The civic nationalists want to believe it. A part of me would like to believe it, too. I've mostly shed that part of myself over the last couple of years as it has become obvious that Trump's Authentic American (whites and blacks) vs Fake American (invaders) paradigm isn't going to materialize, but I'm a pragmatist. If we can believe it, let's use it.

Believe what? This:

Trump at 36 percent approval among African-Americans, new poll finds

Rasmussen's approval tracking has consistently been better for Trump than others have been, but the differences are often exaggerated. We're generally looking at about 5 points better from Rasmussen than the rest of the field. Reuters-Ipsos is my go-to even though it has a moderate left bias because it's free, user-friendly and allows all kinds of fun filters for cross-tabbing purposes.

For the 2016 presidential election, the last poll Rasmussen put out had Clinton +2 in the popular vote while the last R-I poll had Clinton +3. The official results were Clinton +2.1, so these two outfits were the cream of the crop.

R-I currently has Trump's approval among black voters at a far more plausible 14.6%. The monthly high point came back in May, when it hit 16.1% on the strength of Kanye West and, by extension, Candace Owens.

One thing that immediately becomes apparent from looking at the month-by-month figures is that the gains in approval have come entirely from picking off "mixed feelings" respondents. The "disapprove" figure for blacks hasn't budged. It's never dipped below 80%. Four out of five blacks are knee-jerk Never Trumpers. It's the Sailer Strategy or bust for the GOP.

Trump's double-digit approval among blacks isn't translating into better prospects for GOP pols in November, either. Just 7.8% of registered blacks saying they intend on voting Republican in their districts' congressional elections.

The following graph shows R-I's results on the expressed electoral intentions of registered voters six weeks from now in two-way races with "don't know" and "will not vote" responses excluded (N = 40,925--another reason R-I is great!):

Every time I look the marriage gap among women gets wider than before.

Without scaling for the size of each demographic group, it initially appears as though an impending blue wave is cresting. R-I estimates, however, that whites will account for about 73% of all congressional votes cast in November. Asians and Hispanics don't vote at rates anything like that of whites and blacks, especially in non-presidential years.

If the GOP could manage to get whites at a 2-to-1 margin, they'd scarcely need a single non-white vote to win, enact an immigration moratorium, repatriate all non-citizens living in the US, get back to making babies of their own, and Make America Great Again.

That said, back in the early Spring I estimated that Democrats would come out of the November mid-terms with a 15-seat advantage in the House while Republicans would retain the Senate. Despite taking flak from people smarter than myself for the prediction, here we are several months on and I don't think that prognostication is in need of any calibration.

Parenthetically, regular prolific commenter Feryl is now running a blog. It is up to this point primarily GSS outputs with limited accompanying commentary. We cannot have too many people look at the primary source data, so consider paying it a visit.