ABSTRACT
The level of educational productivity, like other types of productivity, depends on how well inputs are turned into outputs by the behaviors or processes of individuals and organizations. Teacher accountability can also be conceptualized in these terms. Teachers and schools can be held accountable for inputs such as the use of funds as budgeted or for hiring only licensed teachers, for practices or behaviors such as adopting specified curricula or maintaining an orderly classroom, or for outputs such as student learning or outcomes such as student achievement on state or district tests. Now, teacher accountability is mostly thought of in terms of measuring student achievement using test scores. Although this thinking is in part a healthy reaction against previous overemphasis on inputs, test-based measures of student achievement have some disadvantages as the sole basis for teacher accountability systems. On the technical side, these include limitations on any test’s representation of the subject domain (e.g. by having too few or no questions on some important material), misalignment of tests and curricula, imprecision of average school, or classroom scores because of small class or school sizes, the potential for score inflation, and narrowing of the curriculum. (See Koretz, 2002, for more on these issues.) In addition, many teachers teach in subjects for which only teacher-made tests are available to measure learning, and in urban districts, high student mobility results in many students having more than one teacher or school during any one year. These problems make it difficult to attribute achievement to a specific teacher or school.
