Principal Accountability, Multiple Measures, and Campbell's Law
In his recent post on Reality Check, Walt Gardner rightly points out that accountability for teachers must be matched or exceeded by accountability for principals:
Unless principals are blatantly incompetent, they tend to remain in their jobs. Nevertheless, principals can be evaluated using multiple measures. The most obvious is the progress that their students make on standardized tests from year to year. They can also be assessed by their own faculty. Every year, United Teachers of Los Angeles uses principal surveys for this purpose. Parents can be included in these surveys. Finally, the ratings of those above principals in the chain of command should count.
To a point, I agree - behind every teacher who is underperforming is a principal who shares responsibility for the situation. Accountability starts with principals. Walt is also on track in suggesting that multiple measures are essential. (Walt is referring specifically to a situation of apparent principal misconduct, but his comments apply to performance in general.)
Whatever multiple measures are used, principal accountability requires that principals themselves be well-supervised, whether by education directors, assistant superintendents, or the superintendent. Only when such a relationship exists can the "multiple measures" be put in their proper context.
For example, a principal may have substantial work to do in improving teacher performance and evaluating out chronic low performers. This work is necessary, difficult, and universally unpopular. Ask any principal who has gone through the process of putting a teacher on probation - or implemented virtually any difficult but necessary change - and they will report a dip in their teacher survey scores. If the principal's supervisor knows that this is the situation, data such as teacher survey scores can be interpreted in an entirely different light.
Student test score data also needs to be interpreted in context; principals know how to drive up test scores by requiring test prep, catering to high-income parents, and encouraging harder-to-teach students to find another school. Only if a supervisor knows what's going on in a school can she or he judge whether the principal is raising scores with integrity and quality instruction, or with underhanded tactics.
Principals need to know they will be held accountable, but not subject to brain-dead accountability. When we forget that data exists in context, data itself can become the goal rather than an indicator of progress. This creates perverse incentives, as described by Campbell's Law.
Campbell's Law states that making a social-sector metric into a goal will distort the process it's intended to measure. We want teachers to have a high level of trust in their leaders, and it's worthwhile to assess the level of trust in a school, but we don't want to incentivize principals to pander to their staff to get good marks.
In baseball, a player who strives to maximize his batting average will stop trying to hit home runs and simply try to hit singles to get on base - it's easier to achieve better numbers with inferior play. Of course, baseball players aren't judged by a single statistic, which is why we track RBIs, home runs, and countless other measures of player performance. As in leadership, there are no shortcuts to excellence in baseball.
Evaluating principals with single or out-of-context measures also makes it difficult to attract talented leaders to struggling schools. If the leader's mandate is to turn things around, the leader should be judged on the basis of their actions and accomplishments, not on the basis of how people feel about having to change. If leading a struggling school is likely to lead to a negative evaluation, who will take on this work?
In other cases, though, it does make sense to take student performance and teacher perceptions into consideration. If a school experiences a consistent trend in student performance over a number of years under one leader, this does reflect on (while not precisely measuring) the quality of leadership.
The question is, who should turn this data into a judgment? Who should decide if a principal is doing the right work and doing it well? Data from a variety of sources can be helpful, but a formula (such as 25% test scores, 25% parent feedback, 25% teacher feedback, and 25% director judgment) can't account for the variation in the challenges principals are tackling. This takes expert judgment from a supervisor who knows the realities of the school, just as teachers must be evaluated by principals who actually spend time in their classrooms.
The LA Times recently took the stance that publicizing out-of-context data on teachers is a valid approach to improving schools (or at least selling newspapers). The public is hungry for data and eager to jump to facile conclusions about why our schools are they way they are. We must be quick to put this data in context.
Principal accountability is the leading edge of school improvement - schools will not improve unless their leaders ensure higher-quality teaching and learning. We principals can and must be held accountable for doing this work, and for doing it with the highest levels of integrity. It seems much more common that principals are dismissed for misconduct of the types Walt describes rather than for failing to tackle the real work their jobs require.
How do you think you and your peers should be evaluated? What incentives (wise or ill-advised) are in place in your district?
Justin Baeder (@eduleadership) is a public school principal in Seattle, Washington. He speaks and writes about principal performance and productivity, and is a doctoral student at the University of Washington in Educational Leadership & Policy Studies.