People say the VP choice doesn't matter, per research. That seems to be true, to an extent, but that research is limited along several parameters, sample size being a primary one. I have sympathy for the argument that we care too much, but also for the argument that it matters.

We just don't have that many elections to look at in the modern era. This is especially true when incumbency is taken into consideration. Regardless, though, Biden is on the older side. He's also a white guy. These factors combined do mean the VP choice might matter more.

All that said, I think Democrats have it within their power to win this election no matter what. I'm not saying the VP choice will necessarily be a tipping factor. But, we're talking about a close race around the margins.

Also, and I think this is somewhat important, a lot of "Did the VP make a difference?" literature may be somewhat incompatible with our current considerations, which mainly concern winning the midwest and, more importantly, boosting black & latino turnout.

We haven't had a black or latino/a VP candidate in a pres race. So, research on how a specific VP helped w/ other demos or their home state might not be pertinent. Whether or not Jon Edwards helped in NC may be irrelevant to whether Val Demmings would boost black turnout overall.

Again, this all comes back to sample size. We have very few data points. We have even fewer when we consider political polarization. & so many demo factors vary between elections. I'm not saying the VP choice is definitive. But I don't think you can say it's meaningless, either

And, I'm not an expert, but a lot of the VP research concerns states. It's more difficult to look at something like how Joe Biden (paired w/ Obama against McCain/Palin) helped w/ white working class compared to Edwards (paired w/ Kerry against Bush). Too many variables.