The Teacher Followup Survey (TFS) consisted 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 and obtains data on educational activities and future plans for all groups; and to collect data on attitudes about the teaching profession and job satisfaction.

The TFS sampling frame consisted 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), level (elementary and secondary) and years in teaching (new and experienced). A large majority of teachers in all categories were stayers and it was necessary to oversample movers and leavers in order to reduce the sampling errors of estimated differences among groups.

In January 1992, questionnaires were mailed to the sampled teachers' home addresses. 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 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. A general clerical edit checked for out of range values. Inter-item consistency was also checked and items were blanked when respondents answered questions unnecessarily because they did not follow the skip instructions correctly.

There were three stages of imputation used for the 1991–92 TFS: logical imputation and "hot deck" and a proration imputation procedure was applied to the current teacher questionnaires to adjust components of earned income so that they added to the total earned income. In addition, a clerical imputation was used when no suitable donor record was found, when the computer imputation produced an out of range entry, or when there were very few cases where an item was unanswered.

A ratio adjustment was used that forced the weighted sample counts from the TFS to conform to the weighted number of teachers from SASS. In 1991-92, another weighting factor was used to adjust for the fact that preliminary SASS final weights were used in computing the TFS basic weight.

Weighted response rates are defined as the number of in-scope responding questionnaires divided by the number of in-scope sample cases, using the basic weight (inverse of the probability of selection) of the record.

Reinterviews by telephone were conducted to obtain some indication of measurement error. The reinterview sample consisted of 1,498 teachers, approximately half current teachers and half former teachers. Reinterview results were analyzed using one index of inconsistency for items with a dichotomous response that is accounted for by response variance. A separate index of inconsistency was used for response items with more than two response categories.