Modeled Scores

The RNC data co-op assigns every California registered voter a set of modeled probability scores in the range 0.00–1.00. Each score is the model's estimate of how likely the voter is to fit a given description (e.g. supports the Republican party, will turn out in a general election). Higher values mean higher predicted probability; the underlying inputs are voter-file attributes, voting history, and commercial demographic data.

Each chart below shows how the matched cohort distributes across ten 0.1-wide buckets — for example, the 0.80–0.90 bucket counts voters whose score falls in that range. A right-skewed shape means most of the cohort scores high on that metric; a left-skewed shape means most score low; a flat shape means the cohort is mixed.

Caveats: these are model outputs, not source-of-truth attributes. A voter with a Republican Party Score of 0.80 hasn't necessarily registered Republican and isn't guaranteed to vote that way — the score reflects the model's estimate, which can be wrong on individual voters. Treat them as relative ranking signals across a cohort, not absolute predictions about any one person.