Yi Li @einsley23 Twitter profile | Pikagi

Pikagi

Yi Li

@einsley23

228

Followers

181

Following

34

Media

335

Statuses

PhD candidate @McGillU . @Harvard @UNC alum | Epidemiology, biostatistics, causal inference, machine learning

🇨🇦

https://t.co/2wlD4gUYrD

Joined July 2016

Don't wanna be here? Send us removal request.

Pinned Tweet

@einsley23

Yi Li

1 year

The causal mediation analysis methods paper with Maya Mathur, @DanielHSolomon , Paul Ridker, Bob Glynn and @kaz_yos is finally published in Epidemiology!

Tweet card media

Effect Measure Modification by Covariates in Mediation:... : Epidemiology

ts may be modified by baseline characteristics and are interested in how the resulting mediation effects, such as the natural direct effect (NDE), the natural indirect effect. (NIE), and the propor...

journals.lww.com

4

6

34

Last Seen Profiles

@StwGendut

@BinorRaja

@violzetas

@stw_pdg

@bokeplokalmalam

@stw_pdg

@AutomaticEpps

@stw_pdg

@rrtyd124

@JhonatanLo41767

@BehlmJani

@LPSaMSumireSai1

@narindraadithya

@hochmm22

@DG_TAKAMATSU

@imgbyj

@r1ta___

@angvsthewrld

@StwGendut

@sunbombb

@trader_girl1112

@whoistiajha

@BobbyLiche

@ShannonDom95614

@AhmedBat44

@bokeplokalmalam

@ibudesu0407

@AzatCarolin

@Sbongakonkenzi5

@stw_yang

@vhsbled

@AnyhYmnyh40403

@GameGeneQuiz

@lelvrxy

@yeonzzmin

@Pti108

@einsley23

Yi Li

6 months

昨日のロサンゼルス空港でのファンのビデオレポを見て、Number_iの身の安全について心配になった。海外のファンも日本と同じようなルールを守ることが大切だと思います。Number_iさん、すべてのファンのリクエストに応えないのは、まったく理解できることなので、遠慮なくしてください。

12

340

2K

@einsley23

Yi Li

2 years

If I propose a novel idea (e.g. new method or new perspective) *orally* when chatting w someone, and that person wants to use my idea in a future paper, should I be a coauthor on it? Or should they ask for my permission first? #AcademicTwitter @AcademicChatter

65

31

311

@einsley23

Yi Li

2 years

A review paper on causal mediation and R package 'regmedint' tutorial with Maya Mathur, @kaz_yos , and Jay Kaufman is now online: A brief primer on conducting regression-based causal mediation analysis. @APA_Journals @APA_Databases #EpiTwitter

5

23

87

@einsley23

Yi Li

6 months

@kaiyunomm1 感想を聞かせてくれてありがとう。彼らや滝沢さんを含め、より多くの人にこのメッセージを見てほしい。彼らが去ろうとしている空港がどうなってしまうのか、とても心配です。

0

4

38

@einsley23

Yi Li

6 months

@jun425508561 はじめまして、コメントありがとう。もっと警備を強化することは確かに必要だ。こんなに少なかったとは驚きです。

0

2

32

@einsley23

Yi Li

2 years

On a different note, just received the SER abstract acceptance notification. And it’s an oral presentation! @robertwplatt @kaz_yos #SER2023

Tweet media one

2

1

31

@einsley23

Yi Li

6 months

@nennr_77 コメントありがとう。サインをもらうために子供を連れてくるなど、一部のファンがやっていたことは馬鹿げていた。あれはクレイジーだし、受け入れられない。

0

0

29

@einsley23

Yi Li

2 years

@drmichaelhooper @AcademicChatter I hate doing this too. But if everyone is crystal clear about authorship (i.e. always ask for others' approval), I believe more people will be more willing to share

2

0

29

@einsley23

Yi Li

6 months

@Maiiiii809 それはわかるが、警備員が少ないのも合理的ではなかった。number_i さんの周囲には何のプロテクトもなく、ファン側にも何の制限もないため、状況は制御不能だった。

0

0

26

@einsley23

Yi Li

6 months

@appy65540165 ありがとう。この投稿をしてよかった。心配しているのは私だけだと思っていた。ファンとして他に何ができるかわからない。

0

1

19

@einsley23

Yi Li

2 years

The article will be publicly available next year. Feel free to email me if you want a copy.

