correlations

I know that having lively discussions on the quality of education, the size of classrooms, and the pay of teachers is not unique to where I live. Yet, there are some enigmatic qualities about this city, within this state. A railroad town to its core, we can throw a stone and hit the golden spike that united this country with a rail line nearly a century and a half ago. The parallel lines of steel brought commerce, materials, and people. World War II veterans passed through here after tours of duty in the Pacific, connecting with routes back home, out East; and the other World War II veterans, Japanese-Americans deposited into internment camps, came through as well. The stories of the railroad and our local downtown are famous. It made this place a hub, and it once made it a place where prostitution, opium, and good times and vices could be found, hidden in dark corners, underground tunnels, and private upstairs retreats. We still have good times though I can’t comment on the sex and drugs. We have a good sushi bar and brewpub, and you can get fresh baked bread and paint your own pottery.

The railroad isn’t what it used to be, in spite of our rich history and the sweeping yard that our streets still need to negotiate space with. Yet, we still have many who pass through and many who stay. I am, myself, one of those who stumbled into this place because there was work, mountains, and the historic downtown with the sushi bar and brewpub. Others have come here for other details, but basically the same essential reasons. We live here because we can. And, our children, for the most part, all go to the schools that have stood decades. Strangely, most all that separates different groups of people is property values and median incomes. Take a look at census data and you realize that a walk through the neighborhood, in the downhill direction, will not only scale you down in elevation but in socioeconomic status. Within the span of about a mile or so, median income changes by a factor of three. Not surprisingly, education background, property values, and other statistics that whitewash the value of a person change significantly through the same walk down the street.

Recently, a local news report looked at elementary schools statewide, and collected test score data in language arts and math in the second, fourth, and sixth grades for all. Then, in an effort to create a story, they ranked the schools, based on a kind of averaging of these scores for each school. My interesting old town with the rich diversity and many different populations living within had 5 of the 10 “worst performing” schools, based on these test scores, in our entire state.  There are lots of explanations, and lots of ways that those explanations get twisted by reporters and district officials. The superintendent said that we have “unique demographics,” and that could mean something about parents’ first language, economics, or, implicitly and too often a euphemism and a way to blame parents or even the kids themselves. That’s when my hackles start to go up.

Suppressing my hackles, I started to wonder more about these test scores. While I don’t like them entirely and I don’t think they’re measuring the same things that I’m going to value in someone’s education (there’s a whole other essay), they’re something we have to pay attention to, because it’s what the school is sending to the state and gets reported on the news. But there are lots of other things to measure as well, and the National Center for Educational Statistics (NCES) gives us a wide variety of these to consider, looking not only at test scores, but also at languages spoken, number of students on free and reduced lunches, number of books in the home of students, etc. These data paint a more complete picture. One other source of data that is fun to work with is census data, easily explored with interactive maps. These actually plot census tracts in different colors to show racial distributions, education levels, income levels, etc.; and actual numbers for each of these tracts can be pulled out of these maps. It’s all fascinating on its own, and it’s exactly the tool that made me realize the disparity in the economies of families as I walk just a matter of blocks from my home.

So, I just started playing around with these tools, and I did the following. In spite of the fact that I knew that any really accurate portrayal of what was going on was complicated and multivariate, I decided to just look at test scores in schools — in particular, the percentages of students, as reported on the news, who met the standards of the test at each school — and the median income of homes in the same census tract as the school. This falls short in lots of ways. First, it reeks of the self-assuredness of a physicist that tends to oversimplify any problem. Just arm one of us with some numbers and a spreadsheet, and we’ll explain your entire discipline for you and wonder why you even bother having your own academic journal. Second, there are some funny approximations here, including the fact that a school pulls from more than just one census tract, but I was trying to simplify things (and make it easier). Third, even if there is some pattern that emerges, there are so many other root causes for the pattern, you should be considering way more variables at the same time. But, I like a clean, clear graph, just to get a first glance at what I’m facing. Like this one:

