Recent articles regarding teacher performance evaluation have shed light on the phenomenon of leniency bias. Also known as leniency error and performance or interview bias, leniency bias comes when an interviewer or manager inappropriately or unfairly rates an employee in a positive light. It also comes into play when jurors [opens PDF] develop favorable opinions of a defendant in a court case outside of the facts presented.
With regard to evaluation, leniency bias has been described using the famous, fictional Minnesota town Lake Wobegon: everyone there is above average in some way. When assessing teacher effectiveness, as many as 98 percent of teachers were rated effective or better, regardless of how their students were performing, and relatively few (3 to 4 percent) were given the lowest possible effectiveness ratings and subsequently fired for poor performance. This comes despite calls for greater accountability from teachers and education administrators and claims that “5 to 10 percent” of teachers are ineffective. Outside of education, private companies with multiple tiers of performance tend to rate most employees in their top two echelons, above average and higher.
Why, when given the opportunity to identify ineffectiveness in a system and rectify the “problem” of ineffective teachers, does it happen so infrequently? In the example of teacher effectiveness, one reason is that the effectiveness (or lack thereof) changes depending on the measurement. At one point, Florida teachers were being evaluated half on student test scores and half on classroom evaluations, but the acceptable test score range has changed frequently in recent years, while the observation assessment at one point consisted of a single 20 minute observation in one year. Because of these shifting measurements, from year to year teachers can fluctuate between being labeled “effective” to “ineffective” and back again, rendering the designations moot.
Another factor that has been discussed are the risks that principals and other supervisors must take into account when determining effectiveness beyond performance. The demands on and for a teacher can differ widely depending on socio-economic and geographic impact, as does the likelihood of quickly replacing a teacher. As previously discussed on ACET’s blog, the state of Minnesota’s own desire to measure more complicated indicators of student growth and success has led to the adoption of its own standards. Principals who must contend with low turnover and a low supply of available teachers in the local job market risk a lengthy, expensive search for replacements in a field where up to 60 percent of professionals leave after their third year (per the Georgetown Public Policy Review). Rather than undertaking such searches, some principals may feel that it is better to stick with an ineffective but established teacher than expend resources on an unknown quantity that could potentially be even worse.
Another proposed reason for this leniency bias is that there is no bias at all, that there just isn’t “5 to 10 percent” worth of ineffective teachers to remove. Foundations and publications have reviewed the evaluations that produced so few identifications of ineffectiveness and found the resulting data, while not yet ready for mass consumption, conforms to their own suspicions of the true level of teacher ineffectiveness (3 to 4 percent, rather than the proposed 5 to 10).New evaluation systems continue to be developed to address these factors. However, it seems appropriate to conclude that as soon as a solid set of measurements has been decided on, it will be easier to determine if teacher effectiveness is the indicator of student success that it has been held up to be.