Graduation rates aren’t a very good measure on which to hold high schools accountable.
There are a couple reasons for this. First, graduation decisions are mostly left up to the schools and districts that are supposedly accountable for them. If you hold schools and districts accountable for their graduation rates, they have an incentive to just pass more students along and out the door.
Second, graduation rates also have a long time lag between when a student begins his or her high school career and when he or she finally finishes. Although schools and districts have quite a bit of control over these things, there’s still a long time lag, and it will be hard for a superintendent or principal to show any immediate improvements.
Both of these issues can be mitigated in various ways. States adopted high school graduation exams to require all students across the state to meet the same bar. But those exams come with their own problems. To avoid the time lag problem, states could adopt retention measures calculated on an annual basis. But few states have done so. Averaging across multiple years could also help, but few states do that either, and schools still can’t move the needle very quickly.
But worst of all, graduation rates don’t tell us very much about whether students are prepared for life after graduation.
The graph below comes from my new paper on high school accountability systems. The graph plots graduation rates against college-going rates for each high school in Tennessee. Each dot represents one school, and the solid diagonal line represents the relationship between the two variables. As the line suggests, there is a trend that schools with high graduation rates also have high college-going rates.
But the general trend misses a lot of variation. Schools with nearly identical graduation rates can differ by 50 or 60 points in terms of the percentage of graduates who go on to college (compare all the dots in the red circle).
If you construct an accountability system based purely off graduation rates, you’re going to get some perverse results. Using the same data on Tennessee’s high schools, I created lists of “low-performing schools” based on how the schools ranked on different measures of school quality. Of the schools with the lowest graduation rates, none were also in bottom tier on student growth, and only six percent were in the group of schools with low college-going rates.
Another way to look at this question is to take a step up the scale and look at states as the unit of measurement. To show what this looks like, I pulled college-going rates and high school completion rates for each state from the National Center for Education Statistics. The most recent state-level college-going rates come from 2009-10, so both variables use that year as the reference point.
The graph below shows the results. Each dot represents one state, and the solid line represents the trend, such as it is. As the flat trend line suggests, a state’s graduation rate tells us basically nothing about how many of its graduates go on to college.
Now wait, you may say, students have to graduate from high school before they can go to college. These two trends have to be linked.
Well, yes and no. It’s true that students have to make that linear progression, but it isn’t at all clear that this same linear pattern holds for schools or districts. That’s because some schools with low graduation rates may do a really good job for their high-achieving students, and all of that sub-population goes on to college. But the reverse of that can also be true. A school may just be bad across-the-board, bad on graduation rates and bad on college-going rates.
The same is true at “better” schools. They might be great at getting kids through to graduation but terrible at providing counseling or advice on how to go to college. Or they might not value higher education. Schools in rural areas tend to have lower college-going rates, suggesting geography and values play a role in these things too.
Either way, high school graduation rates just aren’t a great measure for accountability purposes. They are easy to count, and we now have a more standardized way to measure them. But if policymakers care about concepts like student growth or college-going, they can’t rely on graduation rates alone to measure them.