Tweet media one

@einsley23

Yi Li

2 years

A review paper on causal mediation and R package 'regmedint' tutorial with Maya Mathur, @kaz_yos , and Jay Kaufman is now online: A brief primer on conducting regression-based causal mediation analysis. @APA_Journals @APA_Databases #EpiTwitter

5

23

87

0

2

15

@einsley23

Yi Li

5 months

"The Complex Estimand of Clone-Censor-Weighting (CCW) When Studying Treatment Initiation Windows" on arXiv: Led by: Michael Webster-Clark w/ Sophie Aniello and @robertwplatt Studying the effect of a regimen like "starting a treatment within a specific

Tweet card media

The Complex Estimand of Clone-Censor-Weighting When Studying...

Clone-censor-weighting (CCW) is an analytic method for studying treatment regimens that are indistinguishable from one another at baseline without relying on landmark dates or creating immortal...

1

1

14

@einsley23

Yi Li

6 months

世界中のアーティストたちは、旅行や公の場に姿を現すとき、常に何人ものボディーガードを従えている。私たちも同じレベルに達することを切に願う。そうでなければ、危険はそこにある。 @h_Takizawa329 @tobeofficial_jp #Number_i #Number_i_Coachella2024 #平野紫耀 #神宮寺勇太 #岸優太

1

4

13

@einsley23

Yi Li

2 years

@PWGTennant @AcademicChatter Thank you! I find this extremely tricky (fuzzy?) in some so-called casual stats consulting matters. Glad to hear from expert's advice!

2

0

12

@einsley23

Yi Li

6 months

手紙はOKですが、プレゼントはNGです。念のため、手紙以外のものは受け取らないようにしてください！このメッセージをすべてのアーティストに広めてください。 @tobeofficial_jp @_yuta_kishi_ @number_i_staff

1

4

12

@einsley23

Yi Li

1 year

This paper is now publicly available on PubMed:

Tweet card media

A Brief Primer on Conducting Regression-Based Causal Mediation Analysis

We provide an overview of regression-based causal mediation analysis in the field of traumatic stress and guidance on how to conduct mediation analysis using our R package regmedint.We discuss the...

www.ncbi.nlm.nih.gov

@einsley23

Yi Li

2 years

A review paper on causal mediation and R package 'regmedint' tutorial with Maya Mathur, @kaz_yos , and Jay Kaufman is now online: A brief primer on conducting regression-based causal mediation analysis. @APA_Journals @APA_Databases #EpiTwitter

5

23

87

1

1

10

@einsley23

Yi Li

1 year

Agreed. I met a statistician at a conference. He greeted me bc he saw me on the train days ago. I asked him if he's a professor, he said "no, I'm a failure" just bc he's not good at writing papers! Many PhDs think not becoming a professor makes them a failure, which is untrue.

@robertwplatt

Robert Platt

1 year

Very well said. Every individual case is different, depending on the person, field, jobs available, other constraints, etc. I always encourage trainees to figure out the best path for them, wherever it leads them. There is no “perfect career”.

0

2

13

0

0

8

@einsley23

Yi Li

1 year

"Linear algebra is a beautiful subject --- and teaching is a beautiful life."

Tweet media one

0

0

7

@einsley23

Yi Li

2 years

Discussing authorships openly before working on projects saves lots of trouble. Once authorship is clear, everyone does the work based on the credit they’ll receive. I find this an easier & fairer process than assigning authorships after work is done based on % of contribution

@DrMargaretCrane

Margaret E. Crane, PhD

@DrMargaretCrane

2 years

PSA: if you ask a friend or a student for stats help on a paper that requires more than an hour of their time, consider adding them as a co-author to the paper

23

115

2K

1

0

8

@einsley23

Yi Li

2 years

My reviewer journey starts like this. #AcademicTwitter @GrumpyReviewer2 @AcademicChatter

Tweet media one

Tweet media two

1

1

8

@einsley23

Yi Li

2 years

Super excited about the bayesian adaptive trials seminars lined up over the next few weeks at McGill EBOH

Tweet media one

1

0

8

@einsley23

Yi Li

6 months

@2529Aki 滝沢社長がこのツイッターをご覧になったかどうかはわかりませんが、今日のパフォーマンスでボディーガードを増やすのは遅すぎると思います。

1

1

8

@einsley23

Yi Li

2 years

@WalidSoneidar @kaz_yos @APA_Journals @APA_Databases The original Valeri & VanderWeele (2013) paper only have SAS, Stata and SPSS macros. @kaz_yos and I implemented the program in R, but beyond that we also extended Valeri & VanderWeele's method by accommodating effect modification terms

