Observers committed to reducing error should consider multiple measurements for teacher evaluation.Yes, Evaluations Can Be Fair and Accurate
In this month’s ASCD, Robert Marzano discusses ways to minimize error and maximize accuracy and fairness when principals, coaches, or other administrators are conducting classroom observations. Marzano recommends using multiple measures for teacher evaluation, and urges administrators to conduct as many observations as possible, given the realities of most educators’ schedules.
Think about it: Teachers teach different types of lessons for different purposes. Obviously, a principal who steps in when a teacher is introducing content (Marzano Design Question 2) is going to find a very different classroom from one in which a teacher is practicing and deepening knowledge (Marzano Design Question 3), or when students are generating and testing hypotheses (Marzano Design Question 4). And in fact, teachers spend an average of 60 percent of their time introducing new content – and only five percent helping students generate and test hypotheses. If an observer conducts only two or three observations a year, the chances of hitting a classroom when students are working on those kinds of higher-order thinking skills are rather low. And there’s another difficulty: Observing for Design Question 4, the cognitively complex tasks, is going to become more and more important as schools begin to implement new Common Core State Standards. In short, administrators will need to see how teachers are doing with teaching skills that in the past, they frankly haven’t spent much time with.
To reduce the likelihood of sampling error (we can never eliminate it entirely), supervisors would also do well to consider moving beyond the traditional classroom observations as evidence for teacher evaluation. Marzano recommends the following:
1. Teacher Self-Evaluation
What’s that, you say? Won’t teachers score themselves too high?
Perhaps. But a teacher’s self-evaluation can work as a useful comparison. Does is match the evaluators’ score? Is it too low? Is it too high? If the scores match up, chances are the evaluation is fairly accurate. If the teacher’s score is low, there’s a chance the evaluator’s may be inflated. And if a teacher scores himself too high, he may have an inaccurate picture of his own classroom practice, which may help focus the administrator’s feedback.
2. Announced Observations for Different Lesson Types
Try scheduling three announced observations: one for each of the three lesson types discussed above.
Even though the teacher won’t be “surprised” when you drop in, and may have spent extra time preparing to impress, he or she will still score fairly accurately on the skills being rated. Nobody learns the specific skills needed to score at “innovating” when “organizing students to engage in cognitively complex tasks” overnight, no matter how much cramming they’ve done.
3. Use Brief Walk-Throughs as Unannounced Observations
The three to five minute walkthrough can be a great way to resolve uncertainties or reconcile differences in scores between lessons, or between a teacher’s self-evaluation and a formal observation.
4. Record Teachers’ Classes on Video
Big brother isn’t watching. But different raters can watch video and score independently for improved inter-rater reliability. Bring in the teacher to score his or her own video for a further measurement.
5. Let Teachers Challenge Scores
It may sound like a headache. But no observer is perfect, and a teacher might provide evidence of their expertise in specific elements: videos, student artifacts, student surveys. Remember, evaluation should empower teachers; it should give them a stake in their own professional development. No teacher will improve if they are denied agency.
Administrators, do you have tips you’d like to share about how you improve observer accuracy? Leave a comment in the comment section below.