I started blogging regularly in the Fall of 2008. When I set up this website I put a reminder note in my calendar: on my fifth anniversary, I would step back and reflect on this work, and decide if it should change, or continue.
That reminder popped up last week, and time has come. I’m going to take a break from blogging at least through the end of 2013, and perhaps longer. Distance always creates perspective, and I’m interested to see if a sabbatical makes me feel like I should return with more vigor, or move on to other things.
I’m grateful to all of you for reading, and hope we intersect in different ways.
The Denver Post published an Op-Ed I wrote, but did not have space for the accompanying graph. Here it is:
I find most political commentary spurious – either so far in advance that no one remembers how inaccurate it was, or done only in hindsight where one is merely suggesting explanations for outcomes already known. Far more interesting are predictions just prior to events — particularly when there is not enough polling data to attract Nate Silver. So as in previous years, I am publishing my thoughts just as voting closes (and then slipping back into sabbatical).
There is also a facet of the 2013 elections that add to the general uncertainty. In Colorado, the election is by mail, which make previous results less relevant and predictive. I’m pretty sure something here will be horribly wrong. Your guess is as good as mine as to what it will be.
A66 continues to posit a question that we are unlikely to answer anytime soon: how much money should we spend on K-12 public education? Reasonable people disagree, often substantially. Per student, Colorado spends less than the national average, which is in turn far higher than any other developed nation. There is no clear answer on what amount is sufficient (in no small part because a correlation to quality is not evident), so will voters support an extra billion dollars each year for uncertain outcomes?
Denver Public Schools has released their 2013 School Performance Framework (SPF). The summary is mixed — the distribution of performance has stretched, so there are more kids in both the highest-quality and the lowest-quality schools. When aggregated, the gains are positive: more kids are in quality schools in Denver than at any point in the past five years.
However beyond the initial rays of light are two deeply troubling trends. First, the performance of the best district-operated schools continues to depend predominantly on the demographics of the kids they serve. Second, there appears to be no connection – particularly in the secondary grades — between the district’s efforts to improve its traditional schools and their subsequent performance.
At some point very soon, DPS will issue its 2013 School Performance Framework (SPF). The SPF performs a valuable function by aggregating multiple metrics on school performance and presenting the data so that both parents can understand it at a high level and with enough depth that educators can drill down into specific areas. Performance mechanisms are often difficult, and the positive news is that we are now having conversations in Denver about how to make good things like the SPF better (instead of questioning their basic premise). But it is also time to give the SPF a hard look and work to improve its credibility and usefulness.
Here are my top suggestions for improving the SPF. I hope others will weigh in as well.
1. Raise the Proficiency Bar: The low levels at which proficiency is rewarded on the SPF is shocking. Examine the 2012 SPF rubrics (Under “2. Student Achievement Level”).
High schools get maximum points for having 20% of their students at proficiency in math. That’s not a misprint – if one in five kids is doing math at grade level, that school receives all possible points in the category. Other subjects are not much better: maximum points are awarded for 30% proficiency in science, 40% in writing, and 50% in reading. Middle schools are similar: maximum points for proficiency levels of 30% (science), 40% (math and writing), and 50% (reading). Elementary schools are 30% (science), 40% (writing), and 50% (math and reading). Not one level or subject has a proficiency goal above 50%.
Let’s start with something on which everyone should agree: metrics are neither inherently good nor inherently evil. They are simply calculations — ticks on a stick that measure time, distance, speed, place, or other attributes. All measurements can be used well or poorly and most disagreements are not about the number, but about the use.
For many years education metrics were predominantly based on proficiency: how close is a student to a specific standard. However the misunderstanding and misuse of proficiency metrics have led to a number of consistent abuses: assigning relevance to trends over time using different cohorts (e.g. this year’s 6th grade vs. last year’s 6th grade) or discrete standards (3rd grade math compared to 10th grade math). In particular, many users of proficiency data commit the cardinal sin of comparing the performance of schools with vastly different demographics.
The new favored metric in education is growth percentiles (which measures change in an individual student or cohort over a year’s time). And the ability to measure student growth is a significant milestone and provides an incredible amount of value and insight. Growth percentiles are also adjusted to group students with similar academic histories, making comparisons far more valid. But, like proficiency before it, growth is often misunderstood and misapplied, and is routinely cited as evidence for conclusions that it does not support.
My criticism here is not the metric, it is how the metric is being used (and my comments are limited to the Colorado Growth Model, where “growth” as commonly used refers to median growth percentiles). Here are some of the primary issues with the use of growth percentiles in Colorado.
TCAP data was released last week, and the focus in Denver was an emphasis on median growth percentiles. Now I think academic growth is really important, and I’m on record as saying the Colorado Growth Model is one of the most significant developments in education in the state. But I also believe that our obsession with growth has gone too far.
Growth is a relative measure, so it is critical to understand not just the scores, but their distribution. For both anecdotal evidence and some data imply that it takes growth scores of at least one standard deviation from average (and possibly as much as 1.5 sd) before there is any real impact on proficiency. Allowing growth to obscure other metrics means that we are in danger of celebrating incremental progress in growth year after year without seeing any serious proficiency gains. And that does not serve students well.
For better or worse, life is set up to reward accomplishment more than improvement: nobody selects a surgeon or a lawyer based on how much better they are now than they were last year. Nor do colleges and organizations look at a student’s growth scores when making decisions about admissions or hiring. In truth, outside of school systems the designation of “most improved” is usually a blue ribbon, and as any kid can tell you, it’s a bit of a letdown. Growth is a valuable metric which we should use wisely, but it has to be viewed as a means to the desired ends of mastery and skill inherent in proficiency scores.
So how would I measure school performance in Denver? I believe the most meaningful measure for any school or system is exit-level proficiency (ELP). In Colorado, the best metric we have for ELP for high schools is 11th grade ACT scores; for middle schools it is 8th grade TCAP scores.