0

0

7

@einsley23

Yi Li

2 years

I wonder how others think about really speedy peer reviews. The one so speedy that doesn't have patience to collect all 3 reviews even when it is still in the 14-day review window. #epitwitter @AcademicChatter

3

1

8

@einsley23

Yi Li

2 years

#epitwitter Impact of lowering fine particulate matter from major emission sources on mortality in Canada: A nationwide causal analysis | Proceedings of the National Academy of Sciences

Tweet card media

Impact of lowering fine particulate matter from major emission sources on mortality in Canada: A...

Emissions of fine particulate matter (PM2.5) from human activities have been linked to substantial disease burdens, but evidence regarding how redu...

1

2

6

@einsley23

Yi Li

2 years

Charlie Poole: a way to avoid misspelling "causal inference" is to search "casual" when you finish the doc.

1

1

6

@einsley23

Yi Li

2 years

@furkann_kir @AcademicChatter If not entirely relevant, cite the book in Discussion where you can talk about irrelevant things.

0

0

7

@einsley23

Yi Li

2 years

@AmanBughio @AcademicChatter But it's impossible to avoid brainstorming ideas together w friends and colleagues from time to time. It's not like I know someone is writing a paper and they consult me. In that case, I'd confirm authorship first.

1

0

6

@einsley23

Yi Li

1 year

Greenland, S., Pearl, J., & Robins, J. M. (1999). Confounding and collapsibility in causal inference. Statistical science, 14(1), 29-46.

0

2

7

@einsley23

Yi Li

2 years

Valeri & VanderWeele (2013) approach do not support EMM in mediator or outcome model. In this paper (), we generalize their approach by including exposure-covariate and mediator-covariate product terms.

Effect Measure Modification by Covariates in Mediation: Extending Regression-Based Causal Mediation...

In this paper, we generalize the closed-form regression-based mediation analysis approach proposed by Valeri and VanderWeele (2013, 2015) to accommodate effect measure modification by the covariates....

@kaz_yos

Kazuki Yoshida

3 years

OSF Preprints | Effect Measure Modification by Covariates in Mediation: Extending Regression-Based Causal Mediation Analysis Our corresponding R package is available on CRAN:

Tweet media one

0

4

21

1

1

6

@einsley23

Yi Li

2 years

Dear @yudapearl , I am reading your 2015 paper: I have a hard time understanding why Y_x still depends on S, given Z? This seems not supported by d-separation rules. For “natural dependence of salary on training”, should there be a path from S to Y? Thx!

Tweet media one

1

0

5

@einsley23

Yi Li

2 years

Just reviewed a meta analysis manuscript. The authors said why they didn’t use the valid measure is because the original individual studies didn’t use the valid one. Me: then you probably should not have done this meta analysis.

@GidMK

Health Nerd

2 years

6/n And remember, ALL systematic reviews are only as good as the research they include If all of the papers on a topic are bad, the eventual results that you get from aggregating those bad papers together are pretty meaningless

5

16

266

0

1

6

@einsley23

Yi Li

2 years

If you can't access the full article for free at this point, please see the OSF preprint version:

A Brief Primer on Conducting Regression-Based Causal Mediation Analysis

© 2022, American Psychological Association. This paper is not the copy of record and may not exactly replicate the final, authoritative version of the article. Please do not copy or cite without...

0

1

6

@einsley23

Yi Li

2 years

It's a pity I can't make it this year because of the super long waiting time for a US visa, but glad all my brilliant advisors, i.e. @kaz_yos @robertwplatt and Jay Kaufman are there and will keep me updated!

@kaz_yos

Kazuki Yoshida

2 years

#SER2022 Poster 612 "Impact of model misspecification on proportion mediated in causal mediation analysis" by @einsley23 will be displayed on Thursday (unfortunately by me). I'll sleep under the poster to reach Yi's depth of understanding.

