Why Hillary Will Win Big: Unskewing Nate Silver with Social Science Methods 101


Because nothing is more important than made up numbers.

Because nothing is more important than made up numbers.

Everyone is on edge at the prospect of Donald Trump assuming the Presidency, but it ain’t gonna happen. Trump is the candidate of a declining minority of US voters. Over half of  US voters are women, and the majority will vote Clinton. Depending on turnout, roughly 20% are African American or Latino, and way over 80% or those will vote Clinton. About of quarter of Americans hold a college degree, and the majority of them will vote for Clinton. This is a much better indicator of who will win than any of the bullshit non-scientific polls that are the fascination of the media, the public, and even supposedly educated social scientists.

Why do the “polls” show what they do? Well, we need to unskew them using our basic understanding of undergraduate social research methods. We know that estimates of population parameters require a random sample taken from the target population—or, I guess we used to know this but now this has become contested knowledge since shitbag scumfucks have started bullshit polling firms and taken over the AAPOR. Gauss is rolling in his grave. But, let’s just start with the reality of these polls from a production standpoint. Hands on research methods. Let’s get our fingers dirty and lick them off.

The standard “poll” conducted by our many for-profit consulting, media, and “research” firms currently have no formal sampling frame. They MAY randomly select respondents, but they will have 1500 or 750 or however many respondents by 9pm on Monday. The average response rate of these polls is likely under .1% that is point one percent. The “top” polls from places like Pew are probably under 1% for originally targeted respondents. They simply move on to the next household.

Second, the interviews are now all conducted by minimum wage workers or are robocalls—or worse, they are internet surveys. The quality of data obtained by these “polls” is worse than exceptionally low.

Third, many of the “polls” don’t even use random samples, but idiots who sign up to be respondents for all kinds of surveys and get paid for it. Do you really think the average American would do such a thing? No. Do you think these cheetos eating morons will be disproportionately for Trump? Duh? The people who sign up for such bullshit are often the same people who think that Hillary parachuted into Benghazi to burn down the embassy.

Fourth, What kind of fucking moron answers their phone when someone from an unknown number calls? Are you fucking kidding? The problem with the response rate issue is that non-response is, on its face, systematic. Young people and educated people don’t answer the phone for a stranger. Busy working class people, African Americans and Latinos….they may even be being charged if they answer some bullshit call. If the caller ID doesn’t say “mom” or “spouse” or “friend” or “kid” nobody picks up on that shit. Who does? People like my mom, god love her, elderly people who think when the phone rings you must answer it. And, she has a landline, which most of these polls sample from exclusively. Paranoid psychotics who wonder who got their phone number, they answer the phone. Normal people don’t answer the phone. If Hill didn’t have a huge elderly female following, she’d be down by 10 points in these bullshit polls.

So, go vote. And, don’t worry. Our worries will come later when the Ted Nugent types start shooting at us. Make sure to remove your “I’m with Her” bumper sticker, remember the Fraternal Order of Police endorsed Trump, which means your friendly neighborhood cop is probably a fascist.


3 Responses to “Why Hillary Will Win Big: Unskewing Nate Silver with Social Science Methods 101”

  1. Casey Says:

    Would your solution to the problem be a poll that uses essentially the same sampling procedures as the GSS?

    • sherkat Says:

      Because of the shit poll industry, it is impossible to do a scientific pre-election poll. The GSS takes months (and sampling design would have to be different if you want to predict electoral outcome instead of represent the population). I’d like to see the entire industry go belly up. Horse race crap is not a way to run a democracy.

  2. Casey Says:

    Now that it’s November 9th, what are your thoughts?

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: