I've been on debate boards with scientists who defended evolution with a careful, honest appraisal of the evidence, but as soon as the discussion turned to politics, they were happy to quote misleading—sometimes ridiculous—statistics to defend their political views.
I was told that 50% of homeless people are families who have fallen into hard times and are diligently seeking jobs, not drug addicts, alcoholics, and people who simply prefer the freedom of street life.
Anyone who has ever spent time with homeless people knows this isn't even close to true, but there are statistics that can be skewed to back that up.
I've spent a lot of time with homeless people. As a Christian, ministering to people is supposed to be second nature to us. I know very well that almost no people on the streets are family people out of luck with work. So I decided to research the statistics I was given.
As it turned out, 50% of those admitted to homeless shelters are indeed family people in hard times. However, most of those find a job and a place to stay within 3 days, while almost 100% of the people who turn up on park benches and begging on the streets are drug addicts, alcoholics, or mentally ill.
Why do people—especially if they're supposed to be unbiased scientists—skew such statistics? Easy. It's popular and easy to accuse the American economy of letting people down than it is to take on the real homeless problem.