Friday, December 23, 2016

No True Scotsman

So this is a more opinionated post, but I feel this topic should be addressed, and that those who have already done so have not done it in the best way it should/could be. The topic is the Black Lives Matter (BLM) movement, and the issue of police brutality in America (or, if you prefer, the non-issue thereof). I will discuss how the "No True Scotsman" argument--that no "true" member of a group, as opposed to a traitor to the group, would do a bad thing that a claimed member of the group recently did--can be applied to these issues.

Certainly the name BLM seems pretty uncontroversial. But of course, the fact that they are casting a critical eye on the police and calling some actions taken by multiple individual officers into question, as well as arguing that their practices may be racially biased, is the real root source of controversy here. We have already seen the cycle occur multiple times:
1. Police officer shoots unarmed black man. This shooting can be fatal, but this is not absolutely necessary for the cycle to proceed.
2. BLM protest the shooting and, more generally, police brutality and racism in America.
3. Someone, sometimes (as in Dallas this July) motivated by these protests and criticism of police, decides to murder a (or multiple) cop(s).
4. Conservative critics of BLM pounce on step #3 as proof that BLM "hates cops" and wants them all (or at least the white ones) to be murdered. They then call on these leaders to, at least, renounce violence (as Sargon of Akkad recently did, which partly motivated me to write this post). If they are particularly radical, like Heather Mac Donald, they may demand that BLM and other "anti-cop" individuals and movements stop "demonizing police" by creating a "stream of falsehoods", and allege that these entities' statements have led to "an appalling increase in shootings and murders in many cities across America."
5. BLM and other advocates for police reform respond that theirs is a nonviolent movement.

So I'm gonna talk about the fact that BLM is a movement, not an organization, and how for this reason, anyone can identify with this movement, even if their goals or plans (e.g. murdering cops) are out of line with those that BLM was founded to pursue.

This is where the "No True Scotsman" argument comes in: if someone kills a cop, even if that person says that police brutality and BLM were their motivations to do so (as was the case in Dallas), that doesn't necessarily mean that that person is a "true" member of BLM. 

Now, you could also flip this argument around and apply it to cops, which would look something like this: if a police officer does something they definitely shouldn't have done, then they must not be a "true" police officer, but instead someone who did something contrary to what policing is really about.

So is one of these arguments more valid than the other? Yes, in my opinion: the one about BLM. I am saying this because BLM is a loosely organized movement, meaning that there is no vetting process or authority who decides who "gets in" to this movement. In stark contrast, not just anyone can be a police officer; you must first graduate from high school, pass an entrance exam, and then, among other things, undergo about 600 hours of training (though the exact amount varies from state to state). So this is why it matters more when a police officer does something he/she shouldn't do than it does when a BLM member does something he/she shouldn't do.

I will close by adding a link to an excellent blog post I recently found about this subject.

Thursday, December 22, 2016

Where does "eight spiders" really come from?

For the past few weeks I have been, on and off, searching for the origin of the myth that the average person swallows 8 spiders per year in their sleep. (Note: Lemmino already did the exact same thing as me in a very well made and carefully researched video he uploaded this October, which you can watch here.) First, I should note that there is little doubt that this is a myth, and that it is not true. As Scientific American (and Lemmino in his aforementioned video) have noted, this claim "flies in the face of both spider and human biology, which makes it highly unlikely that a spider would ever end up in your mouth." So I'm not gonna focus on that. Instead I will try to find out where this myth originated. Lemmino and fellow YouTuber CGPGrey have already done that, and I think their videos, Lemmino's especially, are pretty clear proof that the source long cited by Snopes (Lisa Holst's purported "article" in "PC Professional" magazine) doesn't exist and was made up by Snopes' webmasters. So I will try to tackle a different question than either Lemmino or CGPGrey did: where did this myth come from?

So I searched a bunch of online databases, including Highbeam and the Google Books archive, to find the oldest books/articles I could regarding this claim. My searches have identified the probable origin of this myth as a 1990 article in the magazine Cornell Engineer, published by the Cornell University School of Engineering. (That said, I recognize that this may have originated earlier, but I don't think there's much of a chance.)

Here are some things that have since become even clearer than the origin of this myth as a result of these searches:
1. This claim has many variations: it seems to have originated, for example, as a claim about the # of spiders swallowed in one's lifetime. This is true not only of the original 1990 article, but also of multiple other articles repeating the myth from 1999, 1999 again, and 2002, among others. That said, there are others claiming it is "eight per year", but there is also at least one source claiming that the average person swallows 8 spiders every day. Granted, it's a fiction book, but still, I would expect published authors to be a little more careful. BTW, one of the 1999 articles cited above cites this webpage listing "useless facts" as the source for the claim, so we know this page was publicizing the claim since at least 1999.

There are also different versions of the myth with different #s of spiders other than 8. For example, alternative numbers of spiders cited include 7, 4, and 10.

