Choose one debate question, define who is listening, and compare every candidate without losing the underlying evidence.
Important: this is a research-backed simulation, not polling on the debate. Ranges combine published Ontario evidence with transparent transcript coding.
Show me how this landed with a , , in , who is a .
Change any underlined phrase. The confidence range widens as the voter lens becomes more specific.
Choose a candidate in Compare candidates, then inspect regional variation. The mosaic is not riding-level polling.
Select a region to apply it to the voter lens.
Regional mosaic, sized for exploration—not geographic area or riding-level prediction.
Hover or focus a region for context; click to filter.
Select any candidate to reveal the full answer analysis and jump directly to that response on YouTube.
These segments complete the debate record but are kept secondary because they were not five-way, equally timed main questions.
Markdown · approximately 156 KB. Includes speaker-organized main and lightning-round responses, followed by the complete timestamped transcript. Machine-generated; open-debate speaker attribution is not human-certified.
No respondent rated these debate clips. Published Ontario evidence provides the population, issue, and political context; the answer dimensions are analyst-coded from the transcript and reweighted by the selected voter lens.