Biden's spike in the polls is well-documented. His current popularity with black voters is striking as has been noted. Less has been written about how he's doing with white noncollege voters, but the results here are also of significance. Harry Enten notes in his latest CNN column:
"Whites without a college degree still make up a substantial portion of the Democratic Party. Despite much noise about how whites with a degree are the future for Democrats, each group is about 30% of the party. Nonwhites, both with and without college degrees, are about 40% in total. [Note: these figures agree with my own States of Change data.]....
Our latest CNN national poll shows Biden is strong with non-college educated whites, as well [as with nonwhite voters). Biden jumps from the low 20s among whites with a college degree to the mid 30s among whites without a college degree. Comparably, his lead over Sanders increases from only 5 points over Sanders among whites with a college degree to about 20 points among whites without a college degree. Sanders polls in the mid 10s with each group....
Biden's ability to breakthrough with non-college whites marks perhaps the biggest difference between the 2016 and 2020 primaries. In 2016, Clinton actually lost white voters without a college degree in the primary. Sanders beat her by 7 points among this group in the median Democratic nomination contest with an entrance or exit poll, even though she won college educated whites by 7 points in these same contests."
This pattern clearly could help Biden secure the Democratic nomination. But does this mean Biden is the most "electable" nominee against Trump?
Not necessarily and for two reasons, which are frequently confused. One is that superior appeal to one group of voters does not mean that more voters may not be lost by inferior appeal among another group. Using Biden as an example, perhaps whatever he gains among white noncollege voters relative to other potential candidates will be more than counterbalanced by losses from unenthusiastic support and turnout among young voters and possibly some other Democratic constituencies.
Therefore, electability is always and inevitably about trade-offs. No candidate will be without them so people should be asking: who has the highest net, given potential gains and losses, and therefore the best chance of beating Trump overall and in the states that are likely to count the most. This is always hard to do given data limitations but that is really the argument we should be having.
The second reason to be cautious about Biden's electability is that, even if one is taking proper cognizance of the various factors that may increase or decrease a candidate's appeal, it is hard to make these judgments and harder still this far in advance of the actual general election campaign, when data are sparse and less reliable. Put more bluntly, a lot of early judgments about electability, by voters and pundits alike, turn out to be wrong.
Does that mean we shouldn't care about electability and, since no one knows anything (as they say in Hollywood), just roll out who we like the best and hope they win? I don't think so, especially given the 2020 stakes. But we do need to think about it in a sophisticated and, to the extent possible, in a data-driven way.
That's it for my plea for rationality. Back to the usual heated and intemperate debate who's the best candidate!