(This October poll also nailed the actual 4 point spread in the congressional popular vote.)
Pew is of course the gold standard in polling organizations.
> The finding that a low response rate leads to substantial bias on some topics (e.g., volunteering) but not others (e.g., partisanship or religious affiliation) underscores the importance of having high response rate in-person surveys, which make such knowledge possible.
The above quote is saying “we can’t get accurate detailed demographics via phone” but using positively-oriented PR speak to do it. This is another way of admitting “in detail our demographics do not match the underlying population.” This could mean that their blue skews young, their red skews old, or their blue skews uneducated or vice versa, etc.
The article is from 5.5 years ago when they were just transitioning from landline polling to cell phone polling. Call spam has increased significantly since then.
Pew was founded in 1990 to do telephone research so of course they were going to defend it, but 2017 is when questions were just starting to be asked about it (in the context of a shift to mobile).
After all, the jokes about no politics at Thanksgiving has a lot to do with much more conservative old people having different ideas than the youth
And then to the right as they get older. A surprising number of British politicians on the right were Marxists in their youth.
> Now, though, there is no confusion: Generation X is safely Republican. One model from 2014 measuring only white voters through the 2012 election shows those born in the mid-to-late 1960s being the most Republican-leaning of all, more so than the older Boomers and Silent generation.