I put together a handy reference chart of the differing demographic makeup of the electorate across various battleground states, to give an idea of how different groups matter more in different places. This is derived from the Census Current Population Survey questionnaire after the 2020 election (like exit polls, it has its own methodological issues).
I’m using the racial categories offered by the U.S. census, plus adding a line that subtracted white, black, and Asian voters from the total to identify the remainder (a number that tends to be larger in Western states with major Native American populations).
As you can see, there are some very large variations: Hispanic voters are a very big factor in Florida and the Southwest (from Texas to Nevada), but a much smaller group elsewhere. Black voters remain heavily concentrated in the old South, but also in Northern cities. For example, Pennsylvania and Michigan have black populations three times the size of those in Wisconsin and Minnesota. That’s one reason why a shift or falloff among African-American voters could have a big impact in two of the “Blue Wall” states without making a dent in the third, Wisconsin.
Many of these constant patterns are already apparent in early-voting data. But there are signs of shifts. In Georgia, where nearly 3.5 million people have voted already, the gender breakdown is marginally more female, but black turnout appears to be strikingly down so far, at just 26 percent. (Then again, the data we have lists 0.8 percent of the vote as unknown gender and 9.1 percent as other/multiple/unknown race, so take all of this with a grain of salt.)