Given the prevalence of supports reported by control teachers, we explored the relationship between induction supports and outcomes independent of group assignment (treatment or control) and district type (one-year or two-year). Using data from the first three Induction Activities surveys, we created a variable that reflects the number of years (0, 1, or 2) the beginning teacher had an assigned mentor and constructed three other new measures10:
Overall, we found that induction measures were not significantly related to math test scores (p-value of F-test = 0.068) or reading scores (p-value of F-test = 0.651). However, we found that the association between the years the beginning teacher had a mentor and math test scores was statistically significant (regression coefficient = 0.12, p-value = 0.015). For measures of teacher retention, there was a statistically significant relationship between the induction activities variables and retention (p-value of F-test = 0.016 for remaining in the district; p-value of F-test = 0.001 for remaining in teaching). One measure—the Induction Services Index—was positively related and no measures were negatively related to teacher mobility for both remaining in the district and remaining in teaching. The estimate of the regression coefficient on the Induction Services Index for remaining in the district was 0.02; for remaining in teaching, it was 0.01. This implies that, for example, if the retention rate in a district were 80 percent, then an additional induction service, such as meeting with a study group in one semester, would be associated with a district retention rate of 82 percent, all else equal. All results were robust to alternate methods of constructing the indices and alternate model specifications.