Remember that Big War Study I’ve mentioned a couple of times? No? Well never mind; I’m going to tell you about (one of the studies in) it now!
In the first few studies, we’d given participants a war with a “just” side and an “unjust” side*, and asked them to make judgments about soldiers fighting in that war. We found that, as you may or may not expect**, people said that when a soldier on the just side kills a soldier on the unjust side, this is more morally defensible than the other way around. We refer to this as an “asymmetric” pattern of judgments.
So in this next study, we wanted to know how people would feel about soldiers on either side of the war when they do something really horrendous, like defile enemy corpses. Or what about when they do something really great, like go out of their way to protect enemy civilians? Would these acts still elicit asymmetric judgments, or would they be equally wrong (or right) for soldiers on both sides?
To test this, we still had a war with a just and an unjust side, but this time we had 24 different scenarios, in which the soldier acted in a wide variety of ways. Participants judged a random 8 of these scenarios, and they again indicated how morally defensible they thought the act was.
This is what the data look like:
Ahem. That is what the data looked like before I had any idea what I was doing when trying to use R for data visualization. (Mistake #1: too much jitter.)
A few*** hours later, I’d wrangled it into this shape instead:
It’s still not perfect (and in the actual paper we’re not even trying to present all the results in a single figure), but it’s getting closer to something I talk you through!
The 24 scenarios are now arranged in a 4×6 grid, from overall most wrong (top left), to overall most right (bottom right). Each scenario has two violins, representing judgments of soldiers on the just side (left) and the unjust side (right) of the war.
You can see that in the top and bottom row of the scenarios (i.e. the worst and best scenarios), there is very little difference in judgments of soldiers on either side.**** The black dot represents the mean, with bootstrapped CIs (but I don’t know how R did that part – will have to get back to you!) For the middle scenarios, you can see that in general, the moral defensibility judgments are lower for soldiers on the unjust side.
And those terrible colours? Well, in each scenario, the soldier was either ordered to act (and obeyed the order), or he made a decision to act (and acted on his own decision). This had overall surprisingly little effect on people’s judgments, which is why I’ve overlaid the yellow and purple violins.
And that’s it! A picture is meant to be a 1000 words, but I’ll leave the remaining 500 for another post. (Perhaps to explain what each of the scenarios involved? Or how to analyse this in a mixed model?***** Such exciting possibilities….)
* Yes, this is contentious. Yes, we go into detail in the actual paper.
** And yes, we discuss reasons (philosophical, political, psychological) why you may expect one thing or the other in the paper too.
*** About 6 hours, split over 3 days (1 day a week).
**** So is this a floor/ceiling effect? Are there “real” floors and ceilings in how we feel about morality? Let’s save this for a different post.
***** Thanks to Nate Carnes for helping me figure this out!