About

Carmena’s Overton & Electoral Research is a new, non-partisan US research firm. We model elections, behaviour and the Overton window, the range of ideas the public treats as acceptable at a given time.

Our goal

The most accurate predictive modelling and research for elections, behaviour and the Overton window.

That is a goal, not a claim. A new firm has no record to point to, so we will build one in public. We will publish each forecast, with its range, before the result is known, and score it against the result afterwards, whether it went well or badly. The 2026 midterms are the first test. The scores will be on our track record page.

What we work on

Elections. Forecasts and turnout analysis, starting with the 2026 midterms: the early vote, the party mix of early voters, and new versus habitual voters, from state records where they exist.

Behaviour. What people do, not only what they say: who turns out, who switches, who is new. For example, Democratic early voters in Texas are twice as likely as Republicans to be new voters (8% against 4%).

The Overton window. Tracking which ideas move from fringe to acceptable to mainstream, and how fast. We have not published work in this area yet.

How we work

Every chart names its source in one line. Every modelled figure is labelled as a model. Every caveat that changes the reading travels with the chart. The methodology page explains each method in full.

Team

Team profiles are not yet published.