Comparative Thematic Analysis — Cross-Group Thematic Analysis

Standard thematic analysis produces a unified thematic account of a single dataset or undifferentiated participant group. Comparative thematic analysis maintains group membership as a structuring variable and explicitly interrogates where themes are shared and where they diverge. The comparison is a core analytic and interpretive move, not a secondary observation.

Both approaches are defensible and the choice depends on your epistemological stance. A single unified codebook developed deductively (e.g., from theory or policy categories) supports direct comparison but risks imposing one group's logic onto another. Parallel inductive codebooks developed separately for each group and then harmonized better preserves each group's particularity. The latter is preferred when groups may frame the phenomenon very differently.

There is no universal rule, but each comparison group needs enough participants to support credible within-group coding — typically at least four to six, and preferably more. Groups with fewer than three participants produce patterns too thin to anchor a comparative interpretation. The goal within each group, as in standard thematic analysis, is data sufficiency for the research question.

Yes. Any rich text data — policy documents from different jurisdictions, media coverage from different outlets, social media posts from distinct communities — can be analyzed comparatively using thematic analysis procedures. The comparison units need to be meaningfully distinct and the data within each unit needs sufficient depth.

The most effective write-ups are organized around comparison themes rather than around each group sequentially. Each main section of the findings should address a theme or dimension of comparison, presenting evidence from multiple groups side by side and making the interpretive argument about convergence or divergence explicit. Avoid siloing each group's findings and then bolting on a brief comparison at the end.