Popular Political Myth: Blacks In Prison

[phoenixx]

Political types love to trot out the stats on the number of blacks in prison being 3 to 5 times the amount in college, depending on who's doing the talking. Sometimes they don't trot out a number, they just use the mental image that the statistic or statement entails to make their point.

I confess, it had never occurred to me to question that statistic. I don't know why - I question everything else I hear. As this post outlines, accepting those claims at face value was my mistake (and a mistake many other Americans are making as well).

What do you think of the issue?

[drmomentum[$]]

I think people should take better care to get their numbers right, and I'd like to hear what the candidates have to say about their sources.

[mjfrombuffalo[$]]

There's lies, damned lies, and statistics. Depending on how one words the question and what population you look at, you can get different results. [ / mj's usual statistics mantra]

In 2002, the Justice Policy Institute issued a report showing 791,600 Black men incarcerated with 603,000 in college, but a rebuttal said the college number was underestimated and that "people of all ages are sent to jail, whereas college students tend to be 18-24 year olds" (Are More Black Men In Jail Than College?).

In 2004, the American Sociological Review narrowed both the population studied (Black men born 1965-1969), but also narrowed the criteria (ever been in prison vs. graduated college or served in the military). Their findings were 20% had served time in prison by the time they reached their early 30's. They also looked at the incarceration rate for Black men born 1945-1949 and saw an increase in incarceration rate between the two demographic groups, with the boys of the 60's more likely to have been incarcerated at some time (not including local jail). American Sociological Association

NYTimes summary of various studies by Columbia, Princeton, and Harvard on the issue.

Reason.com/Fact Checker's data is point-in-time data (number in jail at end of 2005 vs. number enrolled in higher ed in 2005), similar to the Justice Policy Institute data. Point-in-time data is interesting demographically, but does not necessarily serve well as a predictor or indicate trends.

Moral: You can almost always find statistics to back a statement. You can also almost always find statistics to refute a statement. Wheat and chaff can look very similar when sorting and it's interesting (at least to me) to try to find the difference, but sound bites often don't contain a lot of wheat.