Hunter College, Arizona State, and WestEd used Evidano to validate findings on teachers’ learning four to five years after math PD

Studies of professional learning often stop shortly after the program ends, which may explain why findings on PL effectiveness conflict. This study asked what long-term impact PL has on classroom practice four to five years later, and what teachers attribute their sustained uptake to.

The team compared three PL programs along the adaptive–specified continuum. Adaptive models build goals and resources from local context and teachers’ shared classroom artifacts; specified models provide published materials with defined teacher learning goals and resources. The paper, titled “Understanding Sustained Teacher After Participating in Mathematics Professional Learning,” appears in the ICERI2024 Proceedings.

Eighteen case study teachers who had taken part in one of the three programs were followed up four to five years later. For each teacher, the team collected six classroom videos and conducted four semi-structured interviews about current practice, beliefs, and their PL experience.

Teachers also discussed video clips they had chosen as evidence of what they had taken up from the PL.

Two researchers first coded the data independently for evidence of content, pedagogy, and resources from the PL, then met to discuss the salient themes.

Next, the team entered the transcripts and research questions into Evidano. It identified key themes, answered the research questions for each participant and for groups of participants, and backed its answers with direct quotes, which the researchers checked for bias.

The team used these outputs to validate its own conclusions and identify anything it had missed, so Evidano served as a check on human analysis rather than a replacement for it.

Teachers who took part in specified PL were still using specific content, resources, and pedagogy four to five years later, in ways aligned with the program’s goals.

Teachers from adaptive PL emphasized the community and collaboration it built, but were less clear about specifically how their instruction had changed.

The authors argue that understanding these differences can help teacher leaders, districts, and teachers choose PL models.

This case shows a clear validation workflow: independent human coding first, then Evidano answers to the research questions, grounded in quotes, used to test the team’s conclusions and surface oversights.

For education researchers working with long-term, multi-interview case studies, it shows how per-participant and group-level answers can make cross-case comparison faster without giving up human judgement.

The study follows 18 teachers across three programs, so it illustrates patterns rather than estimating effects for all PL models.

This summary is based on the paper’s published abstract; the full proceedings paper is available to IATED library users.