Transparency
Methodology
Informed Texan does not endorse candidates.
Candidate alignment measures similarity to answers you provided. It does not represent Informed Texan's opinion.
Your politics are not for sale.
How candidate positions are determined
Researchers collect public sources — voting records, official statements, campaign materials, debate and forum transcripts. Each position is placed on a 0–4 issue scale with a written summary, a confidence label and linked sources. Positions move through DISCOVERED → AI EXTRACTED → NEEDS REVIEW → VERIFIED → PUBLISHED, and every edit creates a new version; old versions are preserved.
How scores are calculated
Your answers map to the same 0–4 scale. For each issue: issueFit = 1 − |your position − candidate position| ÷ 4. Overall alignment = Σ(issueFit × weight) ÷ Σ(1 × weight), shown as a percentage, with each issue shown as 0–5. Weights: Not important 0, Somewhat 1, Important 2, Very 3, Critical 4. This math runs in your browser with ordinary code — no AI.
Dealbreakers
Dealbreakers are never hidden inside the score. If a candidate's position is two or more steps from yours on a dealbreaker issue, we display an explicit conflict notice next to the score.
Evidence confidence
High: multiple consistent primary sources such as votes. Medium: clear statements in one strong source. Low: indirect or dated evidence. Unknown: no documented position. Unknown positions are excluded from the calculation — never treated as zero — and lower the coverage figure shown beside the score.
Standards & Conduct
Conduct is reviewed separately from policy fit. You choose which standards matter; we show documented records with their legal or factual status. We never declare that a candidate fails a standard.
Source selection
We prefer primary sources: official records, court documents, filings, transcripts. Secondary reporting is used for context and labeled as such.
News clustering
Articles are normalized, deduplicated and grouped into one story per event using headlines, entities, timing, similarity and source links. One synthesis is written per story, not per outlet.
AI usage
AI helps extract and summarize evidence asynchronously, only when a source materially changes or an editor requests it. Results are cached and reviewed by people. Page views, scoring, sorting, filtering, saved results and printed guides never use AI.
Corrections
Anyone can request a correction. Editors review each request; changes are logged publicly with the date.
Funding & independence
Informed Texan is free and nonpartisan. We accept no political advertising and no money from campaigns or parties, and we do not sell data.