Tweet media one

2

3

43

0

0

6

@einsley23

Yi Li

2 years

The coldest day (so far) since I came here

Tweet media one

0

1

6

@einsley23

Yi Li

2 years

Mar. 13 McGill Epi seminar with Eric Tchetgen Tchetgen — An Introduction to Proximal Causal Inference

0

1

4

@einsley23

Yi Li

2 years

I have 30 conference abstracts & 2 manuscripts pending review assignments to complete. #EpiTwitter

0

1

6

@einsley23

Yi Li

2 years

Excellent paper on selection bias by my UNC classmate!

@LuHaidong

Haidong Lu

2 years

Selection bias is often a challenging concept in epidemiology. One feature that distinguishes selection bias from confounding and measurement bias is, selection bias results from a change in the sample under study. It leads to a critical but, surprisingly, unsolved question. 1/

4

39

209

0

0

5

@einsley23

Yi Li

9 months

#Number_i_GOAT 以前とは全く違うスタイルはどうだろう、jpop全体の中でも異彩を放っている‼️ 一日中一曲ループしている。メロディーもどんどん良くなっているし、ダンスも相変わらず上手い。野心と強さを感じるし、これからも応援していきたい🥰

Tweet card media

Number_i - GOAT (Official Music Video)

Number_i - GOAT (Official Music Video)Stream “GOAT” now.►https://tobe-music.lnk.to/GOAT_EP The members of Number_i are Sho Hirano, Yuta Jinguji and Yuta Kish...

www.youtube.com

0

1

5

@einsley23

Yi Li

2 years

Last week I tried using #ChatGPT to write a literature review for an assignment. I asked it to explain a certain mechanism and list some references. The explanations worked well, but none of the references actually exists.

3

0

4

@einsley23

Yi Li

2 years

@slavov_n Both are important. A good school provides a good platform equipped with more resources that allow you to do the work in a better way.

0

0

5

@einsley23

Yi Li

10 months

Hi @yudapearl I'm reading your 2012 technical report on effect heterogeneity. Could you elaborate on why ETT and ETU can differ in RCTs, and what "the choice of treatment" means? Why are E[Y|X = 1]≠E[Y_1], and E[Y|X = 0]≠E[Y_0]? Thanks!

Tweet media one

Tweet media two

0

0

4

@einsley23

Yi Li

1 year

I did most of my thinking of methodological problems when walking outside or taking a shower. My brain just doesn’t function when I sit in front of a computer.

@ShengwuLi

Shengwu Li

1 year

One of the beautiful things about doing theory is how much of the work can be done on a walk outside.

9

25

400

1

0

4

@einsley23

Yi Li

8 months

Reviewing the STAT 286/GOV 2003 material from 2021. Key papers on ITR for my own reference: 1. Imai, K., & Li, M. L. (2023). Experimental evaluation of individualized treatment rules. Journal of the American Statistical Association, 118(541), 242-256.

Tweet card media

Experimental Evaluation of Individualized Treatment Rules

The increasing availability of individual-level data has led to numerous applications of individualized (or personalized) treatment rules (ITRs). Policy makers often wish to empirically evaluate IT...

www.tandfonline.com

1

2

4

@einsley23

Yi Li

2 years

The accompanying R package: Yoshida K, Li Y (2022). regmedint: Regression-Based Causal Mediation Analysis with Interaction and Effect Modification Terms. R package version 1.0.0, .

@kaz_yos

Kazuki Yoshida

2 years

"Effect Measure Modification by Covariates in Mediation: Extending Regression-Based Causal Mediation Analysis." This work by @einsley23 , which laid the foundation for the R package regmedint, was finally accepted by a journal!

0

4

22

1

0

3

@einsley23

Yi Li

1 year

@EpiEllie Same in MTL: Wed/Thu/Fri 34°C, Sat – next week: 20°C

0

0

2

@einsley23

Yi Li

2 years

@olga_basso The timeline is messed up and new tweets might be buried. "Following" says "show N tweets" every few seconds, and clearly those N tweets are posted hours ago, not in those few seconds

0

0

4

@einsley23

Yi Li

2 years

@olga_basso "For you" shows what people you follow do, e.g. comment/reply/retweet. "Following" shows posts only, but includes some retweets as well, so not sure what's the distinction

1

0

4

@einsley23

Yi Li

2 years

Tree handicrafts made by yesterday's freezing rain

Tweet media one

0

0

3

@einsley23

Yi Li

1 year

@PWGTennant I arrived at PDX! Looking forward to seeing you around soon!

