2013 Identifying Effective Teachers Policy
The state should have a data system that contributes some of the evidence needed to assess teacher effectiveness.
Minnesota has a data system with the capacity to provide evidence of teacher effectiveness.
Minnesota has all three necessary elements of a student- and teacher-level longitudinal data system. The state has assigned unique student identifiers that connect student data across key databases across years and has assigned unique teacher identifiers that enable it to match individual teacher records with individual student records. It also has the capacity to match student test records from year to year in order to measure student academic growth.
Minnesota defines teacher of record as the educator responsible for awarding a mark and credit for the section. The state's teacher-student data link can connect more than one educator to a particular student in a given course, but it does not have in place a process for teacher roster verification.
Minnesota does not publish data on teacher production that connects program completion, certification and hiring statistics. In its annual "Education Statistics Summary," Minnesota publishes only the total number of new teachers (newly licensed and first teaching assignment) for that particular year.
Data Quality Campaign www.dataqualitycampaign.org Summary http://education.state.mn.us/MDEAnalytics/Data.jsp
Develop a definition of "teacher of record" that can be used to provide evidence of teacher effectiveness.
To ensure that data provided through the state data system are actionable and reliable, Minnesota should articulate a definition of teacher of record and require its consistent use throughout the state. The state's definition should reflect instruction rather than grading, and Minnesota should develop a process for teacher roster verification.
Publish data on teacher production.
From the number of teachers who graduate from preparation programs each year, only a subset are certified, and only some of those certified are actually hired in the state. While it is certainly desirable to produce a big enough pool to give districts a choice in hiring, the substantial oversupply in some teaching areas is not good for the profession. Minnesota should look to Maryland's "Teacher Staffing Report" as a model whose primary purpose is to determine teacher shortage areas, while also identifying areas of surplus. By collecting similar hiring data from its districts, Minnesota will form a rich set of data that can inform policy decisions.
Minnesota declined to respond to NCTQ's analyses.
Value-added analysis connects student data to teacher data to measure achievement and performance.
Value-added models are an important tool for measuring student achievement and school effectiveness. These models measure individual students' learning gains, controlling for students' previous knowledge. They can also control for students' background characteristics. In the area of teacher quality, value-added models offer a fairer and potentially more meaningful way to evaluate a teacher's effectiveness than other methods schools use.
For example, at one time a school might have known only that its fifth-grade teacher, Mrs. Jones, consistently had students who did not score at grade level on standardized assessments of reading. With value-added analysis, the school can learn that Mrs. Jones' students were reading on a third-grade level when they entered her class, and that they were above a fourth-grade performance level at the end of the school year. While not yet reaching appropriate grade level, Mrs. Jones' students had made more than a year's progress in her class. Because of value-added data, the school can see that she is an effective teacher.
The school could not have seen this effectiveness without a data system that connects student and teacher data. Furthermore, multiple years of data are necessary to enable meaningful determinations of teacher effectiveness. Value-added analysis requires both student and teacher identifiers and the ability to match test records over time.
It is an inefficient use of resources for individual districts to build their own data systems for value-added analyses.
States need to take the lead and provide districts with state-level data that can be used for the purpose of measuring teacher effectiveness. All states have longitudinal data systems, but not all states are yet able to connect student data to individual teacher records. Such data is useful not just for teacher evaluation but also to measure overall school performance and the performance of teacher preparation programs.
State Data Systems: Supporting Research
The Data Quality Campaign tracks the development of states' longitudinal data systems by reporting annually on states' inclusion of 10 elements in their data systems. Among these 10 elements are the three key elements (Elements 1, 3 and 5) that NCTQ has identified as being fundamental to the development of value-added assessment. For more information, see http://www.dataqualitycampaign.org.
For information about the use of student-growth models to report on student-achievement gains at the school level, see P. Schochet and H. Chiang, "Error Rates in Measuring Teacher and School Performance Based on Student Test Score Gains", July 2010, U.S. Department of Education, NCEE 2010-4004; as well as The Commission on No Child Left Behind, Commission Staff Research Report: Growth Models, An Examination Within the Context of NCLB, Beyond NCLB: Fulfilling the Promise to Our Nation's Children, 2007.
