Drugs, alcohol, suicide, and Trump

Perhaps they just wanted to even things up by making the rest of the country suicidal.

True, but I feel she's* putting forth a false implication with the bulk of her study's focus. She clearly states early on that:

1) Trump performed best in counties with high economic distress and a large working class.

2) Drug, alcohol and suicide mortality rates are higher in counties with more economic distress and a larger working class.


So, the mediating variable here is socioeconomic status, which she seems to largely ignore for the remainder of her report. I'm not saying she ignored the role of economic status (indeed, she devoted 2+ pages to discussing it), but she does seem to dismiss it when discussing the correlation between drugs/suicide and Trump voters.

She states that:

Bolding mine.

So, basically, she found that Trump's election correlated with a bunch of things but she decided to emphasize the drug, alcohol, and suicide connection...but then coyly say that "no single factor" can account for his election on its own.

This is an academic form of clickbait. The largest influencing factor is actually socioeconomic status. 

The linear relationship is, basically:   SES ---> bunch of other stuff related to SES ----> Trump.
                                                        \___________________________________>

That little diagram I tried to make probably won't show up well, but basically, SES is directly related to Trump votes. SES is also directly related to the other stuff. The other stuff also correlates with Trump votes, but only because they are related to SES. The relationship between all that other stuff and Trump would likely not exist without the SES connection. 

The variables in the middle are more strongly related to SES than they are to Trump. Indeed, the SES relationship is the important one. However, she keeps intentionally highlighting drugs/suicide throughout the paper, even though there's no real theoretically logical or statistical reason to do so.

To me, the implication that she's attempting to make is clear; she's trying to imply that Trump voters are a bunch of suicidal drug users. It's clever in that it relies on invoking the biases of the reader. Trump is a deplorable person. The "Trump voters are all racists" card has already been played, and debunked. So, what other way can we say that "only miscreants voted for Trump" without playing the "racists!" card. Oh, I know! Everyone just *knows* how *drug users*...you know...*those people* are, amirite? So, if it can be shown that geographic areas with a bunch of drug users and suicidal people were more likely to vote for Trump, then we can then imply that Trump voters are a bunch of grubby drug users with a death wish!

But, she's not directly saying that. Oh, heavens no. She's just finding the somewhat spurious correlation and then claiming, "I'm not saying that there's a causal link here; I'm just saying that the connection is "interesting"."

That's very similar to the Fox News, "I'm just asking questions" approach.

Like I said, this is the academic version of clickbait. She's invoking the drugs/suicide ---> Trump connection because it will get her paper attention, not because it is theoretically or statistically important or substantive.


*I have no idea if the author is male or female. However, every time I see the name "Shannon," I automatically picture a female because I've known far more female Shannons in my life than male. So, for convenience's sake, I'm going to use "she" rather than "he/she."