0

0

3

@einsley23

Yi Li

2 years

Agreed that the time for lecture prep is extremely underestimated. Preparing for a 1-hr journal club (I'm not the presenter, just a moderator) already took me 3 hrs since this afternoon, and I haven't quite finished yet.

@PWGTennant

Peter Tennant is on BlueSky

2 years

The suggestion that it takes 1-hour to prepare a 1-hour lecture is a complete fantasy. University leaders use this fantasy figure because it allows them to pay below the minimum wage, by claiming that it 'should only take an hour'. It's morally repulsive. And should be illegal.

43

282

1K

1

0

3

@einsley23

Yi Li

2 years

@kaz_yos @louisahsmith @AdanZBecerra1 @societyforepi Thanks Louisa!!

1

0

3

@einsley23

Yi Li

2 years

@robertwplatt You can impute the outcome by lecturing without slides

0

0

3

@einsley23

Yi Li

2 years

@na_tonge @AcademicChatter Only list the top-tier ones

1

0

3

@einsley23

Yi Li

2 years

@Cao_Sifan @AcademicChatter It's tricky if the other person is senior to us. I was in similar situation before: I worked (solo) on a project for someone for free. After I sent my results via email he never replied. What's worse is later that person published my results without even listing me as a coauthor

4

1

3

@einsley23

Yi Li

2 years

@re_sieber @mcgillu 9 Celsius is very hot! Since I moved to Canada any temperature above 0 = hot.

0

0

3

@einsley23

Yi Li

2 years

I still keep these receipts from 3 years ago

Tweet media one

0

0

2

@einsley23

Yi Li

2 years

@ewrigleyfield @Afaf_Saliba @AcademicChatter I feel Afaf. Especially for students, coming up with a new idea is difficult. We don't want it to be easily grabbed by others. There is nothing wrong with protecting ourselves in a tough environment

0

0

3

@einsley23

Yi Li

6 months

もう少しボディーガードを追加するに関するいくつかの考え👇: 1. アーティスト1人あたり、平均して1人のボディーガードを確保してください。例えば、Number_i さんには3人のボディーガードが、IMP.

0

1

3

@einsley23

Yi Li

1 year

I stopped confusing between G-formula and G-estimation after taking EPI 289.

@PronouncedKeil

Alex Keil

@PronouncedKeil

1 year

If you know someone who (rightfully) confuses g-estimation with g-computation or the g-formula, then you can send them this screenshot of footnote 46 in Richardson and Robins 2013. It will transfer the confusion toward your insult that compares them to a Turing award winner.

Tweet media one

3

1

27

1

0

3

@einsley23

Yi Li

2 years

Watched a few years ago. Love the silence after “Three patients per group or in total?” “Yes”.

@Richard_D_Riley

Richard Riley (R²)

@Richard_D_Riley

2 years

I think this may be the greatest video I've ever seen - resonates very strongly! And I love the ending.

20

25

133

0

0

3

@einsley23

Yi Li

2 years

Joint work with Health Canada, UofT, UCSD, etc: Chen et al: Using Parametric g-Computation to Estimate the Effect of Long-Term Exposure to Air Pollution on Mortality Risk and Simulate the Benefits of Hypothetical Policies @EHPonline #epitwitter

Using Parametric g-Computation to Estimate the Effect of Long-Term Exposure to Air Pollution on...

Background: Numerous epidemiological studies have documented the adverse health impact of long-term exposure to fine particulate matter [particulate matter ≤2.5��m in aerodynamic diameter (PM2.5)] on...

ehp.niehs.nih.gov

0

1

2

@einsley23

Yi Li

6 months

私も同じ気持ちだ。しかし現実は、無作法な人たちにより良い行動を期待することはできない。彼らには何の制約もなく、何をやっても何の結果にもならない。私たちファンが改善するためにできる唯一のことは、次回はもっとボディーガードを派遣するようTOBEにお願いすることだ。 @h_Takizawa329

@RiunaSnow

劉那❄︎

6 months

@einsley23 @_yuta_kishi_ @number_i_staff 引用失礼しました。押し付けられ、差し出された物を反射的に受け取ってしまってる感もありますが、手紙以外のものは国内外関係なく渡すのも受け取るのもNGだと思います。渡す側も“全く知らない人”から物を受け取るのが恐怖に感じるように、彼らも恐怖にさらされている事を知って欲しいです。

0

0

0

0

0

3

@einsley23

Yi Li

1 year

2 yrs later I logged in Piazza again. Miss Gil Strang so much

Tweet media one

0

0

2

@einsley23

Yi Li

2 years

Every sentence ChatGPT said about our mediation package is wrong, but the funniest is when it said it's an old package and last updated more than 10 yrs ago @kaz_yos

Tweet media one

0

0

2

@einsley23

Yi Li

2 years

lag() is the most underrated function.

0

0

2

@einsley23

Yi Li

2 years

Hi @fishingwithzack , I'm using robumeta package and predict() gave an error. Someone else also ran into this issue last July . Would you mind fixing it? Thanks a lot.

Tweet media one

1

0

2

@einsley23

Yi Li

2 years

1

0

2

@einsley23

Yi Li

2 years

@PWGTennant @AcademicChatter That's not good😥 I hope there is a clear rule to avoid this.

0

0

2

@einsley23

Yi Li

2 years

@AcademicChatter The first time, I didn't submit my review in 6 days, and editor said they already have enough information from 2 reviews. The second time, I submitted in 5 days, although slower than reviewer 1, and they again didn't wait for reviewer 3 and proceeded with the two of us.

2

0

2

@einsley23

Yi Li

2 years

It’s a real danger to use ChatGPT when we ourselves aren’t able to evaluate if the answer is correct or not. Dumping a bunch of terminologies together makes it look like it knows a lot which in fact can be BS.

@PHuenermund

Paul Hünermund is on Bluesky

2 years

My attempt with the same query leads to an answer that is significantly worse. GPT is throwing together various concepts (backdoor, mediators, etc.). How could you ever get something useful out of the response if you don't already know the answer?

Tweet media one

8

2

21

1

0

2

@einsley23

Yi Li

2 years

@PWGTennant @robertwplatt @ehsan7x Can I join your chat at SER2023😃

0

0

2

@einsley23

Yi Li

2 years

@statsepi He and the other guy with a weird profile pic often appear together. I saw his post yesterday in the “For you” tab, and then I immediately switched back to the “Following” tab.

0

0

1

@einsley23

Yi Li

2 years

Wager, S., & Athey, S. (2018). Estimation and inference of heterogeneous treatment effects using random forests. Journal of the American Statistical Association, 113(523), 1228-1242.

Tweet card media

Estimation and Inference of Heterogeneous Treatment Effects using Random Forests

Many scientific and engineering challenges—ranging from personalized medicine to customized marketing recommendations—require an understanding of treatment effect heterogeneity. In this article, we...

www.tandfonline.com

0

0

1

@einsley23

Yi Li

1 year

Another greatest MIT professor who taught me a lot is Patrick Winston. Sadly I didn't have a chance to meet him in person. Talk on 'How to speak': 6.034 lecture videos:

Tweet card media

1. Introduction and Scope

MIT 6.034 Artificial Intelligence, Fall 2010View the complete course: http://ocw.mit.edu/6-034F10Instructor: Patrick WinstonIn this lecture, Prof. Winston in...

www.youtube.com

0

0

2

@einsley23

Yi Li

2 years

@statsepi @SarahDRasmussen Just found now has a twitter handle @diddukewin 😁

0

0

1

@einsley23

Yi Li

2 years

Excellent advice on presentations. "Strive to be the most informative person in the room about the limitations of the model and the results". "If you have enough equations that Beamer is easier to use, you have too many equations."

@ShengwuLi

Shengwu Li

2 years

Giving good talks is a core skill for academics. It's learnable! Here's how I approach mine.

17

274

1K

0

0

2

@einsley23

Yi Li

2 years

It’s just like human — throwing together fancy terminologies doesn’t make one more correct. On the contrary, I’m more cautious when I see someone talks like that because that is a sign of bluff.

1

0

2

@einsley23

Yi Li

2 years

Really excited about tomorrow's journal club (I didn't know Jamie's first job was assistant professor in occupational medicine)

Tweet media one

1

0

2

@einsley23

Yi Li

2 years

@rik_epi Glad you are also enjoying Bayesian! I’m learning Bayesian adaptive trial this semester. Really fascinating area!

1

0

2

@einsley23

Yi Li

2 years

@PWGTennant @Doc_Courtney @Tfeend Degree of transparency is the same if one knows what he is doing. I used google slides to daw DAGs because they look prettier than Dagitty (mainly because of the color). Aesthetics doesn't matter for understanding the concept, but is important for presenting to others.

