At New York this week, I repeated a bit of strategic advice I offer now and then:

Those of us in the political analysis industry love nothing more than slicing and dicing the electorate into its constituent parts and divining via polls and election results which are moving where at what velocity. That is often the best way not only to predict future elections, but to understand their implications, and also to evaluate political parties as coalitions.

But it’s easy to get carried away with such distinctions, and act as though this or that “key” group literally holds the key to victory. In the end (with an exception I will get to in a minute), a vote’s a vote, and candidates who do poorly in a “key” constituency can make it up elsewhere. Indeed, it’s especially dangerous to pretend that winning some voter group matters most; sometimes losing a group by less than the expected margin is just as important. For example, the conventional wisdom is that Democrats made big gains in the 2018 midterm by winning college-educated white voters (who leaned Republican in 2016). But it was also important that Democrats cut their margin of loss among non-college-educated white voters from 37 percent in 2016 to 24 percent in 2018 (according to exit polls).

There are times, of course, when harping on one group makes sense because polls are underestimating their numbers (one reason white working-class voters have gotten so much attention since 2016, when polls clearly under-sampled them), or have ignored them altogether as a distinct group (some polls and analysis lump together disparate voters with imprecise categories – are voters under or over the age of 45 really a “group”? – or failure to make obvious subdivisions such as by gender, or by the various identities of “non-white voters.”).

“In the most recent polls, white college graduates back Mr. Biden by a 20-point margin, up four points since the spring. It’s also an eight-point improvement for the Democratic nominee since 2016, and a 26-point improvement since 2012.

“Mr. Biden has also made some progress toward redressing his weakness among younger voters. Voters ages 18 to 34 now back Mr. Biden by a 22-point margin, up six points from the spring and now somewhat ahead of Hillary Clinton’s lead in the final polls of 2016….

“Remarkably, Mr. Biden still leads by seven points among voters 65 and over in the most recent surveys, despite the kind of racial unrest that led many of these voters to support Republican candidates at various points in their lifetimes.”

In other words, there are multiple paths to a popular-vote plurality nationally and in any one state. But that does call to mind the biggest exception to the doctrine that a vote’s a vote. The Electoral College makes huge numbers of voters irrelevant in presidential elections, and reduces the influence of various groups who are or aren’t situated in battleground states. The single biggest reason for the recent focus on white working-class voters is that they are disproportionately represented in the Rust Belt states where Donald Trump pulled his 2016 upset. Conversely, even though Latinos are the fastest growing racial/ethnic group in the electorate, their clout in presidential contests is reduced by the large number living in states that have not been competitive recently (Arizona, California, Illinois, New Jersey, New York, and Texas). If, as some Democrats hope, Arizona and/or Texas do become competitive this year, you will hear a lot about Latino voters in the aftermath.

But even in battleground states that are easy to stereotype, there’s a lot going on under the surface. Was Trump’s 10,704 margin in Michigan in 2016 attributable to underestimated white working-class voters, or low turnout among African-Americans, or a late minor-party trend among younger voters? You can make a case for any of those propositions, or for any number of combinations of them. So beware over-simplification.