Does MSP Participation Increase the Supply of Math Teachers? Developing and Testing an Analytic Model

Abstract

An important feature of the Math and Science Partnership (MSP) Program of the National Science Foundation is to increase K-12 student achievement in math and science by increasing the quality, quantity, and diversity of the nation's K-12 math and science teachers. Because the underlying supply of math and science teachers is never directly observed, the central premise of this article is that an examination of the extent to which the MSP Program might impact the quantity and quality of math and science teachers requires careful thought and modeling.

With that starting point, this study first develops a model that supports a premise that shifts in underlying supply can be inferred from shifts in the percentage of certified math teachers employed when (a) salaries are constrained to be below market clearing salaries and (b) uncertified or “out-of-field” certified teachers can compete as substitutes for certified math teachers. The study then tests the plausibility of the model using data from Texas and in so doing provides preliminary estimates of the extent to which a school or school district's MSP participation affected the supply of certified math teaches available to that school or district. The results, although inconclusive on the question of the labor supply effects of MSP participation by a school or school district, do suggest the reasonableness of the model for future work when more appropriate data will be available.

ACKNOWLEDGMENT

This study is one in a series of substudies for the Math and Science Partnership Program Evaluation (MSP-PE) conducted for the National Science Foundation's MSP Program. The MSP-PE is conducted under Contract No. EHR-0456995. Since 2007, Bernice Anderson, Ed.D., Senior Advisor for Evaluation, Directorate for Education and Human Resources, has served as the National Science Foundation Program Officer. The authors, from Brown University, are John H. Tyler and Svetla Vitanova.

The MSP-PE is led by COSMOS Corporation in current partnership with George Mason University (GMU) and Brown University. Robert K. Yin (COSMOS) serves as Principal Investigator and Jennifer Scherer (COSMOS) serves as one of three Co-Principal Investigators. Additional Co-Principal Investigators and their collaborating institutions (including discipline departments and math centers) are Patricia Moyer-Packenham (USU, formerly GMU) and Kenneth Wong (Brown).

Notes

¹The minimum data requirements are data that contain employment information by school (campus), by year, and by subject area on the number of (a) employed full-time equivalent certified teachers teaching in their field, (b) certified teachers who are teaching out of field, and (c) uncertified or alternatively certified teachers. The necessary data would have both MSP and non-MSP schools, preferably in the same state and would cover both pre- and post-MSP participation years. Although the Schools and Staffing Survey from the National Center for Education Statistics has the necessary teacher employment variables, the fact that this survey is not longitudinal means that we cannot make pre- and post-MSP comparisons, rendering this data set unsuitable for the purposes of this article. We are still searching for other potential sources of data.

²The Alliance for Improvement of Math Skills and the Texas Middle and Secondary Mathematics Project are targeted at math, whereas the El Paso Math and Science Partnership is concerned with both math and science. A fourth project, the Rice University Mathematics Leadership Project was developed at a latter date than the other three MSP project in Texas and so we do not consider that project in our analysis.

³See, for example, Broughman and Rollefson (2000).

⁴We note that there is one MSP that entered into negotiations with the local teachers' union to change teachers' salaries.

⁵In both the modeling and the empirical analysis that follow, the focus is on high schools, because it is at the high school level that teachers are hired and appear in administrative records by subject.

⁶For narrative simplicity we use the term uncertified to represent both of the following categories: math teachers who lack standard certification and certified teachers who are teaching math as an out-of-field subject.

⁷Quantities along all dimensions in Figure 1 are increasing in the direction of the arrows of the axes.

⁸We note that the elasticity of supply of certified teachers with respect to salary is important, as the more elastic the supply, the greater the ability of a shift in supply to show up in employment levels. Given that we think that individuals with training in math and science are often “lured” away from teaching by better outside wage offers, the anecdotal evidence at least suggests that the supply of math and science teachers is relatively price elastic.

⁹Although there are some few districts nationwide that offer an extra stipend to math or science teachers, these stipends tend to be low relative to base salary.

¹⁰The condition of w ^k _c = w ^k _u is not required for the analytics to hold, but we note that in the case where all “uncertified” teachers were in fact certified teachers teaching math as an out-of-field subject, this would be the case because salaries are not generally a function of certification area.

¹¹If demand for education in the district is a function of teacher quality, then it is a simple matter to show that in the long run total teacher demand, which we are holding fixed, is a function of the ratio of certified to uncertified teachers and is therefore endogenous. Given the recent implementation of the MSP Program relative to available data, we focus on the short run analysis in this article.

¹²Of course, a thorough study of the amount of newly certified high school math teachers that resulted from an IHE's participation in the MSP Program would capture all newly certified teachers, both those who took employment in participating MSP high schools and those who took employment in non-MSP high schools.

¹³The different states of certification status in the data are certified and teaching in field, certified but teaching out-of-field, and uncertified.

¹⁴A fourth Texas MSP, the Rice University Mathematics Leadership Institute, did not start until 2004, so schools from this MSP are not included in our MSP group.

¹⁵Information on AIMS is found at http://www.delmar.edu/aims/Files/AIMS%20PR%20project%20summary.doc

¹⁶Information on the El Paso MSP is found at http://epcae.org/msp/msp.htm.

¹⁷Information on the Texas Middle and Secondary School Mathematics Project is found at http://www.faculty.sfasu.edu/kchilds/nsf2.html.

¹⁸Because we are looking at the percentage change in the number of math teachers within a school, this construct makes sense only for schools that have at least one certified math teacher.

¹⁹The exact interpretation is that the difference is e^β − 1, which is very close to β for the small values of β in most all of our results.

^a Alliance for the Improvement of Mathematics Skills, Del Mar College.

^b Texas Middle and Secondary Mathematics Project, Stephen F. Austin University.

^c El Paso Math and Science Partnership, University of Texas at El Paso.