There's a big debate about quality of election polls. Both sides have good points, and a track record of success and failure. But in too many election races these days, where races are competitive, polling quality can be poor and partisan.
In November 2021, a handful of pollsters saw a red wave coming, and a potential win for Glenn Youngkin in his bid to become Governor of Virginia. Network and partisan polls missed by a mile. That November 2021 red wave crested not only in Virginia, but in both New Jersey and New York rattling the Democratic establishment.
I don't mind election forecasters getting an election wrong, as long as it's within reason. What's wrong with election industry polling is threefold:
1. Too many pollsters don't poll competently, missing key demographic segments, then the political establishment floats these published polls as supposedly credible evidence of success for their political side; hijacking the narrative on false data.
2. There is a partisan polling and corporate agenda problem, and can often see this in network and cable news polls that overstate support for one side or another. Usually in Democrat races, like Fox News is notorious at doing.
3. Aggregated polling (like Real Clear Politics, 538) references frequent Democrat bias pollsters to project a generic ballot average that isn't likely going to become reality.
Based on track record of election wins, or near wins, I'm more confident in polling from Trafalgar, Rasmussen Reports, People's Pundit, Emerson, and a few others, who either nail elections near right on, or forecast it close enough. These pollsters take the time, money, and care to tap into the electorate, and report their polling that usually stands up as fairly credible once the election is over.
After Labor Day, more honest polling will better narrow 2022 Midterms down, as pollsters shift from registered voters to likely voters in their samples. In the polling industry, everyone knows likely voters are more likely to partipate in elections, and this shift in likely voter polling for the general election season could add 1-3+ points to a pollster spread.
On Presidential approval, President Biden has enjoyed some rough polling trends for the last year, with approval ratings often in the 30% range.
Rasmussen Reports saw Biden's approval improving before anyone in the last 2 weeks, which is typical for Rasmussen, and they've been spotting trends first for years. As of Friday, Rasmussen has Biden's approval at 44%. This is important because a President's approval rating historically correlates to how well the President and his party does in Midterm or in Presidential elections.
With 2-1/2 months until Midterms, today's polling averages tell us Republicans are likely going to take control of the House, but Democrats are favored to win the Senate.