Backtests against published survey data.

Pew Research — How Americans View Weight-Loss Drugs

Predicting public opinion on GLP-1 drugs across demographic and partisan segments.

For this backtest, Intera builds a synthetic U.S. adult population calibrated to the demographic profile in the Pew survey on weight-loss drugs, including age, gender, income, and region, plus self-reported party identification. The model is asked, before seeing results, to predict support, concern, and willingness to try GLP-1 medications for weight loss by segment. Predicted response distributions are then compared to observed survey opinion to evaluate whether the synthetic population captures the right directional shifts across partisan and demographic lines.

Source: Pew Research Center, Feb 26, 2024

KFF Health Tracking Poll — Knowledge and Views of Medication Abortion

Predicting stated views on medication abortion access across partisan and demographic groups.

Intera constructs a synthetic panel that mirrors the KFF tracking poll sample, with conditioning on age, gender, race and ethnicity, income, and region, alongside partisan identification and abortion self-identity. The engine is then tasked with predicting support and opposition for medication abortion, perceived safety, and views on specific access restrictions by segment. We compare predicted opinion distributions to the KFF topline to assess whether the synthetic population anticipates the observed partisan and demographic splits in comfort, access preferences, and uncertainty around medication abortion.

Source: KFF, Dec 5, 2025

KFF — Health Information Trust and COVID-19 Vaccine Attitudes

Forecasting trust in health information sources and vaccine sentiment by segment.

In this backtest, Intera simulates how different demographic and partisan segments trust doctors, public health agencies, and political actors as vaccine information sources. The synthetic population is conditioned on age, race and ethnicity, education, region, and partisan identification, then asked to predict trust levels and likelihood of receiving a COVID-19 vaccine in the coming season. We evaluate how well Intera matches KFF's observed patterns — for example, higher trust in personal physicians across parties and diverging views on CDC guidance and vaccine uptake among Democrats, Republicans, and independents.

Source: KFF, Aug 1, 2025