When election forecasting, I think one needs to look at LV or likely voters and RV, registered voter polls. RV or registered voters are 20% less likely to go to the polls and vote. Look at the RCP generic congressional polls. Their average puts the democrats up by a slim 0.3 points

https://www.realclearpolitics.com/epolls/other/2022-generic-congressional-vote-7361.html

Now look at just the RV or Registered voter polls. There’s 6 of them and everyone of them have the democrats in the lead except one, which is a tie. The Economists, Political, NBC, FOX, NY Times, Harvard-Harris. Then look at the LV, Likely Voters which are 20% more likely to vote than Registered voters. 5 of them which all but one has the Republicans in the lead. The exception is a tie. ABC, CBS, Emerson, Rasmussen, Trafalgar.

The bottom line is LV or likely voter polls are more accurate in predicting elections than RV or just registered voters. For most, the difference is just a mundane stat. But LV is a much more accurate indication on how an election will turn out than RV. Once you get within 2-3 weeks before an election, you throw out all RV or registered voter polls as they’re much less accurate.

This is another reason why the Democrats may be leading in the generic congressional ballot, but projected to lose 14-16 house seats.

It's high past time that we start electing Americans to congress and the presidency who put America first instead of their political party. For way too long we have been electing Republicans and Democrats who happen to be Americans instead of Americans who happen to be Republicans and Democrats.