[daveaiello]

When FiveThirtyEight asserts that voters don't think either party deserves to govern, can we reasonably assume that their survey respondents are actually voters?

If you drill through the link in the second paragraph, to the discussion of the composition of the Ipsos KnowledgePanel, https://www.ipsos.com/en-us/solutions/public-affairs/knowled..., it says:

"Built on a foundation of address-based sampling (ABS), KnowledgePanel provides a statistically valid representation of the U.S. population as well as many under-researched and often harder-to-reach populations, such as..."

One of the group sampled is "Young adults (ages 13+)".

You have to be 18 years old and a U.S. citizen to be a voter in the United States (see U.S. Constitution, Amendment 26.).

[cameron_b]

I am always disappointed in a survey's methodology when they try to assert their "truthiness" and explain their methodology.

The part that got me is the note about weighted responses to match the population. That needs mathematical explanation if its to be included, otherwise just say it's graded on a curve of "I'm pretty sure this response is more important."

Toward the article and similar authors:

Surveys are hard, sure but don't hide a 2006-respondent note and assert it actually represents anything. Sure I agree with the top line assertion, but I don't think this is reporting, it's just an opinion piece with a larger expense account.