2. Many articles and books cite the claim as though it were fact. (E.g. 1, 2) Others cite it more skeptically, like a 1997 Chicago Sun-Times article where a journalist actually called an entomologist at the Field Museum to ask him whether this was true or not. Also, in 2001, the LA Times reported that "The San Fernando Valley Folklore Society, which investigates urban folk tales, says it is not true that "the average person swallows eight spiders a year." Someone made up the stat as a joke, and it quickly became gospel on the Internet." Still others cite it but make it clear that it's an internet legend and/or they just read it in an email, so it might not be true: 1, 2, 3.

Tuesday, December 6, 2016

Did Johnson & Stein cost Clinton the election?

In light of the recent and almost completely unexpected presidential election results, I've decided to see whether third party candidates were responsible for this outcome by stealing votes from Clinton.
First, I will list the margin of victory for Trump (in %) in key swing states, according to CNN (i.e. %Trump-%Clinton):
Florida= 1.3
Pennsylvania= 1.2
Ohio= 8.6
Michigan =0.3 (weird stuff is going on here, with an appeals court recently rejecting an already-underway recount initiated at Jill Stein's behest.)
Wisconsin=1
Georgia=4.7
North Carolina=3.8

Now let's compare that with the % of people who voted for Johnson (J) and Stein (S) in each of these states (in GA and NC, Stein was not on the ballot):
FL: J=2.2, S=0.7 J+S=2.9
PA: J=2.4, S=0.8 J+S=3.2
OH: J=3.2, S=0.8 J+S=4
MI: J=3.6, S=1.1 J+S=4.7
WI: J=3.6, S=1.1 J+S=4.7
GA: J=3.1
NC: J=2.8

Suppose we then divide the margin of victory in each state by the % of voters who voted for J+S combined. (Bear in mind that if the number we get is >1, there's no way third party voters could have cost Clinton the election in that state.) Then this is what we get (results rounded to 2 decimal places):
FL: .45
PA: .38
OH: 2.2
MI: 0.06
WI: 0.21
GA: 1.52
NC: 1.36

Now we can rule out OH, GA, and NC because there's clearly no way Johnson and Stein could have cost Clinton those states. So what about the four remaining states--FL, PA, MI, and WI? Well, if Clinton had won all of them, she would have gotten 75 more electoral votes than she actually did, for a total of 307, which is not only enough to win, but also even more than Trump actually got (albeit by only one).

Exit polling data give us a good idea of how third-party voters would have voted if the race was only between Clinton and Trump. According to CBS, 25% of Johnson voters said they would have voted for Clinton otherwise (C), 15% said Trump (T), and 55% said they would otherwise not have voted (O). Similarly, about 25% of Stein voters said C, 14% said T, and 61% O. So let's multiply the % of each third-party candidates' voters in each of these 4 states.

FL: J: 2.2*.25=.55% more votes for Clinton, 2.2*.15=.33% more for Trump
S: .7*.25=.18% more for Clinton, .7*.14=.1% more for Trump
FL total: .73% more for C, .43% more for Trump
So C would have been .3% closer to winning FL without J and S. But this is not enough, because she would still have been 1% behind Trump there.
PA: J: 2.4*.25=.6% more for C, 2.5*.15=.38% more for Trump
S: 0.8*.25=.2% more for C, 0.8*.14=.11% more for Trump
PA total: .8% more for C, .49% more for Trump
Again, in PA, C would have been about .3% closer to winning without J and S. But this is still not close enough, as she lost the state by 1.2%.
MI: J: 3.6*.25=.9% more for C, 3.6*.15=.54% more for T
S: 1.1*.25=.28% more for C, 1.1*.14=.15% more for T
MI total: 1.18% more for C, .69% more for T
Without J and S, C would have been .49% closer to winning MI than she actually was. Because she lost the state by only .03%, she would have won MI without J and S.
WI (all the same as in MI, weirdly enough): J: 3.6*.25=.9% more for C, 3.6*.15=.54% more for T
S: 1.1*.25=.28% more for C, 1.1*.14-/15% more for T
WI total: 1.18% more for C, .69% more for T
Without J and S, C would have been .49% closer to winning WI than she actually was. But this would not have been enough, because she lost the state by 1%.

So it looks like Clinton would only have won the ridiculously close Michigan race without Johnson and Stein. So she'd have gotten 248 electoral votes, compared to Trump's 290. But Trump would still have won because you only need 270 electoral votes to win.

So the answer to the question "Did Johnson & Stein cost Clinton the election?" seems to be "No."

I should note that Vox looked at this same subject already and concluded that the answer to the title of this post is no. The Washington Post, in contrast, concluded that without Johnson & Stein, "Clinton might have won, based upon these data, but only by winning both Pennsylvania and Wisconsin. If Trump held onto even one, he would have kept an electoral college majority." The WaPo also noted that in this hypothetical 2-candidate scenario, "In Pennsylvania and Wisconsin, Clinton and Trump would have tied at 48 percent apiece."