Learning From Stakeholders’ Perceptions of Teacher Preparation: Using Latent Class Modeling in Program Evaluation

• In Event: Reforming Teacher Preparation and Accreditation: The Importance of Research-Based Evaluation of Teacher Education

Abstract

Purpose
This study investigates how candidates’ and mentor teachers’ ratings on program effectiveness compared across: (a) type of learning outcome; (b) group membership, (c) program affiliation; and (d) time in program.

Background
Self-studies are an effective tool for program evaluation, especially when alternative views from stakeholders are sought (Berry, 2004; He & Levin, 2008; Loughran, 2006). Candidates’ perceptions of programs can be very different from the intended design (Loughran); mentor teachers, although largely committed to candidates’ successful field experiences, may be less focused on issues that extend beyond their own school context (Torrez, 2012). In this research, these two stakeholder groups were surveyed to examine the consistency between the vision and practice of one urban college of education.
Methods and Data sources
Candidates and mentor teachers responded to an anonymous survey constructed on the ten learning outcomes of the college’s conceptual framework, using a four point scale to rate the effectiveness of the teacher education programs. The survey also included two open ended questions on the strengths of the programs as well as recommendations for improvement. A total of 273 candidates and 342 mentors responded to the survey. Data were analyzed using Latent Class modeling. Post-hoc thematic analysis of the open ended responses explored the unique characteristics of the latent classes.
Results
Latent class analysis grouped responses into classes sharing similar features: class 1 included 9.4% respondents who gave low ratings, class 2 included 33.7 % respondents whose ratings fell in between, and respondents in class 3 (56.9 %) gave the highest ratings. Overall, candidates and mentors did not differ in how they rated the learning outcomes; however, there were significant differences in ratings across the ten learning outcomes (p<.00). Program affiliation, type of respondent, and time in program had inconsistent interaction effects across three latent classes (p<.05). For instance, in class 1, mentors’ ratings were higher at the midpoint than endpoint, whereas, in class 3, their endpoint ratings were higher. Thematic analysis suggested a mismatch between program expectations and accountability driven reality of the classrooms. A mentor at the end point said, “…with the push for more accountability in schools, when students miss class instruction, teachers and administrators get stressed out. When the GSU students can't get the group time or classroom time they need for a TIP project, they stress out”. Another mentor recommended, “...coordinating with school officials about realistic expectations.”

Significance
Results showed that both stakeholders had a similar understanding of program goals and learning outcomes, which was valuable feedback for program coherence. Overall, respondents seemed to provide high ratings (57% belonged to class 3), which indicated that mentors and candidates alike found the programs effective in emphasizing these learning outcomes. The outcomes focusing on analysis of educational policies and providing interaction with local and global communities were rated the lowest in all 3 latent classes, while the outcomes related to appreciating individual differences among students and creating engaging learning communities were rated the highest.

Authors

• Ruchi Bhatnagar, Georgia State University

• Jihye Kim, Kennesaw State University

• Kim Barker, Georgia State University

• Carla Tanguay, Georgia State University

• Marissa Ball, Georgia State University

• Joyce E. Many, Georgia State University