+++ title = "Resources" +++
These scenarios when the election could be called if swing states are called in the same order as in 2020.
The Pew Research Center reports on how public polling has changed in recent years
-
61% of national pollsters in the U.S. used polling methods in 2022 that differed from the methods they used in 2016, and pollsters also made changes to their methods after the 2020 election.
-
In 2022, 17% of national pollsters used at least three different sampling/interviewing methods, up from 2% in 2016.
-
Use of live telephone polling has declined rapidly since 2012. In 2022, 10% of national pollsters used live phone as their only method while 32% used it alone or combined with other methods.1
-
Online polling grew tremendously between 2006-2020, but that growth appears to have slowed or stopped in 2022 for unclear reasons.
-
Use of probability-based online panels recruited from address-based sampling has become more prevalent, tripling from 2016 to 2022.
-
Some newer methods like text message polling and registration-based sampling have seen upticks in usage in recent years.
-
Overall, the report finds the polling industry has entered a period of unprecedented diversity in sampling and interviewing methods compared to the live telephone polling that dominated in the early 2000s. However, whether these changes will translate into improved accuracy remains to be seen.
<img src="/assets/img/turnout.png" style="width: 100%; display: block;">
Julia code used in the preparation
- Shirani-Mehr, H., Rothschild, D., Goel, S., & Gelman, A. (2018). Disentangling Bias and Variance in Election Polls. Journal of the American Statistical Association, 113(522), 607–614.
- Cognitive bias
Footnotes
-
Contemporary reluctance to accept unscheduled phone calls or phone calls from unknown numbers may be a contributing factor. ↩