What the Scores Measure

Every film we score is analyzed across five categories of on-screen representation. Each category receives its own 0–10 score and a written note explaining the score, and the five feed an overall diversity score.

• Gender representation — the prominence, depth, and agency of women and gender-diverse characters.
• Racial and ethnic diversity — casting breadth and whether characters of color have substantive, non-stereotyped roles.
• LGBTQ+ representation — the presence and quality of queer characters and storylines.
• Disability representation — visible and invisible disability portrayed with dignity and specificity.
• Religious and cultural representation — how faith traditions and cultural identities are depicted.

Reading the 0–10 Scale

A low score (0–3) means a category is essentially absent or handled through stereotype. A mid score (4–6) means meaningful but limited representation — present, but peripheral to the story. A high score (7–10) means the category is woven into the film with depth: characters who drive the plot, written with specificity rather than as tokens. Scores describe what appears on screen; they are not a judgment of a film’s overall quality.

How the Analysis Is Produced

Each analysis starts with an AI model that evaluates a film using its full metadata and reference material about the production. The model produces category scores, written justifications for each category, overall strengths, and areas for improvement.

Limitations

Representation is contextual, and reasonable people disagree about it. Our scores are an interpretive starting point for discussion, not a definitive verdict. AI-assisted analysis can miss nuance. A film made in 1955 is scored on what is on screen, not graded on a curve for its era; context like that belongs in the written notes, and we encourage reading them rather than relying on the numbers alone.

Disagree With a Score?

If you believe an analysis is wrong or missing important context, contact us with the film and the issue. Flagged analyses go back through review.