Notices by Cat Hicks (grimalkina@mastodon.social), page 8

It's so wild that in education right now there's a huge swirling conversation about how kids don't read enough or well enough and on the other hand that large blocks of text from "personalized tutors" are gonna be the interface that works

*shrugs* sorry I'm not some edtech genius or an Ed thought leader I'm just a person who has sat elbow to elbow with a lot of kids (some of them yelling at me, or crying at devastating news about their test scores) and actually did tutoring, for free, for standardized tests, for years

Personalization IS an interesting and useful idea (that's why.... We've cared about it.... For eons....) but it's actually SO funny to compare the types of "AI tutors" in classrooms to slowly generating textbooks and ask yourself if they share as many features with textbook learning as they do with anything else

Turning over rocks (reading papers) and just finding "male and female brains just are wired differently in how they problem solve" underneath them (in the citations) 💀

don't judge a book by its cover but tbh DO judge if your core foundation for a gender differences hypothesis is mostly found in consumer marketing journals like................L I K E

oh god "Twenty-eight practicing auditors (16 males and 12 females), averaging three years audit experience, performed a planning analytical procedures task at one of two complexity levels"

Yes from this let's definitely generalize about group diffs in how human beings work in code

I'm tapping the sign so much today

https://mastodon.social/@grimalkina/113210817267521576

I just know a reference list hates to see me coming

Definition of heart's in the right place/mind isn't I'm sorry

one of the studies cited as a basis for gender differences FINDS NO GENDER DIFFERENCE in their outcome of interest

HCI has a LOT to fucking answer for

I'm mad

I have really really avoided doing this kind of thing but this one honestly might need a step by step blogpost about how unsupported claims get "made real" by being mentioned in an influential paper and then just referenced again and again with no one ever going back down the chain of evidence and when that is happening about something where we should have a high bar, it's unacceptable.

oh good. an fMRI study. I'm sure that'll be cited in a way that's super reasonable and absolutely informed and isn't at all in a weird consumer business journal oh wait it's about Ebay offers

I just read this out loud to my neuroscience wife and my research friend and we are all SCREAMING

It's more complex than that obvi ("how do we move past/get more people past/ the fact that the early introduction of diversity can heighten subgroup identities and increase bias and feel worse before it feels better"? an important question of responsibility here) but STILL

Reading papers about interventions to decrease bias in STEM fields is like, so funny because they'll be like: "on the one hand, the evidence is clear that we benefit emotionally, cognitively, behaviorally, and in overall achievement from solving this equity issue. On the other hand, every fiber of our being screams against progress for some reason"

This is so fundamentally key to understanding the next wave/current frontier of truly changing environments. Realizing that a lot of evidence that has looked contradictory is because key causal elements of the social context were not included. This is exactly one of the points I tried to make in my review paper here, about "missing layers" that reconcile seemingly paradoxical outcomes for interventions: https://osf.io/preprints/psyarxiv/qz43x

- Binning, K. R., Doucette, D., Conrique, B. G., & Singh, C. (2024). Unlocking the Benefits of Gender Diversity: How an Ecological-Belonging Intervention Enhances Performance in Science Classrooms. Psychological Science, 35(3), 226-238.

"despite many studies indicating that gender-diverse environments can improve [...] performance, meta-analyses suggest that, on average, the performance benefits of diversity tend to be insignificant or non-existent (Mannix & Neale, 2005; Post & Byron, 2015).

This raises a question: when do gender diverse contexts enhance performance...

In the present research, we propose an answer: the benefits of diversity [to performance] depend on the social context that shapes intergroup interaction"

Within-field disparities are not investigated as much as between-field (e.g., we have documented a million times "women don't major in CS - why?!") but within-field disparities are striking and teach us a lot:

"[students in physics classrooms] engage in patterns of gendered task division (e.g., women perform secretarial work while men work on problems) that result in inequitable learning opportunities between men and women"

This is fundamentally the thing we already know, "you are the only one who can clean up this awful mess you made, don't bring the people you've excluded in to clean it up for you", but worked out in scientific evidence