I love the media reaction to polls of "adults" as though they mean something. First, at least 30 percent aren't even registered to vote. Of those that are, half don't even bother voting. Even fewer in most mid-terms.
So right off the bat in a poll of adults 65 percent of the sample aren't going to show up on Election Day, so they can't even do anything about their opinion.
But 35 percent of your sample vote. That means something, right?
Of that 35 percent, about a third are going to automatically inclined vote for you because of party label and a third will automatically inclined to vote against you for the same reason.
So in reality, when the media poll "adults" only about 12 percent of them actually affect an election.
That's why, whenever possible, I cite the portion of the poll measuring the opinions of unaffiliated likely voters. THAT'S where the meat is.
When you see a news article bleating you can't do the right thing because a poll of "adults" or even "registered voters" disagrees, toss it in the trash.