Is firing bad teachers really the key to fixing education?

Education historian/commentator Diane Ravitch points out that the states with non-union teachers (who thus have little or no job security) tend to have lower
academic achievement than the states with strong teachers' unions.

That should put to rest the myth that bad teachers with ironclad job security are the cause of the challenges facing public education.

As Ravitch adds, the state reported to have the consistently highest academic achievement is Massachusetts — a strong union state. (It’s also widely called "Taxachusetts" by the right — could there be a connection?)

Ravitch emphasizes that she's not necessarily saying that unionization and job
security lead to higher academic achievement, but the facts show that unionization and job security clearly don't work against higher academic achievement. They are correlated.

I thought it was worth looking for some data. But not officially being a statistician, I wasn’t really sure of the best measure of state-by-state academic achievement.

So I decided to look at one measure that interests me. That's the list of "cut scores" for National Merit semifinalists. National Merit recognition is based on
the PSAT scores of 11th-graders. The cut score for recognition varies from state
to state. That's explained this way on Wikipedia:

The minimum Selection Index for recognition as a Semifinalist is set by the NMSC [National Merit Scholarship Corporation] in each state at whatever score yields about the 99th percentile.

The organization FairTest has posted a list of the cut scores for the high school graduating class of 2010, which range from 201 (Wyoming) to 221 (Massachusetts, Maryland and New Jersey). California's is 218.

The National Right to Work Legal Foundation posts a list of Right-to-Work states (which don't allow workplaces to require union membership, meaning unions are toothless) and what the Foundation calls Force Unionism states. I took those lists, added each state’s Class of 2010 National Merit cut scores and averaged.

The results:
Right-to-Work states: average cut score 208.4545
Forced Unionism states: average cut score 213.6897

That result seems to show that unionized teachers correlate with higher academic achievement, and non-union teachers correlate with lower academic achievement.

If I’m missing confounding factors, I can’t see what they would be. It's true that not all 11th-graders take the PSAT, and the culture probably varies state by state as to whether taking the PSAT is more widely encouraged or less. But that wouldn’t seem to confound the basic finding.

By the way, the lowest-cut-score state — Wyoming at 201 — is a Right-to-Work state, and the three that are tied for highest — Massachusetts, Maryland and New Jersey — are strong labor states.

It seems conclusive: Teachers' job security — and, if you will, "forced unionism" — correlate with higher academic achievement.

— Caroline Grannan,
San Francisco public school parent and advocate

Is firing bad teachers really the key to fixing education?