,
28 tweets,
11 min read
Read on Twitter
From gender bias to media bias? A thread on how our study looking at adult perceptions of children's pain got misconstrued. A reminder of the importance of taking media coverage with a grain of salt, reading original studies when possible, and guarding against confirmation bias. 
"Parents take young boys' pain more seriously" (@FatherlyHQ); "People don't care as much when girls cry as when boys do" (@Cosmopolitan); "Pain often missed in girls" (@usnews); People see females as "less credible" when reporting pain (@NYTHealth). Where to start?
First, the effect in our study was observed *only* in female participants--not "Americans" or "parents" or "people" in general. In fact, male participants rated girl pain *higher* than boy pain (albeit not to a statistically significant degree).
Second, we did not measure "sexism" or "credibility" or the extent to which adult raters "cared" about the pain of boys vs. girls--or even whether they took the pain of boys "more seriously" than that of girls. And we certainly did not show that pain is "often missed" in girls.
Here is what we did. We took a single video of a 5-year-old child getting a routine blood-draw who was dressed & filmed in such a way that their actual sex/gender was not apparent (you couldn't tell if it was a boy or girl). We showed the exact same video to participants...
across two conditions. In 1 condition they were told the child was "Samantha" & asked to rate "her" pain; in the other condition they were told the child was "Samuel" & asked to rate "his" pain. Overall, participants rated "Samantha" @ 45.9 on a 100 point scale, "Samuel" @ 50.4.
As we state in the paper, this difference is *statistically* significant, but we don't yet know whether it is *clinically* or *practically* significant (e.g., we don't yet know whether it would translate to different pain-treatment decisions or expressions of sympathy, etc.).
But even more importantly, we go on to note that the difference was driven entirely by female participants: when you look at men only, there is NO statistical difference in ratings; when you look at women, it's 45.7 for Samantha vs. 53.1 for Samuel (effect size d = .34).
We did not predict this. So as of right now, we are not sure how to explain the observed female-only bias in ratings of child pain based on perceived gender. What we do know is that when we control for the explicit belief that boys, in general, display less pain than girls ...
... the overall effect that "Samuel" felt more pain than "Samantha" disappears (by "display" we mean, showing or expressing pain, e.g., by crying out or saying "it hurts"). So one possibility is that the general belief/expectation that "boys don't cry" affects the inference
that (some) adults draw about the felt experience of pain in boys when they DO in fact cry out or express pain. The inference might go something like this: "Boys in our culture are taught to act tough & strong when they are in pain & not to cry out, so if a boy DOES cry out...
& express he is in pain, then perhaps he is experiencing quite a lot of pain--enough to behave in a way that boys are stereotypically taught not to behave." If that is the right explanation, it would say nothing about not taking girls' pain (as) seriously, dismissing their pain,
caring less about their pain, etc. It would just mean, given real-life socialization differences between boys & girls, for a boy to display X level of pain, he might need to feel X+ pain, whereas for a girl to display X pain, she might need to feel X (or possibly X-) pain.
In other words, our finding is neutral between 2 explanations: (1) the girl was seen as "overreacting/being dramatic" (so felt less pain than displayed), (2) the girl was seen as expressing pain *accurately* but the boy was seen as underplaying pain (feeling more than displayed).
Either explanation would be consistent with the data (or some combo of both), and we are doing follow-up studies to tease those issues apart. But this still wouldn't explain why it was that *only* female participants rated the boy as in more pain than the girl. When reporters
asked me about this, I emphasized that the finding was unexpected & we did not have direct evidence in favor of one hypothesis over another. But I also suggested that work by @kate_manne could provide a plausible framework to be tested @ some point qz.com/1535889/yale-s… …
Specifically, while noting that it was entirely speculative as a potential explanation for our female-only finding, I floated the idea that if women are socialized to be more alert to men's needs under various conditions, it could be that they are extra-sensitive to
(among other things) signs of pain or discomfort among men, & this could translate down to greater sensitivity toward expressions of pain even in young boys. I was pleased when I saw @kate_manne told a @washingtonpost reporter that she found this notion plausible too (see below).
Other publications, like @Cosmopolitan & @Jezebel, picked up this explanation, suggesting that our finding might be evidence of "another manifestation of sexism" jezebel.com/girls-pain-tak… …. And so it might be. Though, again, we have no direct evidence of this from our experiment.
But this got me thinking. Suppose we'd found the exact opposite of what we found--namely, that only men showed the effect of rating "Samuel's" pain higher than "Samantha's." It seems likely that this, too, would have been interpreted as a plausible manifestation of sexism, or as
"making sense" given the logic of misogyny. And I can imagine that the headlines would not have said "People..." or "Americans..." but rather, "Men take the pain of girls less seriously than the pain of boys." I might be wrong, but this seems plausible to me.
If I am right, though, that opposite findings would both have been interpreted as evidence of the same phenomenon (sexism or misogyny), then we may have a situation where a given explanatory framework, as characterized, is not meaningfully falsifiable: researchgate.net/publication/32…
This doesn't mean a more fleshed-out theory of sexism/misogyny, w additional auxiliary assumptions built in, would also have counted opposite findings as theory-supportive; but it suggests such auxiliary assumptions need to be spelled out & carefully tested in the present case.
In short, our finding provides a peek into a fascinating & complex issue that raises many questions & deserves extensive study & careful testing of different hypothesis. Clinical & practical significance remains to be seen, as does generalizability beyond the video we used.
Here is a free link to the pre-print of the paper with open access data and materials psyarxiv.com/pqg3a/
And here is a little follow-up thread answering some questions that have come up from tweeters:
And here is an open access paper of mine discussing another instance of media misrepresentation of a primary finding (in this case, truly egregious misrepresentation): onlinelibrary.wiley.com/doi/abs/10.100…