Those pass rates, on the vertical axis, are percentages; and median incomes for a census tract in which a school is located is the horizontal axis. To me, a few glaring things pop out. Look closely at the lower left corner. There are two dots there, each representing a school in a spot where median incomes are lower than $20,000 per household. (These kinds of conditions are revealed in other ways, too. For example, the NCES data shows one of those schools as having three (yes, 3) students who do not qualify for free or reduced lunches. And, only two (2) students qualify for reduced lunch. 736 of the 741 students qualify for a free lunch.) Emphasizing the severity of this statistic is the fact that the upper right corner has a datapoint for a school in an area where median income approaches $90,000 per year for a household. And then there’s everything in between.

To a physicist, what’s really compelling about this picture is how well the dots all line up. Ignoring all the other factors playing into this, you could look at this graph and basically be able to predict how likely a kid is to do well in school based solely on how much their family earns in a given year. It makes some sense. Other statistics that we look at that seem to influence success in school include number of books in the home and education background of parents, and these would clearly tend to vary with income (the former a potential result of income, and the latter a potential cause of income). Moreover, others have pointed out to me that kids on the lower end of this graph are the kids showing up to school hungry and worrying about whether or not they have a warm coat in January. There are more things to attend to than studying multiplication tables.

Those data were amalgamations of entire schools and multiple subjects. As we look at just the sixth grade and just at one set of scores, the results look even more linear:

In second grade, the trend is still there, but the scatter is wider. What does this mean? I think it shows a cumulative effect: If you’re set up to succeed in one year of your life, you’re likely set to succeed the next year; but if you’re not… well, then things just keep getting worse.

I’ve been thinking about these statistics a lot lately because teachers have been told that they’re going to being evaluated based largely on these test scores, and their potential pay raises will be solely based on these evaluations. These kinds of data suggest that this just can’t be fair. I know what school I’d want to teach in if salary and test scores were my only criteria. But, more importantly, there’s a cycle here that seems to be at the heart of these correlations. As we’re creating educational settings, our hope shouldn’t be just to help the kids who are already clearly being helped by their circumstances. Education, at its heart, should be to level the field, raise all the boats, provide some justice and opportunity. Yet, it turns out that opportunity begets more opportunity. We need to work harder to make sure that all these kids have that opportunity. It’s important to these neighborhoods and our community, but most of all it’s important to the life of that one kid.

Advertisements

8 thoughts on “correlations

  1. First, I’m really enjoying the sudden increase in posting.

    Second, I have too much to say about this to start saying anything. I’ll limit myself to saying that the data doesn’t surprise me. It makes me heave a huge sigh. What a mess, what a shame.

    1. Here’s something that’s on my shelf, yet I still have yet to read:
      http://amzn.com/0807749621
      However, I’m confident it’s worth reading because Darling-Hammond (through other work and an address I heard her give) made me think a lot harder about these kinds of issues. Her work made me realize simple things like the need for kids to have food, health, etc. in order to succeed in school.

      Glad you’re reading! It’s fun and helpful for me to be writing more again.

      1. That book looks intriguing, and that you know and recommend the author’s other work is helpful. There are so many books on American education (and ed in general) out there that it feels impossible to know what’s worth reading. I’m gonna inter-library loan this one.

        Just read your post on running. Let’s do a destination half-marathon together! I need some motivation to do that distance–serious motivation–and I think a rendez-vous with your family would provide just the kick in the pants I need.

      2. Item available from Reed’s library! Love working in an academic setting. Will grab the book later today when I need a break.

  2. A destination for a 13 mile run + book discussion(s) would be fantastic. Our current plan is to do the Half that’s here in May (beautiful course — we ran it as a marathon relay this year), but “plan” is a very loose description. Open to lots of potential destinations. Maybe UT in May and Portland in October?

    1. I like that “plan.” I think we should work on making it a Plan, capital P. By “work on” I mean that we should just say it’s a Plan and act accordingly. That’s all that we need to do, right?

      May seems like a long time to wait to see you guys. Will you pass through Portland this summer?

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s