Wednesday, November 12, 2014

Depression at Blackstone Valley Prep

I've got a story in the new issue of Common Ground on student survey data on depression and suicide at Blackstone Valley middle schools, focusing on Blackstone Valley Prep. The centerpiece is this table of 2013 SurveyWorks! data from RIDE: blackstone-valley

Read the whole thing, but here are some additional thoughts on the piece:

  • This was a lot tougher to write than the standard test score snark. Much more sensitive, and frankly, greater opportunity to look like an ignorant jerk if you get it wrong.
  • The data set is weird. First, there is no way to determine the validity (i.e., whether the kids really did feel sad or hopeless for two weeks in the past year). The completion rate is very high for a survey and the question, so if this was a random sample the margin of error would be extremely low. But it is not a random sample, it may be very biased, and the bias may vary by site. A large percentage of non-responders may be depressed.
  • On the other hand, this may be the only school level depression data ever published for a "high-expectations, high-support" or "no excuses" charter school, so it is worth a look!
  • There have been anecdotal reports about stress and depression at "no excuses" charters. For example:

    Hello. My name is Katie Osgood and I am a teacher at a psychiatric hospital here in Chicago. I am here today as a concerned citizen and an educator.

    In my hospital, we are seeing a disturbing pattern among patients coming from the Noble St Charter School Network of schools. We’ve seen an alarming number of students being admitted to the hospital with depression, severe anxiety, and increasingly with actual suicide attempts all directly tied to these schools’ discipline, academic, and retention policies.

  • I just focused on the Blackstone Valley schools for four reasons:
    1. Most of the schools with the highest reported depression scores were there.
    2. The completion rates were relatively consistent and high across those schools (10% or more above most PPSD schools, for example).
    3. It is treated as a discrete market for school choice.
    4. The number of schools is small enough that you don't have to rely on what would be extremely complex statistical analysis (you'd have to try to correct for participation rate and selection bias at each school) to make sense of the entire state data set. With 10 schools, you can just look at all the numbers and draw your own conclusions.
  • If having non-experts look at all the data and draw their own conclusions is not sufficient, then data-driven parental and student choice can't work.
  • This data has been consistent over the past three years. It isn't an anomaly. I actually sat on this for over six months waiting for the 2013 data to come out.

Ultimately, all the caveats about this data only apply to comparison. The survey data about student reports of depression and suicidal thoughts among students at BVP (and Segue, at least) is clear, consistent, complete and disturbing. We don't know why -- what collection of out of school factors, in school factors, and selection bias among students choosing the school -- but the fundamental issue cannot be dismissed without explanation, especially if the schools in question are considered models to be emulated and expanded.

No comments: