@letsnotbehayest It is a pain to explain what "negative binomial" means to readers. A Poisson-gamma mixture is way better for understanding😉.
English
Haoran Li | 李浩然
36 posts
Haoran Li | 李浩然
@HaoranQME
Assistant Professor @umn_edpsych | MLMs, SCEDs. https://t.co/ZYiasjkeZN
Minneapolis, MN Katılım Haziran 2017
205 Takip Edilen86 Takipçiler
It is only our hubris that ever led us to name it “the negative binomial distribution” instead of “the Poisson distribution with wiggly gamma means.”
English
Haoran Li | 李浩然 retweetledi
This year I have the honour of serving the Psychometric Society as its President, and we have been working hard on the program for the International Meeting of the Psychometric Society (IMPS) psychometricsociety.org/imps-2025 You can submit an abstract here: imps2025.exordo.com
English
Haoran Li | 李浩然 retweetledi
Up to 169 IES contracts have been canceled, threatening critical education research. This is a major setback for educators & children with disabilities who rely on evidence-based practices. Urge Congress to protect IES research! #4SpecialEducation
exceptionalchildren.org/takeaction?vvs…
English
Haoran Li | 李浩然 retweetledi
IMPS 2025 Abstract Submission
psychometricsociety.org/post/imps-2025…
The Psychometric Society Invites You to Submit an Abstract for IMPS 2025!
Submission System Opening: January 29, 2025
Abstract Submission Deadline: February 28, 2025
English
See more racist clowns in academia.
Zhiding Yu@ZhidingYu
Mr. @pmddomingos This is a country whose leader blatantly says "We lied, we cheated, we stole… we had entire training courses." And thus there's conceited clown like you to spread China hate everywhere. Your self-imagined star-spangled awesomeness doesn't change the fact that Chinese researchers have become a major force in the AI community. China is also leading in industry general autonomy, robotics and AI applications. Your word can't change this fact and the successes don't come with fraud. If you think there’s a problem with this, there’s a problem with you.
English
@SolomonKurz Our model is very basic in the introduction paper 😉. Both TMB and Adaptive packages have convergence issues. brms have great potential to fit far more complex zero-inflated models for SCEDs!
English
Here's an example of four models applied to the data (5 cases in a multiple-baseline AB design). I fit the models with #brms and computed the fitted lines with #tidybayes.
English
Check out @HaoranQME et al for a nice introduction to fitting zero-inflated count models to single-case reversal data (researchgate.net/profile/Haoran…). The authors went the extra mile and made their data open on the OSF (osf.io/e4tp5/). #RStats
English
Very excited that we are awarded with this NIMH R01 grant to study adolescents' social media management strategies and their mental health! Thanks @jcmichener for working on this news story!
fsos.umn.edu/fsos-assistant…
English
Haoran Li | 李浩然 retweetledi
Every person who you mentor is a unique individual, and none of them are you.
Mentoring isn’t about sharing what you would do, but exploring what’s best for them to do.
English
Thrilled to share that I’ve been honored with the Distinguished Early Career Faculty Award from #VCUSOE. Thanks to everyone who has supported me along the way! facebook.com/share/p/2tKK4X…
English
SO HAPPY for our home team!!!! First time ever in history for the Timberwolves to have a 4-0 in a playoff season. Such a good time to be a Minnesota Timberwolves fan
Minnesota Timberwolves@Timberwolves
🧹🧹🧹🧹
English
Who all is going to M3 this summer? I'll be there!
Betsy McCoach@BetsyMcCoach
Nope! Not that kind of model! The 2024 Modern Modeling Methods conference is June 24-26, 2024 @UCONN. For more information or to register for the conference, go to modeling.uconn.edu @quantitudepod @JkayFlake @doinkboy - Can you help spread the word?
English
Unleash the full potential of GLMMs with SCED count data, now dealing with both overdispersion and zero-inflation. Check it out in Behavior Research Methods #sced link.springer.com/article/10.375…
English
@BrittanyHottPhD Thank you for you interest in my work! I would like to chat about the opportunity for a guest lecture. Fell free to email me: haoranli@umn.edu.
English
@Haoran0915 This is cool! I am excited to read and add to our advanced single case design course. Any chance you might be interested in a guest lecture?
English
Finally out in Psychological Methods. Interested in using GLMMs to handle count and proportion outcomes in SCEDs? Check this out!
psycnet.apa.org/record/2024-00…
English
@xiaoransunpsu As it converges, you don't need to use a sandwich estimator. But you can play with type = "CR2" or AR1 structure to see if the vcovCR works. It seems that the function only works without specifying any R side covariance structure.
English
@Haoran0915 Thanks! I ran the three-level model based on your suggestions and it actually worked!!! Although with vcovCR (model_try, type="CR0") I still got the same error message...
English
@xiaoransunpsu If not converge, try not specify the working corr matrix at all (comment out in the model as you did) but keep the three level model. If still not converge, then consider remove the family level. The original error may be due to inconsistent grouping factor in Corr and VcovCR.
English
@Haoran0915 That's a good suggestion -- thanks! I chose Sandwich estimator at first because the family level nesting is very rare (only occasionally two participants nested within a family) so worried the 3-level models may not converge. I'll try that!
English
@xiaoransunpsu vcovCR(model_try, type="CR0"). I think It can extract grouping factor(s) of your fitted model automatically.
English
@xiaoransunpsu I assume you have repeated measures nested within participants. Some participants are from the same family. So the following code may work.
random = ~ -1 + W + S | family_id/participant_id
corr=corCompSymm(form = ~time | family_id/participant_id)
English