Sorry, forgot to reply to this.
I'm not sure how much they ever really matter
And, while one might argue that she won three million more votes, it was also an irrational certainty in polling data which prompted the Clinton (and now Warren) team of geniuses to ignore their people on the ground in Michigan and Wisconsin and unsuccessfully reach for Iowa (whose local staff had long since reported was a lost cause).
Polls in isolation are nearly useless in a system as convoluted as America's. They have some value in tracking momentum, but they do not generally measure turnout rates - which these days is arguably the most important factor - and they also tend to under-represent younger voters and non-habitual voters.
When somebody quotes polls without considering their context, it is a sure sign that they do not understand politics (not referring to you, I hasten to add!) So, for example, the polls that showed Biden ahead in Iowa all year were nearly worthless because virtually no voters were actually paying attention yet. When they started to tune in, about a month before the caucus, we saw a sudden surge for Sanders, which is also what happened in 2016. The polls mostly caught that, which is where their true value lies. But to regard polls from, say, July 2019 which showed Biden in the lead as 'data' or 'evidence' of anything was always silly. Basic observation and inductive reasoning made it painfully clear that he is a blubbering, incoherent mess who will severely depress turnout. It is telling, if not surprising, to see posters even now, after Iowa and NH, citing head-to-head polls as 'proof' that Biden still has the best chance of beating Trump.
Looking at swing state polls is a more logical assumption, but they tend to be based on thinner and lower-quality data, and tell us nothing at this stage. It is far, far, too early - we haven't had any head-to-dead debates yet, there are still nine months of messaging and events to go, and nobody except the weirdos who frequent forums like this is really paying attention yet. Polls serve mostly to confirm what has already happened; using them as a predictive tool without very careful consideration of a range of non-quantifiable factors is not generally wise - as Joe Biden helpfully shows us.
I've posted this before, and can't recommend it highly enough:
They’re increasingly unreliable, and they shape politics in ways we may not want.
www.newyorker.com
Also, on the question of polling and electibility, I found this interview quite interesting too, if you have 20 minutes to spare. Makes it pretty clear, reading between the lines, that the Democrats intend to bomb the village in order to save it.
The political scientist Rachel Bitecofer joins Dorothy Wickenden to discuss the best way for Democrats to win the Presidency, and Congress, in 2020.
www.newyorker.com
The people who spent February 2019 parroting the cable news talking point that 'Bernie should drop out because all the candidates share his views and he can't win' would do well to heed their own advice now, if they actually want to defeat Trump.
But it's pretty clear that the catastrophic farce of a contested convention is now the DNC's main objective, and that the chasing pack is competing not to win, but for the right to serve as anointed viceroy, on pain of permanently shattering 'party unity' if not breaking up the Democratic Party altogether, and quite possibly handing Trump the victory then and there.
It's a shame that Warren supporters aren't so cultist, or they'd have long since accepted reality and demanded she call off her kamikaze mission, and do the responsible thing by dropping out and backing the only viable progressive candidate. But instead, the toxic masculinity of her army of rabid online bros is only intensifying,
as shown when one of her most vocal supporters broke ranks ; )
