Teachers are the largest component of school spending, with more funding being allocated to teacher salaries than to any other educational expense (Loeb, Miller, and Strunk 2009). Teacher and staff1 salaries and benefits consume up to 80 percent of current expenditures (Aud et al. 2010). Yet, there is not a wealth of data on teacher compensation. National data on teachers are limited to periodic sample surveys or to simple counts at the district or school level. School districts and states often maintain significant databases on teachers in their jurisdictions, but these databases are rarely comparable across states. Some databases contain personally identifiable or sensitive information (e.g., Social Security Number), thereby preventing them from being available to researchers and the public.
Comparable teachers' compensation data across districts and states are needed in order to address a wide variety of education policy issues. For example, many commentators believe teachers are the key determinant of school quality (Hanushek and Rivkin 2004). Accordingly, the ability of schools to attract and retain high-quality teachers to work in urban districts is currently the focus of new policy initiatives. Recently, school officials in urban districts such as Denver, New York, and the District of Columbia have been contemplating "front loading" teacher compensation by increasing the salaries for new teachers. New York City Schools Chancellor Joel I. Klein stated, "You want to allocate your money in a way that attracts new talent and rewards excellence" (Sawchuk 2009).
In response to the need for individual teacher-level data to address these and other policy issues, the National Center for Education Statistics (NCES) developed the Teacher Compensation Survey (TCS), an administrative records universe survey that collects total compensation, teacher status, and demographic data about all public-school teachers from multiple states. In 2007, NCES launched the pilot TCS data collection, with seven states volunteering to provide administrative records for school year (SY) 2005–06. In the second year of the data collection, the TCS expanded to 17 states reporting SY 2006–07 data.
The TCS offers several advantages over other data sources. For example, much of the teacher compensation research to date has been based upon sample surveys. The TCS removes sampling error and self-reporting bias through the use of a dataset that contains universe data at the teacher level for multiple states.
The TCS file can be merged with the Common Core of Data (CCD) Public Elementary/Secondary School Universe Survey file (referred to as the School Universe Survey, or School Universe, in this report) to obtain such school information as school type, operational status, locale code, number of students eligible for free and reduced-price lunch, student totals and detail (by grade, race/ethnicity, and sex), and pupil/teacher ratio.
The TCS permits comparisons of teacher salaries at various points along the career trajectory according to teacher characteristics (such as teacher's educational attainment, years of teaching experience, etc.). To ensure data confidentiality, the TCS does not use Social Security Numbers as the identification numbers for teachers. The Census Bureau has assigned new teacher IDs to all teachers for the TCS data file. The TCS data have also undergone a perturbation to eliminate the possibility of the data being used to identify individual teachers. The TCS is designed to provide comparable data across states and districts and may shed light on the compensation necessary to attract teachers, the ability to retain teachers, and teacher mobility.
The TCS data collection is a research and development effort to see if it is possible to collect and publish teacher-level data from the administrative records residing in state education agencies (SEAs). This report provides an overview of the TCS data collection for SY 2006–07; a comparison of state administrative records of the TCS with other sources of data; and a discussion of the data availability and quality, as well as limitations, of the TCS. This report also includes findings and descriptive statistics for SY 2006–07.