In 2012, The Post is featuring an expanded look at horse-race election polls, referencing some newsworthy polls that veer from survey best practices. The new approach allows a fuller look at all the polling, but also provides key guideposts on reading polls. One of these is a new “Post preferred” label that we use to highlight polls that meet a high standard for quality.
The Post thoroughly vets each survey we report (and don’t report) for its methodological rigor before featuring its results. We sweat the small stuff, including sample design, question wording, interviewing techniques, likely voter models and weighting schemes (the statistical adjustments pollsters make to their data). Post preferred polls meet the highest standards for reliable survey quality. Independent sponsorship also matters, as partisan polls have a record of systematic bias. Most Internet polls lack random samples and fall below this threshold. Other telephone polls that may be solid in one aspect may not pass muster on another.
Now, when you use The Post’s Politics App for iPad to see the latest state polling, you’ll see a label next to each poll we’ve designated as Post preferred. We’ll also reference the highest-quality polls when they are available, as always letting you know when a poll has an important weakness.
Why differentiate between poll quality? You should have access to the best possible information on who’s up and who’s down, but you should also have information on when to be cautious. Polls that meet The Post preferred standard employ more reliable methodologies.