Blue or Red?

Every answer is a token, like a word in a bag-of-words classifier. Each token carries one weight, positive for blue and negative for red. Your score is a baseline plus the weights of the answers you gave, passed through a logistic function to get a probability. Skipped questions add nothing.

The weights come from a logistic regression fitted to 6,075 respondents in the ANES 2020 Time Series who reported voting for Biden or Trump, using the survey’s post-election weights. Two constraints keep it readable: weights move in one direction along each answer scale, and training randomly hides answers so that the live estimate stays calibrated when only a few questions are answered.

Limits: the model only knows the 2020 two-party electorate. Non-voters and third-party voters are not in the data, and party coalitions have shifted since. It predicts from statistical association, so read the result as a description of who tended to answer this way, not as a judgment of anyone. Your answers stay in this page and are not saved or sent anywhere.