As a general rule, you should mistrust polling percentages until you've looked at the crosstabs and examined the methodology. When I was doing this stuff for a living, the only numbers I relied on were raw totals from past precinct reports and I was usually able to get pretty damn close to the actual election day outcome. That's how I was able to determine months in advance what neighborhoods to canvass, how much money I'd need to raise, etc.
A percentage just gives you a snapshot of sentiment, it doesn't tell you whether someone is going to vote, which a lot of people don't even know themselves until the day of. If they've voted in every election for 10 cycles however, that's much more reliable information. And of course the EC makes most national presidential polls pretty worthless for predicting outcomes.
Sounds really interesting, that would have helped a lot when I was volunteering full time for Bernie and had to deal with campaign staff post-Iowa who were fumbling with assigning canvassing turf (mostly because of the dumbass distributed organizing the campaign decided on). Any interest in making a post in c/electoralism about this, like a 101 type thing? Or at least describe what one would need to learn to do that?
As a general rule, you should mistrust polling percentages until you've looked at the crosstabs and examined the methodology. When I was doing this stuff for a living, the only numbers I relied on were raw totals from past precinct reports and I was usually able to get pretty damn close to the actual election day outcome. That's how I was able to determine months in advance what neighborhoods to canvass, how much money I'd need to raise, etc.
A percentage just gives you a snapshot of sentiment, it doesn't tell you whether someone is going to vote, which a lot of people don't even know themselves until the day of. If they've voted in every election for 10 cycles however, that's much more reliable information. And of course the EC makes most national presidential polls pretty worthless for predicting outcomes.
we love our statistically literate comrades dont we folks
Soooo.... what is your expectation about the current election?
Sounds really interesting, that would have helped a lot when I was volunteering full time for Bernie and had to deal with campaign staff post-Iowa who were fumbling with assigning canvassing turf (mostly because of the dumbass distributed organizing the campaign decided on). Any interest in making a post in c/electoralism about this, like a 101 type thing? Or at least describe what one would need to learn to do that?