The Teacher Follow-up Survey (TFS) consists of two questionnaires, one for current teachers and one for former teachers. The TFS was designed to measure attrition rates and to compare teachers who left the teaching profession, teachers who moved to another school, and those who stayed in the same school as the previous year. TFS also measures the current economic activities of leavers, obtains data on educational activities and future plans for all groups, and collects data on attitudes about the teaching profession and job satisfaction.
The TFS sampling frame consists of all eligible teachers who responded to the SASS Teacher Survey in the previous year.
A primary sample design objective for the TFS was to support comparative analyses of stayers, movers and leavers for teachers classified by sector (public and private), grade level taught (elementary and secondary), minority status and years in teaching (new and experienced). In order to achieve the design objective, the sample of respondents was grouped into 36 strata formed by the cross-classification of sectors mentioned above.
In January 2005, questionnaires were sent to the sampled teachers' home addresses. As part of an experiment on response rates for mailed vs. internet questionnaires, approximately one-third of the sampled teachers were mailed a paper copy of the questionnaire, while the remaining respondents were offered the option of completing the questionnaire on the Internet. In preparation for the mailing, in the previous year, the SASS Teacher Questionnaire asked for the teacher's home address and contact information for two people who would know the whereabouts of the teacher. Teachers who were sent questionnaires that were inappropriate for their status were asked to return them and the correct version of the questionnaire was then sent.
Data processing for the TFS was a multi-stage procedure. Completed questionnaires were submitted to a series of computer edits consisting of a range check (i.e., identifies values that fall outside of an expected range), a consistency edit (i.e., checks for consistency in responses to specified related items), and a blanking edit (i.e., blanks responses that are inappropriate for given skip patterns).
There were three stages of imputation used for the 2004–05 TFS: In the first stage, items were imputed with a valid response using data either from other items in the same TFS questionnaire or from items in the related 2003–04 school or teacher questionnaires. In the second stage, hot deck imputation, subsample ratio imputation, and random subsample ratio imputation methods were used. In the final stage, all remaining unanswered items were imputed manually.
In the weighting procedure, the base weight was used as the starting point. Next, a weighting adjustment was applied to reflect the impact of the SASS teacher weighting procedure. Next, a nonresponse adjustment factor was calculated and applied, and finally a ratio adjustment factor was calculated and applied to the sample to adjust sample totals to frame totals in order to reduce sampling variability. The final weight for TFS is a product of these factors.
Weighted response rates are defined as the number of in-scope responding questionnaires divided by the number of in-scope sample cases, using the base weight (initial basic weight multiplied by the sampling adjustment factor) of the record.
|Sector and teacher type||Base-weighted response rate|