Updated at 2.30pm EST, 15 August 2014 with additional analyses. See bottom of post.
Over the last several days there have been protests in Ferguson, Missouri in response to the shooting of Michael Brown, an unarmed teenager. The Ferguson police have reacted with force to these protests prompting a lot of concern with how the police handled protesters, the media, and the militarization of police departments. This has been – and still is – a tragic and disturbing story.
I added a few questions about this situation at the end of an experiment I just finished collecting on Mturk (N = 341, though varies depending on the analysis). The results of the experiment will wait until another time. Below are the simple and preliminary results of the Ferguson items.
First off, normal caveats apply. The data below are not experimental. The data below are not representative of any population. The data were collected on
April August 14th*, starting at approximately 9am EST and ending about 3 hours later.** Most people said they heard about the events, though plenty of people did not. Of the people who answered yes to the above question, below is a histogram of responses to the question “How much have you been following the events in Ferguson, Missouri?” (0 = not at all to 10 = very much) Of all the participants I also asked “To what extent do you sympathize which each of the following groups in Ferguson, Missouri?” (0 = not at all, 10 = very much). The groups were “the police,” “the protestors,” and “people living in the neighborhood who are not protesting.” Three histograms: Although plenty of people had no opinion, in general people felt less sympathy for the police compared to the protestors and they felt the most sympathy for people living in the neighborhood who were not protesting.
The study included several demographic variables and so I examined how self-reported race, age, political party identification (1 = strong Democrat, 7 = strong Republican), and income (1 = less than $10,000 per year, 8 = over $100,000) were related to the three sympathy items. The differences between Blacks and Whites were always significant (p’s < .01; d = .94 for police and d = .61 for protestors). These results should especially be taken with a grain or seven of salt because there were only 19 people who identified as Black in the entire sample, whereas there were more than 190 Whites.
Ok, now for age… Now for party identification… Now for income… In general, people who are older, more Republican, and earn more money report more sympathy for the police and less sympathy for the protestors compared to people who are younger, less Republican, and earn less money.
I know that this is far from a complete analysis and that there are a lot more questions to be asked.
This is only a start.
*What part of me wanted to think that it was April??
**That is, this morning
[Edited to fix a typo early in the morning on 15 August]
Update: 2.30pm EST, 15 August 2014
Brendan Nyhan, on Twitter, suggested that I look at the difference between sympathy for police and protestors. These analyses are below.
Brendan Nyhan, a political scientist at Dartmouth, argued in a recent column on The Upshot that President Obama should maybe stay out of the events in Ferguson because this will likely polarize the issue. This has the potential to set back the emerging bipartisan discussion about race and the criminal justice system.
In response to a tweet about my initial post Brendan suggested…
@MBrandt05 Interesting. Any way you could calculate *difference* in protestor/police sympathy? Clearest measure of polarization potentially.
— Brendan Nyhan (@BrendanNyhan) August 15, 2014
Good call. I subtracted sympathy for protestors from sympathy for police to create a difference score. Higher numbers indicate more sympathy for police compared to protestors. Below is the scatterplot which shows that Democrats have more sympathy for protestors compared to the police (scores below zero) and Republicans tend to have more sympathy for the police compared to the protestors (scores above zero). The effect is not big, but it isn’t tiny either.
Another way to cut the data is to split people into Democrats (less than 4 on party identification), Republicans (greater than 4 on party identification) or moderate (equal to 4) and to split people into police sympathizers (greater than 0) and protestor sympathizers (less than 0). Below are those results.
All-in-all, in this specific sample, Republicans tend to be more sympathetic to the police compared to the protestors and Democrats show the opposite pattern.
I was relieved to wake up to much more peaceful news about the protests the morning. I hope that this can be the start of a productive dialog in Ferguson and the US in general so that we can all live in safe and just communities.