3

0

2

@einsley23

Yi Li

5 months

@SJKim_Psych regmedint can do this! In emm_ac_mreg, emm_ac_yreg, emm_mc_yreg, you can specify as many moderators as you want in either mediator or outcome model. Please see the example (regmedint_obj2) at :)

Tweet card media

GitHub - kaz-yos/regmedint: R implementation of effect measure modification-extended regression-b...

R implementation of effect measure modification-extended regression-based closed-formula causal mediation analysis - kaz-yos/regmedint

2

0

2

@einsley23

Yi Li

2 years

@Doc_Courtney @PWGTennant @Tfeend Yeah, and before clarity, the first goal is to attract readers to read.

0

0

2

@einsley23

Yi Li

1 year

It's very difficult to stop the coding work once I start it. I'll keep working on it forever. If I'm not typing I'll be thinking about it until I'm disturbed by something.

1

0

2

@einsley23

Yi Li

6 months

速い！

@number_i_staff

Number_i_staff

@number_i_staff

6 months

📺 TV Appearance 📺 April 12th 8pm JST Number_i will appear on TV Asahi's "Music Station" 2-Hour Special! #Number_i_M ステ

Tweet media one

581

10K

41K

0

0

1

@einsley23

Yi Li

7 months

@number_i_staff @_yuta_kishi_ #Number_i #ShoHirano #YutaJinguji #YutaKishi #平野紫耀 #岸優太 #神宮寺勇太

Tweet card media

Number_i · GOAT · Song · 2024

open.spotify.com

0

0

2

@einsley23

Yi Li

2 years

@Apoorva__Lal @lorenzoF_SM How about also len() to avoid confusion about R vs python?😅

0

0

2

@einsley23

Yi Li

2 years

Great but very intense 4-day summer course with Dr. Samy Suissa. Did 2 presentations on Monday and Today. Yesterday I was the discussant. It’s literally like attending a pharmacoepi conference. The only difference is tomorrow afternoon there is still a 3-hour final exam🤣

Tweet media one

0

0

0

@einsley23

Yi Li

3 months

@tobeofficial_jp @number_i_staff こちらから連絡して申し訳ありません。写真が常に売り切れているのは、クリアバッグに入れるのに時間と手間がかかるためでしょうか？未収納のまま送っていただいても構いませんが、売り切れは多くの人が困っています。ご配慮ありがとうございます。 @h_Takizawa329

0

0

2

@einsley23

Yi Li

2 years

@louisahsmith @kaz_yos @AdanZBecerra1 @societyforepi Definitely! Can't wait to meet y'all in person!

0

0

2

@einsley23

Yi Li

2 years

@WalidSoneidar @kaz_yos @APA_Journals @APA_Databases The EMM extended methods manuscript is still under revision, but you could take a look at

Effect Measure Modification by Covariates in Mediation: Extending Regression-Based Causal Mediation...

In this paper, we generalize the closed-form regression-based mediation analysis approach proposed by Valeri and VanderWeele (2013, 2015) to accommodate effect measure modification by the covariates....

0

0

2

@einsley23

Yi Li

2 years

@reperegrination Thx. Yeah I would not actually accuse, but glad to know many academia folks have the integrity and consciousness.

0

0

2

@einsley23

Yi Li

2 years

@yudapearl @soboleffspaces @ObservStudies @HarvardEpi @EpiEllie From SWIG, the conditional independencies can be directly read off, so g-formula can be easily written out, while in DAG the interventional (counterfactual) nodes are not in the graph.

2

0

2

@einsley23

Yi Li

2 years

@olga_basso Perhaps there is a bug or you misclicked something. It's in your Profile setting where you can make your account a professional one. It's free. If you want to promote your tweets, that costs money. I am using it and can see the promote button too.

2

0

2

@einsley23

Yi Li

2 years

Proud of you, @AmBeRnIgAm !

@AmBeRnIgAm

Amber

2 years

Tweet media one

1

0

5

1

0

2

@einsley23

Yi Li

2 years

@ShyCatBanshee @AcademicChatter Good point. But still need to ask the proposer for permission. I'd be happy enough if they ask me if I want to contribute a bit more to be a coauthor

0

0

2

@einsley23

Yi Li

2 years

@megikaya Just Jupyter notebook? Or do you need GPU?

1

0

2