HD34 - Faculty Salary Benchmark Groups


Executive Summary:
Virginia has a long history of using benchmarks to evaluate faculty salaries to ensure that Virginia’s colleges and universities stay competitive in the recruitment and retention of faculty. Prior to 1987, the Council employed outside consultants who reviewed institutions nationwide to develop a set of faculty salary benchmarks to be used in comparing faculty salaries at Virginia’s colleges and universities to comparable institutions nationally.

In 1987, the Council, working with representatives from the General Assembly money committees, the Department of Planning and Budget, the Secretary of Education, and institutions, developed institutional benchmark groups for each of the public colleges and universities in an effort to determine appropriate faculty salaries at the institutions. Since the schools in Virginia are all very different from each other in terms of mission and the range of degrees offered, the schools were compared to similar institutions outside of Virginia. It was determined that each school would have a group of benchmark institutions determined through statistical analysis, and that faculty salary increases would be based on the average faculty salary of the benchmark group. The General Assembly set forth a policy to peg faculty salaries to the sixtieth percentile of the institutions’ benchmark groups.

The benchmarking process has been conducted system-wide three times in the last twenty years – 1987, 1997, and 2007. Updates have been made for select institutions in intervening years. The last update took place in 2001 for three institutions – Longwood University, Norfolk State University, and Virginia State University.

The basic process has remained fundamentally unchanged since it was first initiated in 1987. The process consists of three phases: (1) data identification and collection, (2) statistical modeling, and (3) institutional meetings. The first two phases are quantitative in nature. The final phase allows the institutions to address quantitative and qualitative information not addressed in the statistical model.

The benchmark review began in July, 2006, when SCHEV staff met with staff from the General Assembly money committees. The result of this discussion was general confidence in the process that was used in 1997. The consensus of the group was to leave the basic process intact and to make only minor adjustments to recognize changes in the availability of data and institutional characteristics.

The next three months were spent reviewing the 1997 process and results. SCHEV staff set about matching data used for the 1997 analysis with current data sources. Staff identified several data elements used in 1997 that were no longer available or reliable.

SCHEV staff conducted individual meetings to consult with institutional staff in the review. Staff visited all college and university campuses (meetings with University of Virginia – Wise and the Virginia Community College System were held in Richmond) from November, 2006, through February, 2007. These sessions allowed institutions to involve many more of their staff than would be available for statewide meetings in Richmond. It also gave institutions an opportunity to raise issues that were specific to them – concerns that they might not have raised in a more general setting. These discussions and feedback were very helpful to SCHEV staff.

Once these meetings were completed, SCHEV staff developed a set of data elements to be used in the statistical analysis. This list was shared with the institutions in early March for their review and comment. Institutional feedback was received in late March and incorporated into the statistical model (Tables 1a-1d).

SCHEV staff compiled a set of data on over 3,000 colleges and universities nationwide. These data were shared with institutions in April for their review and analysis. This gave institutions the opportunity to discover any errors or omissions before SCHEV staff ran the statistical model in May.

The model used a statistical procedure called “cluster analysis” to identify institutions similar to a given Virginia institution based on 17 to 19 (10 for Richard Bland College) quantitative characteristics. For each Virginia institution, a cluster was drawn from a pool of institutions nationwide. The statistical model calculated how similar an institution was to the Virginia institution. This procedure was done for each of Virginia’s colleges and universities. The result was a list of the top 75 most similar institutions from across the nation for each Virginia school.

The final step in the process was to narrow down the list to 25 institutions. The 25 benchmark institutions were chosen by a committee consisting of representatives from SCHEV, the Department of Planning and Budget, the office of Secretary of Education, the office of the Secretary of Finance, the House Appropriations Committee, the Senate Finance Committee, and the institution. Meetings with each institution were held during the month of June in Richmond. The final list for each school was selected from the list of 75 based on the information from the cluster analysis and additional information brought by the institution. This final step added subjectivity to the process and was important since the cluster model could not capture all the characteristics of each institution.

Final lists of 25 benchmark institutions for each four-year institution and Richard Bland College are presented in Tables 3a-3p. The process for developing lists for the 23 colleges in the Virginia Community College System is ongoing and these lists will be given to Council in September, 2007.

Table 2 outlines the changes to 60th percentile goals for each institution. The full fiscal impact of the changes will be determined as the budget is developed over the next several months.