Public data set highlights: hate crimes and fake news

Thursday, November 17, 2016

This week's Data is Plural highlighted two things that I have been following particularly closely since the election. The first is hate crimes:
Hate crimes in the United States. Since the 1990s, the FBI has collected data on hate crimes from local law enforcement agencies. On Monday, the bureau released data for 2015, reporting “5,850 criminal incidents and 6,885 related offenses, as being motivated by bias toward race, ethnicity, ancestry, religion, sexual orientation, disability, gender, and gender identity.” Those numbers are based on reports from 14,997 participating agencies. On the FBI’s website, you can view and download summary tables of the most recent data. You can also download incident-specific data for 1992 through 2014 from the National Archive of Criminal Justice Data. Unfortunately, as ProPublica noted yesterday, the FBI dataset is “deeply flawed”; more than 3,000 law enforcement agencies don’t participate in the program.
It is too early to say definitively that hate crimes are in fact on the rise, but there is reason to be concerned. The analysis of the aforementioned data showed that hate crimes against Muslims rose by about 6% last year, and two other groups that collect data on hate crimes, the Southern Poverty Law Center and the Council on American Islamic Relations, have noted a spike in reported hate crimes during election week and in the days since. Reports of a 46% increase hate crimes in post-Brexit UK certainly don't make anybody feel better, either.

It is worth pointing out that police department non-participation also causes problems for tracking information on police shootings and excessive use of force. Vox does a pretty good job of laying out these data problems. (Side note: I've been quite enamored with some of their data reporting lately.) DiP has featured open data on police activity in several editions:
The second, which has been the focus of very intense discussion for some time now, is on media coverage:
Fake news on Facebook. Last month, colleagues at BuzzFeed News and I analyzed and fact-checked 1,000+ posts from hyperpartisan Facebook pages, and found a disturbingly high rate of fake news. Here’s the data. Facebook CEO Mark Zuckerberg has dismissed the possibility that fake news influenced the election, calling it a “pretty crazy idea”. Meanwhile, renegade Facebook employees have now formed an unofficial task force to battle fake news on the platform.
There was more than a little hubbub about Facebook's role (or lack thereof) in driving the election by reinforcing people's self-constructed information bubbles. I personally think the more important ethical question is whether Facebook has a responsibility to do something about it by correcting its users' bad habits driven by the very natural (despite also being very unfortunate) human tendency toward confirmation bias. But I think it is also worth noting that the tendency toward that bias may have been exacerbated by the media's own self-constructed bubble.

I don't have a whole lot of optimism that the polarization of this country, or the tendency of its citizens to only accept information that confirms what they already believe. But on the bright side, at least analysts and the media are paying closer attention now.

No comments :

Post a Comment