A comparison of persistence assumptions for estimating teacher effects

Temurtaş A., ÇOBANOĞLU AKTAN D.

School Effectiveness and School Improvement, vol.37, no.1, pp.27-50, 2026 (SSCI, Scopus) identifier

  • Publication Type: Article / Article
  • Volume: 37 Issue: 1
  • Publication Date: 2026
  • Doi Number: 10.1080/09243453.2025.2547666
  • Journal Name: School Effectiveness and School Improvement
  • Journal Indexes: Social Sciences Citation Index (SSCI), Scopus, IBZ Online, Child Development & Adolescent Studies, Education Abstracts, Educational research abstracts (ERA), ERIC (Education Resources Information Center), Psycinfo
  • Page Numbers: pp.27-50
  • Keywords: generalized persistence, persistence, Value-added models
  • Hacettepe University Affiliated: Yes

Abstract

The persistence of teacher effects is a crucial aspect in value-added models, especially when it comes to high-stakes decisions. This study compared various assumptions of persistence (generalized, variable, complete, and zero) regarding their model fit and consistency of estimates. Longitudinal data from assessments in mathematics, science, social sciences, and Turkish subjects were analyzed. Results indicated that the model with the variable persistence assumption demonstrated superior fit in three out of four subjects for teacher value-added score estimates. The correlation between the estimates for models with generalized and variable persistence assumptions was notably high. However, substantial variances were observed in estimates regarding their magnitude and corresponding rank order. The study deliberated on the implications of different persistence assumptions on teacher value-added score estimates.