Divergent Thinking Scoring

Divergent thinking scoring applies response normalization and category coding to measure fluency counts, flexibility categories, and originality thresholds. This framework ensures creativity assessments remain valid and comparable across raters.

Semantic distance scoring computes cosine distance between response embeddings using validated models to quantify originality. This automated procedure provides an objective originality assessment metric for Alternative Uses Task datasets.

Inter-rater reliability ICC calculation requires rater training protocols and applying reliability benchmarks to category coding. The scoring framework enforces ICC targets to ensure consistent flexibility and originality evaluations.

Fluency confounds originality scores because higher response counts increase the likelihood of generating original ideas. The framework counters this confound using ratios, top-N scoring, or covariates to separate true originality from sheer output volume.

Automation-ready divergent thinking scoring uses validated embedding models to compute cosine distance for semantic originality. This approach includes reporting checklists and documentation references to ensure reproducible creativity analytics.

Response normalization standardizes participant text before category coding and originality thresholding. Applying normalization procedures early ensures that fluency, flexibility, and semantic distance metrics are calculated on clean, comparable data.