Lies and statistics

Whenever someone quotes statistics like “80 percent of gun owners support expanded background checks” or “one third of all women will be sexually harassed,” I ask myself, “Where did that figure come from?” At long last we see a letter (from a college professor) that actually provides the source of his statistics: a poll conducted by Data for Progress and presumably commissioned by the Justice Collaborative Institute.

The letter makes the surprising argument that over 50 percent of Republicans support one of the House of Representatives’ more extreme Democrats’ more extreme proposals. Based on that argument, he then excoriates Republicans as “hypocrites and posers.”

Intrigued, I looked up “Data for Progress” and learned that it “uses data science to support progressive activists and causes.”

Not surprisingly, all of their “scientific” polls demonstrate strong popular support for left-wing and Democratic Party issues. This is inherent in the bias of their questions, along the lines of “Do you believe progressive Democrats are better people than hypocritical Republicans?”

The Justice Collaborative Institute? Truly an “impartial” source. Some of their agenda:

• No cooperation between local police and the U.S. Immigration Service. Or, less politely: support sanctuary cities.

• Don’t force people to leave the United States based on criminal arrests or convictions. Or, less politely: don’t deport illegal immigrants who commit murder.

• Don’t sentence anyone to more than two years unless the crime involved “risk of serious physical harm.” Or less politely: a three-time burglar or con artist shouldn’t get more than two years in jail.

And the list goes on.

So the problem with the professor’s argument is that cherry picking data will always give you the answer you want. In computer terms: “garbage in, garbage out.” I wonder if he passed that message along to his students.



Submitted via Virtual Newsroom


Today's breaking news and more in your inbox

I'm interested in (please check all that apply)