$100–200/hr
The role in one line
Senior political analyst authoring evidence-based election forecasts and political-risk assessments for AI model evaluation.
Written by Training Turk from the public listing; it may be incomplete or out of date. Read the full posting on Mercor.
What you would do
- Author political reports and forecasts from a specified evidence cutoff without using later information
- Make directional judgments about polling trends, win probabilities and market-implied odds with clear reasoning
- Articulate causal mechanisms explaining how events affect candidate positions and voter behavior
- Review AI-generated political analyses and grade them against your professional standards
- Document judgments and reasoning processes for evaluation of AI model performance
Who they are looking for
- Hands-on experience on specific Senate, governor or statewide races in targeted swing states
- On-the-ground knowledge of electorate, candidates and local political environment from university, campaign, pollster or party analytics roles
- Documented record of making live probabilistic judgments about races, not retrospective commentary
- Strong polling literacy and ability to analyze causal event dynamics
- Typically 8+ years direct experience or unusually strong performance evidence at equivalent level
Skills this role asks for
What the interview is likely to probe
1.Making disciplined point-in-time forecasts
Forecasting discipline distinguishes serious analysis from hindsight-biased commentary, critical for evaluating AI judgment.
Expect something like: “It's 60 days before a Senate runoff in your state. Candidate A leads in recent polls by 3 points but has low name recognition. Give a win probability for A and explain what polling, historical patterns, and voter turnout assumptions informed your estimate.”
2.Interpreting polling and evaluating uncertainty
Understanding poll quality, margins of error, and hidden assumptions separates robust forecasts from overconfident projections.
Expect something like: “You see conflicting polls: one shows your candidate up 5 points with 1000 respondents, another down 2 points with 600 respondents. How do you reconcile them and what probability range do you assign?”
3.Causal event analysis
Linking specific events to electoral impacts requires domain knowledge that separates expert judgment from pure statistical extrapolation.
Expect something like: “A major news story breaks 10 days before the election: a candidate's close advisor is indicted. How do you estimate this event's effect on candidate positioning, turnout, and win probability?”
4.Grading AI political analysis
Evaluating AI reasoning about politics requires holding AI to the same rigorous standards applied to professional forecasters.
Expect something like: “An AI model predicts a 72% win probability for a candidate based on 'strong polling' but doesn't mention margin of error, recency weighting, or historical turnout patterns. What's missing from this analysis?”
Exercise you may get
Author a point-in-time Senate forecast using provided polling and electoral data from a cutoff date. Include win probability, confidence interval and causal reasoning. Then grade an AI forecast.
How to prepare
- Document 2-3 races where you made live win probability forecasts and compare your estimates to actual outcomes to assess calibration
- Review recent polling in an active race and write your own forecast incorporating poll aggregation, recency weighting and historical patterns
- List local factors unique to your state that affect electoral outcomes: turnout trends, demographic shifts, candidate coalitions
- Study an event that changed a race trajectory you tracked and write out the causal chain connecting the event to electoral impact
Facts
- Pay
- $100–200/hr
- Commitment
- hourly
- Hours
- 20 per week
- Work arrangement
- remote · Remote
- Eligible locations
- CAN, USA
- Domain
- Life, Physical, and Social Science
- Company
- Jade
- Posted
- 9/21/2026
- Open slots
- 25