Big Data (Mining)

eus-2019-05-30-10As of 10:00 this morning. The fun continues.

Donald Trump

Making Amerikkka Great Again.

Colby Itkowitz and Dan Lamothe, “White House wanted USS John S. McCain obscured during Trump’s Japan visit,” Washington Post, May 29, 2019,

Big Data (Mining)

I’d already realized that “big data” would methodologically rely on correlation as if it were causation.

My first methods professor, back in, I think, 2003 or maybe 2004, Valerie Sue, warned against “data mining,” that is, simply sifting through data, identifying correlations, picking one, and publishing an article on it. There’s pressure to do something like this because eligibility even for tenure-track faculty positions generally hinges on publication credits. Even tenured professors are expected to continue publishing, even as they continue to take on full class-loads.

But there’s no showing of a causal relationship. The variables can be completely random. Which, as I realized what “big data” means, and how it must operate (statistically, of course, and on a massive scale), would be precisely how it would work. And the results will be wrong—or, worse, occasionally right, but only by happenstance. It’s idiocy with statistics. It’s what they warn you against in statistics classes and on a massive scale.

And guess what so-called “artificial intelligence” is going to rely upon? Very heavily, big data mining. Which is why I call it “artificial idiocy.” Welcome to the new world.

Allison Schrager’s attack is a little different. Her argument boils down to that old financial services disclosure: “Past performance is no indicator of future results.” What’s big data relying on? Past performance, that is, data that’s already been accumulated. Again, there’s no causal relationship established: The assumption is the very inverse of that financial services disclosure.[1]

Taken to heart, this could shake up positivism. Yes, there’s always been a research cycle in which theorizing is an integral part. You are supposed to explain that causal relationship. But as I saw when I looked at chronotype research, and have discovered of medical research (yes, this is the stuff people are trusting their lives with) in general, it’s often elided, which means it’s garbage “research.”

Why? Because “why” questions are hard. Statistics can never answer them. I remember being discouraged from asking the “why” questions later on in my undergraduate career. But I never could stop asking those questions, which is part of how I wound up as a human scientist, not a positivist.

Allison Schrager, “Here’s a prediction: In the future, predictions will only get worse,” Quartz, May 29, 2019,

James Comey

Robert Mueller may be done with his investigation. He might not want to testify before Congress.[2] It doesn’t seem like Congress is done with him, as even a Republican wants him to testify:

“Mr. Mueller’s statement…is a case study in pettifoggery, and reinforces my position that he should be compelled to testify before Congress,” McClintock said in a statement provided to McClatchy. “Any president can be indicted after he leaves office and it was Mueller’s job to make a recommendation for prosecution, if one existed. He did not.”

“Instead, he prefers to make innuendoes while hiding behind DOJ [Department of Justice] guidelines and not be questioned,” McClintock continued. “Sorry, it doesn’t work that way.”[3]

Kind of interesting that no one in the Twitter bubble seems willing to let Congress make that decision “on their own” and cognizant of political reality.

The neoliberals (mainstream Democrats) would lose in 2020 because they’re neoliberal and because they can’t stop themselves from nominating a pervert (Joe Biden). The Left would lose because they pushed for impeachment, banging their heads against the wall because it made them feel better, while invoking a backlash that sees Donald Trump re-elected.[4]

So guess how I think 2020 is going to turn out?

Emily Cadei, “Tom McClintock accuses Mueller of ‘pettifoggery,’ says he should testify before Congress,” Sacramento Bee, May 30, 2019,


Tara Bahrampour and Robert Barnes, “Despite Trump administration denials, new evidence suggests census citizenship question was crafted to benefit white Republicans,” Washington Post, May 30, 2019,


Abigail Hauslohner and Maria Sacchetti, “Hundreds of minors held at U.S. border facilities are there beyond legal time limits,” Washington Post, May 30, 2019,

  1. [1]Allison Schrager, “Here’s a prediction: In the future, predictions will only get worse,” Quartz, May 29, 2019,
  2. [2]Robert S. Mueller, III, “Special Counsel Robert S. Mueller III Makes Statement on Investigation into Russian Interference in the 2016 Presidential Election,” Department of Justice, May 29, 2019,
  3. [3]Emily Cadei, “Tom McClintock accuses Mueller of ‘pettifoggery,’ says he should testify before Congress,” Sacramento Bee, May 30, 2019,
  4. [4]Will Bunch, “Trump’s diabolical plan to blow up democracy, get reelected and avoid jail just might work,” Philadelphia Inquirer, May 5, 2019,

One thought on “Big Data (Mining)

This site uses Akismet to reduce spam. Learn how your comment data is processed.