a helpful infographic that’s supposed to make some kind of clever point but also seems to imply that Pete Buttigieg is beating Joe Biden by quite a bit which if we know that’s not true then who knows what else about it is bullshit
5 Likes
I don’t think it says anything about any candidate beating another; it is only supposed to say something about a candidate’s appeal among donors in various economic strata.
However, as @DukeTrout points out, since the data come from one source (ActBlue), it really says more about who is popular with donors to that bundler. Since the vertical axis is a fraction of donor share, every high dot in one chart necessarily forces the corresponding dots in the other charts low, so if this is the natural place for people in lower economic strata to give small donations to Bernie then everyone else will have low dots for that strata for purely mathematical reasons.
The contrast of Bernie’s chart with the others is striking, but doesn’t really say anything very useful.
3 Likes
Yes, because the intuitive approach is to take the area under the curve. However, this is a scale-free distribution, because there are exponentially more low-income occupations than high-income ones. So, the poorer end of the curve is way more important than the richer end, but we can’t tell exactly how much more important because this little detail gets obscured. It’s just not a good plot for this reason.
ETA: No matter how I crunch the curve fitting, I still show support for Sanders in the 40s based on this graph. This is quite obviously not the case. What are these polls, ranked-choice? Favorable opinion?
6 Likes
Right. Sorry, I wasn’t more precise. Each dot is an occupation for that income range bucket. So you might have 5 dots for retail worker, 7 dots for carpenter, 3 dots for school teacher, etc. but only one dot per vertical per occupation.
1 Like
Buttigieg’s curve is much higher than Biden’s at every income level
so it can’t matter which side we weight more, it’s still wrong
2 Likes
Okay, it’s donors, which makes much more sense. I could easily believe that 45% of people who donate to Democratic presidential candidates donate to Bernie, all across the board, and not very many people give to Biden because he’s mostly astroturfed.
Still, if that many people give him money, why isn’t he more popular? Probably because the people voting for the others in opinion polls aren’t really giving them as much money.
This makes sense if you understand that Sanders supporters are really good about donating to their candidate. Everything after that is just lies, damn lies, and statistics.
3 Likes
9 Likes
7 Likes
Interesting thread on Gabbard’s expenses. Why is she paying this guy?
7 Likes
A friend noted that this is not a crowd that looks like Philadelphia- That she’s seen more black people at Trump rallies.
7 Likes
I posted the Civil Beat story when it first came out. There was a follow-up by CB’s editor:
However, it gave no further information about what Tulsi is getting for the money.
5 Likes
7 Likes
It also implies that, of the sample studied (donors via ActBlue), those who donated to Bernie were heavily skewed towards lower income working class occupations, in sharp contrast to the wealthier tendency for donors to all other candidates.
Which is why the data was noteworthy, albeit not surprising.
2 Likes
That’s not really how data works. If the source is biased, you can’t really draw conclusions from the data.
Don’t get me wrong, I think the premise that Bernie draws a lot more donations from working class supporters than other candidates makes sense, but the preponderance of Bernie donors donating through ActBlue could have a lot of other causes. For instance, his team could be particularly good and directing donations through ActBlue, other candidates might have other preferred donation channels, etc.
4 Likes
Pretty much all data sets are biased in one way or another; the trick is in picking a data set where the biases are orthogonal to the question.
Using this data to gauge overall support is misguided; it’s no good for that.
Using this data to analyse the demographics of donors, however, is not invalid. There is no reason to believe that there is a systematic bias that specifically excludes working class supporters of non-Sanders candidates from this sample.
Of those who donated to Democratic candidates via VoteBlue, the Bernie donors displayed a marked working-class bias that was not apparent in any of the other candidates. Unless there is reason to believe that non-Bernie working class donors are disproportionately donating elsewhere, it seems reasonable to assume that this data is not inconsistent with the overall picture.
In broad terms, Bernie has the working class, Warren has the white professionals, Biden has the elderly suburbanites. This is all pretty much as expected.
1 Like
Those expectations make sense. I’m totally with you on that.
But how do people decide to donate via ActBlue? Most people don’t just go to ActBlue.com and look at the menu of candidates, then decide who to donate to. If that were the case, we might be able to take some useful info from this data. But that’s not how it happens.
People get to ActBlue through the candidates website or online advertising. So the people who click through to donate are inherently selected through the candidates, not the other way around. Some candidates might do a better job of driving prospective donors to ActBlue, while others may do a better job at driving donations through other channels.
And Biden supporters might just mail sacks of pennies to:
Joe Biden
Delaware
USA
4 Likes
8 Likes
1 Like
https://www.phillymag.com/news/2019/10/23/kenney-endorses-elizabeth-warren/
Mayor Kenney Endorses Elizabeth Warren for President
He’s the first major-city mayor to throw support behind the Massachusetts senator.
7 Likes
