Here's an interesting article from the Washington Post:
https://www.washingtonpost.com/politics/2022/09/30/how-read-polls-like-an-exper t-or-least-not-like-newbie/
(
https://tinyurl.com/2uuchtaf)
The author asked four different people who analyze polls for a living what
dos and don'ts they would tell laypeople to keep in mind when looking at poll result.
The one answer they all had in common was to keep in mind that polls are not precise instruments. They cannot forecast the future and are intended to be representative snapshots of public sentiment at the time that the poll was taken. This is especially true of close races.
Some other pieces of advice were:
* Double the margin of error.
* If polls disagree, the reason for that disagreement might be interesting in itself.
* Transparency is good. Does the pollster describe who was polled, what questions were asked, and how the raw polling data was processed? Does the pollster's polling process seem fair and unbiased?
* Don't rely on just one poll, especially if it's an outlier. Look at them in the aggregate.
* Past polling errors are not predictive of current polling errors.
* Pay attention to the source of the poll. Perhaps my favorite quote from the article is this: "In the last month of a campaign, every candidate is just trailing his or her opponent, a deficit that can be all-but-erased with your $10 recurring contribution."
Jeff.
--- Mystic BBS v1.12 A46 2020/08/26 (Raspberry Pi/32)
* Origin: Cold War Computing BBS (1:387/26)