For information about the differences between accountability models, including the differences between growth models and value-added growth models, see P. Goldschmidt, P. Roschewski, K Choi, W. Auty, S. Hebbler, R. Blank, and A. Williams, "Policymakers' Guide to Growth Models for School Accountability: How Do Accountability Models Differ?" Council of Chief State School Officers' Report, 2005 at: http://www.ccsso.org/Documents/2005/Policymakers_Guide_To_Growth_2005.pdf
For information regarding the methodologies and utility of value-added analysis see, C. Koedel and J. Betts, "Does Student Sorting Invalidate Value-Added Models of Teacher Effectiveness? An Extended Analysis of the Rothstein Critique", Education Finance and Policy, Volume 6, No. 1, Winter 2011, pp. 18-42; D. Goldhaber and M. Hansen, "Assessing the Potential of Using Value-Added Estimates of Teacher Job Performance for Making Tenure Decisions." The Urban Institute/Calder, February 2010, Working Paper 31, and S. Glazerman, S. Loeb, D. Goldhaber, D. Staiger, S. Raudenbush, and G. Whitehurst, "Evaluating Teachers; The Important Role of Value-Added." Brookings Brown Center Task Group on Teacher Quality, November 2010; S. Glazerman, D. Goldhaber, S. Loeb, S. Raudenbush, D. Staiger, G. Whitehurst, and M. Croft, Passing Muster: Evaluating Teacher Evaluation Systems, The Brookings Brown Center Task Group on Teacher Quality, April 2011; D. N. Harris, "Teacher value-added: Don't end the search before it starts," Journal of Policy Analysis and Management, Volume 28, No. 4, Autumn 2009, pp. 693-699. H.C. Hill, "Evaluating value-added models: A validity argument approach," Journal of Policy Analysis and Management, Volume 28, No. 4, Autumn 2009, pp. 700-709; T.J. Kane and D.O. Staiger, "Estimating teacher impacts on student achievement: An experimental evaluation". National Bureau of Economic Research, Working Paper No. 14607, December 2008.
There is no shortage of studies using value-added methodologies by researchers including T.J. Kane, E. Hanushek, S. Rivkin, J.E. Rockoff, and J. Rothstein. See also T.J. Kane and D.O. Staiger, "Estimating teacher impacts on student achievement: An experimental evaluation". National Bureau of Economic Research, Working Paper No. 14607, December 2008; E.A. Hanushek and S.G. Rivkin, "Generalizations about using value-added measures of teacher quality." American Economic Review , Volume 100, No. 2, May 2010, pp. 267-271; J. Rothstein, 2010. "Teacher Quality in Educational Production: Tracking, Decay, and Student Achievement."The Quarterly Journal of Economics, Volume 125, No. 1,February 2010, pp. 175-214; T.J. Kane and D.O. Staiger, "Estimating teacher impacts on student achievement: An experimental evaluation". National Bureau of Economic Research, Working Paper No.14607, December 2008. S.G. Rivkin, E.A. Hanushek, and J.F. Kain. "Teachers, Schools, and Academic Achievement." Econometrica, Volume 73, No. 2, March 2005, pp. 417-458; E.A. Hanushek, 2010, "The Difference is Great Teachers," In Waiting for "Superman": How We Can Save America's Failing Public Schools, Karl Weber, ed., pp. 81-100, New York: Public Affairs.
See also NCTQ's "If Wishes Were Horses" by Kate Walsh at: http://www.nctq.org/p/publications/docs/wishes_horses_20080316034426.pdf and the National Center on Performance Incentives at: www.performanceincentives.org.
For information about the limitations of value-added analysis, see Jesse Rothstein, "Do Value-Added Models Add Value? Tracking, Fixed Effects, and Causal Inference." Princeton University and NBER. Working Paper No. 159, November 2007 as well as Dale Ballou, "Value-added Assessment: Lessons from Tennessee," Value Added Models in Education: Theory and Applications, ed. Robert W. Lissitz (Maple Grove, MN: JAM Press, 2005). See also Dale Ballou, "Sizing Up Test Scores," Education Next, Volume 2, No. 2, Summer 2002, pp. 10-15.