anjie cao | 曹安洁

I discovered incredible work on menstrual education in China by Ying on LinkedIn, and then realized she's an alum from our own department! Reached out immediately to get her on the podcast. I learned so much and was so inspired! Hope you can enjoy it as well!

oh I didn't know about this! Thank you for letting me know!

thank you :)

thank you!

9/

Special thanks to @thomastalhelm.bsky.social for many thoughtful feedback and conversations. Also thanks to suggestions from @nicholascoles.bsky.social , Xueguang Zhou, Hazel Markus, and members of Culture Lab :)

8/

So, rice theory is a cool idea. But we’re not so sure these two farms give us a clean causal test.

Thanks for reading! And now… I’m craving rice snacks. Strongly recommend these, 10/10 🍘

(It says Milk but it's 56% rice)

7/

And the two farms weren’t identical socially, either. Lianhu (rice) had stronger military roots. You can even glimpse their cultural difference in how they kept records. Lianhu documented every single year; Qukou only every five. At least in the farm chronicles.

6/

What I found was surprising: both farms actually grew both rice and wheat. Sometimes in similar proportions.

By the 1970s (when most participants were born), Lianhu, the “rice” farm, was mostly growing dryland crops.

(Yes I did enter all the datapoints from the book and ggplot it)

5/

Beyond stats, there’s also history. I got curious about the farms themselves and hunted down the primary sources: the farm chronicles.

Thanks to interlibrary loan, Harvard mailed me two actual physical copies across the country (!). TWICE (😅sorry I'm slow and they won't extend my loan).

4/

We found only one of the three psychological measures (Self-Inflation) was robust, showing consistent results across all specifications. The others were sensitive to the specification of the model.

3/

Consistent with rice theory, the rice farmers turned out to be more “collectivistic” than the wheat farmers.

But we took a closer look. Since the analyses weren’t preregistered, we decided to test robustness using a multiverse analysis: rerunning all possible versions of their models.

2/
In Talhelm & Dong (2024), they picked a fascinating case: two state farms in China, one designated for rice, the other for wheat. Young people from the cities were basically randomly assigned to these farms. So voilà, quasi-random rice vs. wheat farming!

1/
Alright, this one’s been sitting in my drawer for a year now, after @mcxfrank.bsky.social and I got turned down for a public commentary.
But before I forget about it completely, here’s the preprint:
Can we harvest insights for rice theory from two state farms in China?

osf.io/preprints/ps...

thank you Mike!!

Thank you Caroline :)

This of course did not mean that children were not getting better with age — but we hope this (somewhat) surprising finding can highlight the need for more robust reporting standards and more large-scale multi-laboratory projects (like ManyBabies!) (9/9)

We investigated each hypothesis, but found none of these explained the lack of age-related growth in most datasets! (8/9)

Hypothesis 4: Positive growth only after infancy. Maybe developmental changes were only observable after some age (e.g. in toddlerhood??) (7/9)

Hypothesis 3: Change in only a subset of conditions. Maybe developmental changes were only supposed to be observed in some specific conditions? (6/9)

Hypothesis 2: Methodological adaptation for older infants. Maybe studies testing older infants were using more difficult methods? (5/9)

Hypothesis 1: Age related selection bias against young children. Maybe studies testing younger infants were more likely to have publication bias? (4/9)

That’s very strange! Shouldn’t the children get better at the tasks as they get older? We came up with 4 hypotheses that can potentially explain the flatness of these curves (3/9)

To our surprise, we found that for most phenomena, there was no (linear) age effect at all — meaning that as children get older, the effect sizes in those tasks did not get larger! (2/9)

MetaLab

Developmental psychology has long studied how constructs change with age, but what are the shapes of these changes? We investigated this question by conducting a meta-meta-analysis over 25 developmental meta-analyses retrieved from metalab: langcog.github.io/metalab) (1/9)

Estimating Age‐Related Change in Infants' Linguistic and Cognitive Development Using (Meta‐)Meta‐Analysis Developmental psychology focuses on how psychological constructs change with age. In cognitive development research, however, the specifics of this emergence is often underspecified. Researchers ofte...

Our meta-meta-analysis is officially out! (w/ Molly Lewis, Sho Tsuji, @chbergma.bsky.social, @acristia.bsky.social, and @mcxfrank.bsky.social!)

Estimating age-related change in infants’ linguistic and cognitive development using (meta-)meta-analysis

onlinelibrary.wiley.com/doi/10.1111/...

6/
We’ll be at CogSci 2025 presenting this work!
Come find us in San Francisco. Happy to chat about all things looking time paradigms :)

5/
Why does this matter?
Habituation and dishabituation are often treated as separate cognitive predictors.
But our findings suggest they may tap into a shared process.

4/
Infants dishabituated more when the stimuli were simpler, and younger infants showed greater dishabituation overall.

3/
Key finding:
Individuals who habituate faster also show stronger dishabituation, in both infants and adults.
For adults, greater volatility in looking behavior during habituation (often seen as noise) also predicted stronger dishabituation.

2/
We analyzed large-scale looking-time data across the lifespan:
– Infants (N = 1986)
– Preschoolers (N = 33)
– Adults (N = 186)
This allowed us to test how attention unfolds over development—and what predicts its recovery when novelty appears.