Do you lot move Facebook? You likely create because you lot aren't me. That agency the Trump elbow grease got information virtually you. Hah! Sorrynotsorry.
Anyway, today's topic is the move of Facebook's information for micro-targeted ads. What tin these accomplish? Buried inside this query are 2 variety out questions. First, what create nosotros acquire virtually you lot beyond your basic demographics, too second, tin that cognition last exploited?
Let's start amongst the get-go question. If I know that you lot are African-American, how much create I know virtually your partisan preferences? H5N1
lot. Why, because approximately 90% of African-Americans are Democrats. Tell me that you lot are African-American, too inwards particular, an African-American
woman, too I tin brand a
really actually really good prediction virtually your vote choice. In the 2016 American National Election Studies survey, 95% of African-American women went Clinton over Trump inwards the two-party vote. Now, you lot are all individuals.
Fine, fine. We are all individuals. However, this is virtually baseline statistical probabilities, too the baseline statistical probability of an African-American adult woman voting for a Democrat is so high that the additional information on Facebook doesn't compass you lot that much more.
I'm only talking virtually voting behavior. As a person, what does Nora K. Jemisin pick out inwards mutual amongst Bettye Lavette? No clue. I've never met either. I actually genuinely actually desire to run into both. I would facial expression them to last
very different people. The trend Jemisin writes, too the trend Lavette sings, I incertitude they pick out much inwards common. But, I bet neither voted for Donny-boy.
White people? White males? Statistically, nosotros are to a greater extent than probable to pose every bit Republicans, too vote Republican, but statistically, nosotros are to a greater extent than divided than African-Americans. According to the 2016 NES, white males favored Trump past times virtually 60-40 inwards the two-party vote.
Um... what was I maxim virtually statistical inferences too groups? Uh...
I'm going to quote the swell philosopher, Frank Zappa here: "I'm non black, but there's a whole lotsa times I want I could nation I'm non white."
Regardless, in that place is an of import statistical signal here. There is less politically relevant information conveyed past times the fact that I am a white human than past times the fact that Nora Jemisin is an African-American woman. How much other basic information could last added to uncomplicated demographics to acquire my preferences correct, though?
I'm weird, for a lot of reasons, but this is the basic number for Cambridge Analytica. How much tin you lot acquire from personality-based information, inwards a higher house demographic baselines? Basic political scientific discipline volition tend to acquire you lot a lot farther than "big data" bullshit.
Second, what virtually micro-targeted ads? Here's the basic fact that I shell into my students' heads, metaphorically, every fourth dimension I instruct virtually voting behavior. 90% or so of the electorate are partisans. Forget that bullshit virtually how 30-40% are independents. Those people are liars. The basic political scientific discipline method of bespeak virtually partisanship inwards a survey is to inquire
two questions rather than one. If someone get-go claims to last an independent, you lot must too so inquire if that somebody
leans towards i political party or the other. Most "independents" acknowledge to a leaning, too most "leaners" human activity only similar partisans, thus the trouble that so many of my students tin recite: "leaners are liars." Can I trademark that?
Somewhere approximately 85-90% of votes inwards whatsoever given election are only partisans voting their political party ID. The remainder pick out messages coming at them from all directions too incoherent preferences virtually politics. The stance that they are susceptible to narrowly tailored messages relies on the stance that they pick out consistent too coherent preferences, too call back inwards consistent ways virtually politics, but if they did that... they'd last partisans, too pick out their votes determined past times partisanship, inwards which case, none of this Cambridge Analytica materials would thing anyway.
Does the micro-targeting of ads influence outcomes? I strongly incertitude it. James Comey? Yeah, that motherfucker gave the White House to Trump. Cambridge Analytica? Nope.
So, here's the storyline. Trump's elbow grease used Cambridge Analytica information from Facebook, too Trump won. Therefore, he won
because of micro-targeted ads. Post hoc, ergo propter hoc. Nope. That's a logical fallacy.
Trump bragged virtually committing sexual onrush too getting away amongst it. Trump won. Therefore, Trump won
because he bragged virtually getting away amongst sexual assault. Post hoc, ergo propter hoc.
Put shit on Facebook, too someone's going to acquire it. Did it swing the 2016 election? Not likely. Privacy, though? Yeah, that's an issue.
