VI. Conclusions

The goal of this project was to learn more about the data collected by state education agencies, specifically, to determine the extent to which individual records of employment and compensation exist for teachers in public schools, how such records might be accessed, and what elements critical for making comparisons are commonly collected. We found that many state education agencies keep fairly detailed records on individual teachers. Specifically, 34 of the 40 participating state agencies indicated that they did maintain individual compensation and employment records of teachers, and many of these states have granted, or indicate a willingness to grant, researchers access to their data.

In some states, procedures to gain access to these data may be relatively simple; in others, the data may not be made available at all. Some agencies have already created research versions of their teacher datasets (with, for example, encrypted identifiers) that are available to researchers for labor market studies. In other states, access may vary according to request, with some data only available to researchers in coordination with NCES. While some SEAs did indicate that their data will not be made available to either NCES or researchers, it is reasonable to expect that over time an increasing number of SEAs will collect and make available data on teacher compensation.

Several critical data elements are common in varying degrees across SEA databases, including job title, contract length, full-time status, base or total salary, and education and experience level. However, definitional issues associated with many of these data elements may require further clarification before true "apples-to-apples" comparisons can be made using data from multiple SEAs' data systems.

Pooling these state teacher files into a large multistate file would permit disaggregated comparative studies of teacher pay. These studies would not only permit overall comparisons of pay, but also comparisons of teacher pay at various points along typical career trajectories (e.g., bachelor's degree with no experience, master's degree with 10 years' experience), with breakdowns by teacher demographics and state or district characteristics. Unique teacher IDs in these files would also permit multistate studies of teacher mobility and turnover.

Finally, analysis of the data elements available in pension fund data from a sample of 6 states suggests that these data do not appear to be as useful as SEA data in making interstate and interdistrict comparisons of teacher pay.