Conclusion
The matched CCES can stand in for the ANES on this vote question.
Rahel
That is the article’s conclusion. The CCES is a cooperative internet sample, meaning many research teams share one online survey. Matching pairs each volunteer with a similar person in Census data. Weighting makes some interviews count more than others. After both steps, the CCES found the same party–vote link as the ANES, the face-to-face survey. So teams can share one large file. They can study states, districts, and groups that a 2,000-person study cannot measure well.
The condition
The result holds only for the processed sample. It also holds only if volunteering adds nothing about the vote once the matching traits are known. The raw panel is still skewed. A bias unrelated to those traits is still in the data.
Rivers founded the firm YouGov bought and was its chief scientist. The article says so. That is why the outside checks matter: the ANES and the certified vote, meaning the official count.
What our rerun adds
The party slope is how much the vote score changes for each step toward Republican on the party scale. Our shared party slope is −0.317 in both surveys. Our full CCES slope is −0.314. The published table gives −0.313. The state check agrees in direction and in size: +1.0 and 3.4 points here, +0.27 and 2.43 in the article. It does not reproduce those exact decimals.