Hacker News Reader

Comment by tome | Hacker News Reader

One of the charges seems to be that the proportion of black voters chosen for "deterrence" is about twice as large as proportion of black voters in the general population. Given that 92% of black voters voted for a candidate other than Trump[1] and about 50% of all voters voted for a candidate other than Trump, it seems that simply selecting "Democrats" for deterrence explains the outcome, nothing to do with race at all.

Am I missing something here, or is this just a huge misunderstanding of statistics?

[The fact that this database exists and is used at all might be suspect but that seems orthogonal to how the numbers are reported in this story.]

[EDIT: Changed "The charge seems to be" to "One of the charges seems to be". The article contains plenty of charges (the existence of the database at all, for example), besides the statistical aspects.]

[1] https://www.mic.com/articles/159402/here-s-a-break-down-of-h...

View on HN

Controlling for party affiliation within the context of a negative ad campaign would've yielded much more interesting results. In all but Wisconsin, this paragraph in particular almost seems counter to their argument with back-of-napkin controls:

> In Georgia, despite Black people constituting 32% of the population, they made up 61% of the ‘Deterrence’ category. In North Carolina, Black people are 22% of the population but were 46% of ‘Deterrence’. In Wisconsin, Black people constitute just 5.4% of the population but made up 17% of ‘Deterrence’.

Right, I too would like to see numbers controlled for party affiliation. I suspect it would look a bit more like this

> In Georgia, despite Black people constituting 60% of Democratic voters, they made up 61% of the ‘Deterrence’ category. In North Carolina, Black people are 40% of the population of Democratic voters but were 46% of ‘Deterrence’. In Wisconsin, Black people constitute just 11% of the population of Democratic voters but made up 17% of ‘Deterrence’.

> In one confidential document seen by Channel 4 News, Cambridge Analytica admitted the Trump campaign did target “AA” (African Americans) with what it called the “Predators video” – spending $55,000 USD in the state of Georgia alone.

That's not a revelation either. The ad[1] is a series of remarks made by HRC ostensibly denigrating multiple groups, and each remark is labeled with its target: "African-American youth", "millenial voters" [this one was about Bernie supporters], and "Trump voters".

It seems pretty obvious to me (and anyone who saw it?) that an ad like this would be targeted towards the blocs that it's about, which would include black voters and young Bernie supporters, to prevent them from voting Hillary.

[1] https://www.politico.com/video/2016/10/trump-campaign-releas...

The parent comment said "it seems that simply selecting 'Democrats' for deterrence explains the outcome". The article makes it clear that's not the case.

Yes you're right, your comment was clearly in reference to this narrow claim. I didn't intend my comment as a rebuttal to yours, but a bit of a tangent: it's clear that the campaign targeted black voters (presumably among other groups), but it's not clear what's extraordinary or abusive about this. Turnout for oneself and one's opponent is a well-known target of campaign strategy, and I'm not aware of any norm against it. Nor do I think there should be one: it's an entirely legitimate choice to withhold one's vote from either candidate, and this ad simply contained remarks of Clinton's that a black (or Bernie) voter could decide makes her a poor fit for their vote even if they think Trump is too.

To be clear, I meant the statistical outcome specifically. That's why I stated "The charge seems to be that the proportion of black voters chosen for "deterrence" is about twice as large as proportion of black voters in the general population". Admittedly it would have been better to have written "One of the charges seems to be ..." and I have edited my original post to do so.

If the super-predators ad was a central example of the voter deterrence campaign I can’t see why people think it should be stopped. On the continuum between someone voting for one group and someone voting for another group I think a lot of voters will not vote if they don’t significantly prefer one option. This criticism seems to imply if you don’t think your message is good enough to flip someone’s voting intention but good enough they have second thoughts about voting for an alternative then you have a moral obligation to stay silent.

Reply on news.ycombinator.com