Guest Editorial: What to make of the power of SES

See all posts

By Dick Startz, University of Washington, Seattle

As Kate Walsh pens in the above, "SES and family background swamp school effects by a 3:1 margin. In spite of these odds, we can still have daily reminders that teachers matter plenty."

Both statements by Walsh are firmly backed by the scientific evidence.

Begin with the idea that parents (or families or communities) are the most important factor determining student outcomes. Well-intentioned people draw the logical-sounding conclusion that if background is the most important factor in determining student outcomes, then background is the key to getting better student outcomes. I see this coming from people who believe that parents drive outcomes (arguably true) and therefore teacher-centric solutions are relatively unimportant (hogwash)--unlike the NCTQ position.

When we look to see which 'input' is most responsible for explaining why some kids thrive in school and others fall by the wayside, we have to look both at how much the input varies across kids and how large an effect follows from that variation in inputs. The comes-to-school-with-the-kid (parents, community) input varies enormously, and the effect is large enough that comes-to-school-with-kid is the greatest explanatory factor for why some kids make it and others don't.

But if we want to change outcomes, we must ask a different question. Which input among those that have a big effect can we change? Changing the input provided by the average parent is a delightful thought, but it's not going to happen. In contrast, we know that teachers do have a huge effect and that teacher input is changeable.

Think of it this way. Suppose we observe outcomes for a group of kids. If we know about their parents' contribution to their education, we could probably do a pretty good job of ranking outcomes for the students. Once in a while, a really spectacular teacher will make enough of a difference to move her students above what you might have expected given the parents. But this won't happen too often.

Now suppose we reformed teacher quality by making every teacher better by an equal amount and then re-ranked the students. (Remember, we're oversimplifying.) Every student would have equally improved. So relative rankings would be unchanged. Post-reform, we'd still conclude that parents are the definitive factor.

But every student would have a better outcome. That, of course, is the goal of reform. In other words, parental contributions may be the explanation for differential student outcomes, but teacher contributions are the route to improve outcomes for all students.

Those who argue for the primacy of parents implicitly believe the parent term varies much more than the teacher term. That may be true, but it has nothing to do with where we should concentrate reform efforts. The right questions to ask are: 1) can reformers change teacher input (the answer is yes), and 2) is the multiplier 'teacher effectiveness' large (the answer is emphatically yes).

I've simplified and put the discussion in hypothetical terms. But the situation I've described is very much how the real world works. Dan Goldhaber and colleagues' recent study on teaching math provides an example based on the effectiveness of Washington State math teachers. The authors provide a pictorial breakdown of the importance of various explanatory factors for student learning of math.

You see immediately that student background swamps the role played by teachers. Heck, even "unexplained" swamps teacher effects.

The authors also calculate the size of the factor from the same data. They find that increasing teacher input quality so that the median post-reform teacher performs as well as the current 84th percentile teacher would improve student outcomes by the equivalent of 2.6 months of extra schooling. That's enormous! If we did this for teachers in all grades and the student effect cumulated, the average student would learn two and a half years more material than she does now!

Moral: Arguing over whether parents or teachers are more important is a distracting irrelevancy.

(An earlier version appeared at