(Guest post by Greg Forster)
Since we’re so deep into the subject of value-added testing and the political pressures surrounding it, I thought I’d point out this recently published study tracking two and a half million students from a major urban district all the way to adulthood. (HT Whitney Tilson)
They compare teacher-specific value added on math and English scores with eventual life outcomes, and apply tests to determine whether the results are biased either by student sorting on observable variables (the life outcomes of their parents, obtained from the same life-outcome data) or unobserved variables (they use teacher switches to create a quasi-experimental approach).
Finding?
Students assigned to high-VA teachers [i.e. teachers who produce high “value added” on test scores] are more likely to attend college, attend higher- ranked colleges, earn higher salaries, live in higher SES neighborhoods, and save more for retirement. They are also less likely to have children as teenagers. Teachers have large impacts in all grades from 4 to 8.
Let’s bring that down to reality:
Replacing a teacher whose VA is in the bottom 5% with an average teacher would increase students’ lifetime income by more than $250,000 for the average classroom in our sample.
But here’s what I want to pay the most attention to. Note the careful wording of the conclusion:
We conclude that good teachers create substantial economic value and that test score impacts are helpful in identifying such teachers.
Note what they don’t say. They don’t say that increasing math and English test scores by itself leads to improved life outcomes. They say good teachers lead to improved life outcomes, and value-add is one relatively good way to identify good teachers.
You’ve heard the saying that the map is not the territory? (If not, that means you haven’t seen Ronin, in which case shame on you.) Well, it’s true. What raises life outcomes is good teaching, and good teaching can’t be reduced to test scores. (See here, here, here, here, here, here, here and here.)
But if you want to find your way around the territory, you need a map. If you want to help those kids stuck with lousy teachers who are out a quarter million, you’re going to need a tool that identifies them. Value added analysis is the best tool we’ve come up with yet – other than parental choice, of course.
And where the tests are freely selected and voluntarily adopted by schools, the tests provide helpful data for parents, so parent choice is strengthened by voluntary testing. That’s why over 90% of private schools use testing in some form. On the other hand, forcing teachers to use a test they don’t believe in is a self-defeating proposal.
But how do you get schools to want to use a test? Parent choice, of course! Choice is what creates the external standard of performance that makes assessment tools seem legitimate rather than illegitimate. So testing and choice are like chocholate and peanut butter – they’re two great tastes that taste great together.
Jay P. Greene's Blog 
Right on , brother Greg.
RE: Replacing a teacher whose VA is in the bottom 5% with an average teacher would increase students’ lifetime income by more than $250,000 for the average classroom in our sample”
DO you know that the best teachers are routinely put in the honors classroom and better performing classrooms for a variety of reasons. (I’m guessing you’re answer is: “No.”
So, the correlation between a high VA and a low VA is predetermined as school administrators make sure that those with the greatest potential will get the best chance to succeed. AHHhhhhhhaaaaaaaaaaa!
PLEASE know all the facts around the situation befor you jump to conclusions. I know jumping to these conclusions make you feel better, but it will not improve schools. The Liberal’s self-esteem movement did the same kind of data non-analysis. PLEASE, lets NOT make the same mistakes as those dolts!
The are Republican answers to schools REAL problems. But if we make policy based on your faulty analysis will delay the time our important reform will be implemented
A Republican Teacher.
Talk about spinning results. The data on 2.5 million students in this study was old, from the 1990s long before the NCLB and the conversion of classrooms into centers of test -preparation and testing, and long before the economic downturn.
The whole point of the statistical analysis here is to control for precisely the variables “Joe” is pointing to here, so this isn’t really a constructive engagement with the study. Do you find their statistical method unconvincing?
Sandra’s point only shows that this study does not vindicate NCLB. Good thing I’m not a fan of NCLB! Otherwise her point might be a problem for me.