10 Percent Freeze-Out? A Debate With Eric Hanushek and Diane Ravitch

My TIME column last week was about our lousy national debate about teaching quality. But although it was toward the bottom this paragraph sparked a lot of debate:

When Stanford University economist Eric Hanushek looked at teacher performance he found that removing even the lowest five percent of performers could boost overall student achievement substantially. There are two key takeaways from this research. First, the lowest-performing teachers have a negative effect on student performance that is disproportionate to their numbers. Second, in practice this amounts to just one or two teachers per school on average. Most workplaces have similar problems.

Historian Diane Ravitch and Hanushek disagree about this point and agreed to discuss it here on Eduwonk next week. Thanks in advance to both of them for that.  So addition to regular content look for that debate to start on Monday.

8 Replies to “10 Percent Freeze-Out? A Debate With Eric Hanushek and Diane Ravitch”

  1. I look forward to the debate. I am sure that poor teachers have a disproportionate effect on their students, but I guess the real question is how do we replace them with better ones.

  2. For me I think the problem is that more of us would be on-board with dismissing poor teachers if we had some guarantee of a system that works. Beyond a few pilot efforts across the nation, a reliable system of teacher assessment remains elusive while so much reform remains in flux. Because it seems like everything is happening at once, it’s hard to focus on one problem and not have it perceived to have some disproportionate importance.

    Seriously, I would doubt anyone in the teacher camp is so biased or entrenched as to not want to dismiss poor teachers but there are decades of suspicions and with the ongoing chaos of reform priorities, it simply makes us question the process.

  3. I look forward to hearing both sides of this issue because it’s not as simple as it seems. The whole movement to quantify student performance and teacher performance isn’t as straightforward as it may seems.

    We’ve reduced student achievement to API numbers that can be compared across schools and within the state, but these numbers aren’t as reliable as we’ve begun to see as stories continue to emerge about how teachers and administrators are inflating these in unethical ways.

    Intuitively, we all “know” who the poor, uncommitted teachers are. But when we start talking about eliminating the bottom 10% of them we’re opening ourselves up to legal challenges. For example, how do we DEFINITIVELY say that bad teachers in the 11th-percentile (as determined by some mathematical formula) are worthy of maintaining their teaching positions when those one percentage point lower should be fired.

    Personally, I think that they should all be removed as a way of improving the lives of students. But how do we determine some arbitrary cutoff line?

    This is tricky business, and I look forward to hearing the two discussants’ views on the topic.

    But this is an excellent starting point in improving public education.

    John

  4. First and foremost we should know why our students are not performing well in schools and where is the problem, if is because of teachers we should suck them and employ the new ones who will help in rising the performance of our schools. On doing this will help in improving our education sector.

  5. Its difficult to argue against removing the worst 5-10% of any profession and not expect improved productivity and results. The problem is how best to determine who those teachers are. The system most advocate is one based upon factoring in student performance testing data. No evidence exists that this is the best measure of teacher effectiveness. A more comprehensive approach utilizing peer review, teacher training, principal evaluation and some student testing data seems the most appropriate course of action.

  6. I’m not even sure why we’re discussing the “novel” approach of removing 5-10% of teachers — As if teachers are never dismissed or encouraged to leave their jobs otherwise. In my experience, struggling teachers often leave of their own accord (especially in the first few years). In addition, principals have the power to fire at will (i.e., for no reason at all) teachers in their probationary period (usually around 3 years) before they get tenured. Further, principals can fire tenured teachers by following the evaluation rules and documenting poor teaching.

  7. It’s interesting to see so many comments popping up here that seem to be linked to spam sites, but the comments otherwise seem relevant and on point.

  8. While many ignore it, accuracy of rankings is an issue, as is stability, i.e., whether teachers who are ranked in one category one year are in the same category the next. The same teachers do not get the same rankings every year. Sometimes teachers, like baseball players, have a bad year. There are lots of reasons for this, including change of administrators and programs, personal problems (divorce, illness and death of a partner all have immense effects), introduction of new curriculum – you name it. And they do not have the same students every year; even under conditions of random assignment of students, there is luck of the draw. Randomization does not always result in a balance between treatment & control groups.(1)

    Still one would expect a fair amount of stability if one makes decisions on this, but even according to Eric Hanushek, “It’s not clear how much year to year variation there is in a teacher’s performance anyway,” but suggested reasons for instability ranging from divorced to getting a master’s degree. (See Frederick Hess, “My Take on the L.A. Times Reanalysis.”) What he does not mention as reasons a teacher’s value-added may vary year to year are school effects, such as having a supportive administration, correct teacher placement, a school culture that promotes respect for teachers and learning, etc.

    Stability and accuracy are not the same things. Rutgers professor Bruce Baker has raised the issue of whether two goals of statistical teacher evaluation systems are potentially in conflict. Those two goals are:

    A preference to isolate as precisely as statistically feasible, the influence
    of the teacher on student test score gains;

    A preference to have a statistical rating of teacher effectiveness that is
    relatively consistent from year to year (where the more consistent
    models still aren’t particularly consistent).(2)

    Baker goes on to “argue that the pressure to achieve the second objective above may lead researchers – especially those developing models for direct application in school districts – to make inappropriate decisions regarding the first objective.”

    (1) See, for example, Springer, M.G., Ballou, D., Hamilton, L., Le, V., Lockwood, J.R., McCaffrey, D., Pepper, M., and Stecher, B, “Section III: Threats to Validity [and] Imbalance Between Treatment and Control Groups,” Teacher Pay for Performance: Experimental Evidence from the Project on Incentives in Teaching. Nashville, TN: National Center on Performance Incentives at Vanderbilt University, 2010, p 7.

    (2) Bruce D. Baker, “The Perils of Favoring Consistency over Validity: Are “bad” VAMS more “consistent” than better ones?” School Finance 101 blog, April 29, 2011; accessed September 2011 at http://schoolfinance101.wordpress.com/2011/04/29/the-perils-of-favoring-consistency-over-validity-are-bad-vams-more-consistent-than-better-ones/.

Leave a Reply

Your email address will not be published.