests and assessments, or the colleges?

Jason said...

Alignment is a good thing, if that alignment is to our ultimate goal for students.

If colleges are happier to get students with the set of knowledge in the Common Core and on the NAEP (i.e. these students are more likely to be successful than those taught a different set of material), then this alignment is good.

If this doesn't prepare students any better to handle college level courses or be competitive in the workplace, then it's not alignment that went wrong but what we're aligned too.

It's true that you can "lose" some external validation data, but that matters much more when the issue is misalignment between curriculum, assessment, and desired outcomes.