The purpose of this report is to suggest ways to improve the school staffing information gathered through the Schools and Staffing Surveys (SASS) currently administered by the National Center for Education Statistics (NCES).  Spending on school-level personnel, including employees and personnel service providers, accounts for more than 85 percent of the expenditures at the school site. Obtaining better information on school staff can provide insights into the patterns of resource allocation in schools and the access of children to instructional and related services.
 
A methodology for collecting accurate school-level staffing information through the SASS was proposed
and investigated.  This methodology involves:

• Modifications of the current Teacher Listing Form,
• Administration of other staff listing forms to cover all school-level staff,
• Administration of a form to gather information on staff salaries and benefits for a sample of school-level staff, and
• Addition of other questions to the current district and school-level SASS questionnaires for the purpose of estimating expenditures per pupil for various educational services.

The other staff listing forms can either be administered concurrently with the Teacher Listing Forms or subsequently (for example, concurrently with administration of the SASS school questionnaire).  The administration of these other forms concurrently with the Teacher Listing Form would increase the comprehensiveness of the universe of teachers listed, increasing the generalizability and validity of SASS Teacher Survey data.  However, the increased respondent burden associated with their completion could adversely affect response rates.

The proposed methodology was evaluated and modified through a multistage process that included:

• Informational interviews with principals and school district superintendents,
• “Pre-pilot” field testing of the proposed listing forms, involving completion of draft forms and telephone administration of a survey debriefing protocol to respondents in three states, and
• On-site cognitive testing and validation of these listing forms and other materials at two schools in each of three districts.  These schools were located in California and Virginia.

The cognitive interviewing revealed several problems with the current Teacher Listing Form which can be ameliorated prior to the next administration of the SASS whether or not the proposed procedures for the collection of better school resource measures are implemented.  The cognitive interviews and validation also demonstrated substantial improvements in accuracy of teacher listings as a result of the concurrent administration of other staff listing forms.  The other staff listing forms served as memory aids by compelling consideration and classification of all staff.  This increased the number of teachers listed in comparison with the number listed prior to completion of the other staff listing forms.

Cognitive interviewing was not designed to determine the impact of the additional staff listing forms on response rates nor was it designed to inform the feasibility of a large scale implementation of the procedures.  To do this, a pilot test is recommended.

The proposed procedures were designed to be linked with the salary and benefits information provided by school districts.  Accordingly, two alternative approaches for collecting salary and benefits information from school districts were investigated through:

• Informational interviews with district staff and
• On-site interviews in three districts (in two states).  These interviews employed cognitive survey research techniques, including modified think-aloud protocols, directed probing, and projective techniques.

The methodology indicated substantial problems with one of the approaches (the collection of information through general salary and benefit information forms).  These problems reflected the tremendous heterogeneity of benefits packages and eligibility for participation, both within and between districts.  Therefore, an approach involving the collection of information about the salaries and benefits of specific district employees seems preferable.  Implementation of this approach requires careful attention to issues of confidentiality and respondent burden.  It should be pilot tested before considering implementation.

To inform decisions about incorporating the proposed procedures into SASS, the following steps are recommended.  First, a more complete testing of the final instruments to assess their impact on instrument and item non-response should be undertaken.  Second, the data should be key-entered into the computer and analyzed to identify any potential problems with the processing and analysis of the data obtained from the field.  Third, one would need to followup with a larger sample of respondents to determine how well the forms and questions were interpreted and completed in the field.  Fourth, the data collection instruments and procedures proposed in this study would have to be adapted for, and pilot tested in, the private school setting.