This is an old article, but important enough that I'm still going to write on it. Basically, it goes into the numbers and debunks the myth that teachers' unions lead to better schools. More interestingly, it uses statistics and facts usually hailed by liberal and progressive groups to do so.
In short, liberals are claiming that unionized teachers create better schools. The argument debunks this, pointing out that, in fact, what makes a lot of the model states for this argument either rank well or poorly on education is actually socioeconomic makeup.
Again, to me, the most interesting part of this is that this is a typically liberal argument about education and equality, particularly racial equality. Statistically speaking, poor and minority students don't do as well, and their school districts don't rank as well. The states at the bottom end of the achievement scale have largely poor or minority students, while those at the top have largely middle and upper class white students.
The blogger actually goes a step further and breaks down test scores within each state by race. In fact, many of the non-union states' white students do quite well. It is the effect of combining these scores with the larger minority populations which make the overall state averages lower.
As the author points out, there are several reasons minority students don't do as well. They tend to come from less wealthy, and less educated families, who have survived without the benefit of a college education. They are less likely, therefore, to think that a college degree, and the success it requires in high school to get one, are entirely necessary to survive in today's economy. They also likely come from communities without a lot of well-educated role models. They are less likely to know that if they do well in high school, they will be able to afford to go to college, or to a top tier college.
Many minority groups (with the exclusion of black and Native American groups), also face the increased challenge of many of their members not living in English as a first language households. Not being raised in an English-speaking home can negatively affect performance both in school and on standardized tests.
Finally, there has long been an argument that standardized tests themselves contain a racial bias. While this may be hard to believe, it is quite plausible. Much of these tests revolves around reading comprehension, mathematical word problems, etc. While these are noble attempts to allow students to apply their knowledge in a real world setting, for those who grew up in less affluent areas, or in an ethnic enclave, these backgrounds can be anything but "real world."
Let's not get started on the issue of school funding. The schools most in need of help often receive the least funding, as they are in poorer neighborhoods.
All of these issues add up to explain why minorities don't do as well on standardized tests. This explains far more about nationwide education rankings than the presence or absence of teachers' unions. Instead of trying to fight for or against teachers' unions to improve student performance, perhaps we should focus on making real changes in our education system. Offering school vouchers to allow students and parents to choose the best school for their child. Improving inner city and minority-populated schools. Rewriting standardized tests to be a fair representation of a student's knowledge and intelligence, rather than biased toward those with a fortunate upbringing. Allowing non-traditional teaching methods into the classroom. Rewarding teachers for their performance, not their length of service. All of these are issues which will close the achievement gap, and raise the bottom-ranked states up the national education ladder. Coincidentally, many of these are ideas which teachers' unions oppose.