Association for Public Policy Analysis & Management

There is a strong policy interest in increasing the diversity of the teacher workforce and examining the teacher production pipeline to find ways to address the relative lack of teachers of color. Evidence from other occupations suggests that the imposition of licensing requirements on previously unlicensed professions may increase minority representation in the labor force. Teaching, however, has long been a licensed profession, and research within education has focused on the specific requirements of licensure laws and the role of standardized tests in particular. One policy lever that can affect the supply of new teachers is the passing standard set on a licensing exam. This paper uses a difference in difference design to assess the impact of changing licensure test passing standards on the pool of teacher candidates. To date, we have found suggestive evidence that lowering licensure standards may increase the diversity of the candidates who attempt the test, yielding general equilibrium effects that are different than those implied by a static analysis. Work is ongoing to validate and extend these results.

Researchers have typically examined the effects of specific passing standards on the racial makeup of the teacher workforce or quality of the teacher workforce through single time period comparisons of outcomes across states with different passing standards. These comparisons rely on the assumption of a constant pool of teachers, and that the only effect of changing passing standards is the direct effects on passing rates. These typical approaches do not allow for general equilibrium effects. In this paper, we explicitly test the stable population assumption.

We use within-state changes in passing standards to examine the effects of a change in passing standards on the score distribution and composition of the pool of test takers. This paper uses a nineteen-year panel of ETS Praxis exam records covering dozens of states and multiple subject area tests. With records of the required passing standard used at the state – test – year level, we estimate the effects of passing standard changes using difference in difference models. We compare the average scores of candidates before and after a passing standard change in a state – test to the change in scores of candidates in other states. Preliminary models that examined a small subset of four tests found evidence that average standardized scores of first-time test takers decreased by a tenth of standard deviation after passing standards were lowered. The effect was most pronounced in the pool of Black teacher candidate where the average score decreased by a half of a standard deviation. We also found suggestive evidence that the proportion of candidates who were Black and the proportion reporting being trained in ‘alternative’ (i.e. not traditional two- or four-year institutions) increased in response to the decrease in passing standards. Licensure tests are an integral component of state teacher certification processes. This research directly assesses the role of the passing standard set on the composition of teacher candidates, and will provide evidence to better inform licensure policy decisions.