Recently, I examined teacher compensation in six urban school systems. Among the six I analyzed was Denver — a system that uses a strategic compensation approach known as ProComp and has done so since 2005. In fact, voters in Denver passed a $25 million tax increase just to fund the ProComp experiment.
According to A+ Denver:
ProComp, Denver’s teacher compensation system, was collaboratively designed by DPS and DCTA. It was developed to incentivize and reward teachers who set and reach high learning expectations for students, creating a differentiated compensation structure that moved beyond teacher tenure and level of education by providing a pay-for-performance incentive system.
And Denver Public Schools says of the program:
In 2005, Denver voters approved funding for our performance-based teacher compensation system. Since then, ProComp has been setting a new national standard for rewarding and recognizing teacher excellence.
ProComp is particularly distinctive because DPS and the Denver Classroom Teachers Association (DCTA) collaborated to develop, build, and test the program. ProComp’s design is also unique, interlinking these central ingredients for impacting student achievement:
- Best practices in teaching and learning
- Tools and data for measuring student growth
- Knowledgeable and motivated professional staff
- Evaluation of teaching practices
ProComp seems to be structured to reward teaching excellence as measured by student performance and is designed to make teaching in Denver a more attractive proposition. Is it working? Some points to consider:
Lifetime earnings of Denver teachers are relatively low
In the National Council of Teacher Quality’s ranking of teacher compensation (adjusted for cost of living), Denver Public Schools ranks 100th out of 124 districts in lifetime earnings for teachers. Teachers in Denver Public Schools receive $1,574,436 lifetime earnings. This is compared to the first ranked district, Pittsburg Public Schools, with lifetime earnings of $2,738,691 and the second ranked district, DC Public Schools with lifetime earnings of $2,641,592. In terms of other districts in Colorado, Jefferson County is 37th out of 124 districts with lifetime earnings of $1,966,895.
So, in terms of changing the value proposition for Denver teachers, the ProComp system hasn’t been an improvement.
In fact, this analysis suggests that the starting salary for Denver teachers is relatively low and that while teachers can move to a top “step” rather quickly, even with incentives, the earnings don’t keep up with other urban counterparts.
Does ProComp promote excellent teaching?
This study looks at the impact of ProComp on teacher retention and student achievement:
The achievement of students assigned to a teacher participating in the ProComp system, however, is significantly higher (by 6 to 7 percent of a standard deviation) than that of non-participating teachers, so much so that the net effect of having a ProComp participating teacher is estimated to be positive relative to years prior to the implementation of ProComp.
The net effect achievement level in reading for students in ProComp years is also positive, but here the ProComp effect is driven by non-participant teachers.
So, wait, the idea is that teachers who opt-in or otherwise participate in ProComp are already or will become stronger teachers as measured by student growth than those who opt-out. Here, however, you see mixed results. In math, ProComp teachers performed better on the student growth measure. But in reading, non-ProComp teachers showed better results.
That’s right: Teachers who didn’t participate in the incentivized compensation system showed better student outcomes than those who did when it came to reading.
Could it be that compensation structures do not significantly impact teacher performance as measured by student outcomes?
The study goes further and the results get murkier:
Finally, we turn our attention to the specifications that include indicators for a teacher’s voluntary status in ProComp (column 3 for math and 6 for reading). These results show no clear pattern. In middle school math, ProComp participants are more effective than non-participants, with compulsory ProComp teachers performing slightly better than voluntary ProComp teachers. In middle school reading, non-ProComp teachers are more effective and there is no statistically significant difference between voluntary and compulsory ProComp teachers. At the high school level, compulsory ProComp teachers are less effective in math than both voluntary ProComp and non-ProComp teachers. In reading, compulsory and voluntary ProComp teachers perform better than non-ProComp teachers.
There’s not a clear pattern in terms of teacher performance relative to ProComp status. The pay system is not showing a consistent, significantly positive impact on student achievement. It could be that other factors impact student performance. Or, that teachers work at their maximum potential regardless of the presence of incentivized pay scales (student achievement increased over the time period of the study, though that could not be attributed to ProComp alone).
Another study noted:
University of Colorado’s Denver ProComp Evaluation Report (2010-2012) finds little impact of ProComp on student achievement, or on teachers’ professional practices, including their teaching practices or retention. The study did find that ProComp generally impacted teachers’ salaries positively but minimally, with the average teacher earning $456 more than they would on a traditional salary schedule. That said, some teachers experienced large variation, earning either $6,000 more or less than they would have on a traditional salary schedule.
For most teachers, the pay system after ProComp did not have a significant impact on their salary. And, the presence of incentives had little discernible impact on either teacher behavior or student achievement.
The goals of ProComp were to make teaching in Denver more attractive and to reward excellent teaching. Even with ProComp, teaching in Denver is no more financially attractive than teaching in similar urban districts. Additionally, two different studies have suggested mixed results (or insignificant ones) in terms of student achievement.
I’m not sure this is a system meeting stated goals, but it doesn’t seem to be doing any harm, either.
For more on education politics and policy, follow @TheAndySpears