Linh Nguyễn

@nguyenllpsych.bsky.social

data science @AllianzLife | personality research @UMNPsych | #rstats statistical consulting @ linkedin.com/in/nguyenllpsych

Yihui Xie, who authored or co-authored many widely used R packages including rmarkdown, shiny, knitr, and bookdown, has been laid off by Rstudio/Posit. (His blog post describes the separation as amicable.) He's looking for work, someone is going to be very lucky to get him yihui.org/en/2024/01/b...

Bye, RStudio/Posit! - Yihui Xie | 谢益辉

Who is down? Me. After more than 10 years at RStudio/Posit, the time has come for me to explore other opportunities. A little over two weeks ago, I was told that I was laid off and my last day would &...

yihui.org

whenever i get pass a new academic/professional milestone im simply relieved to never have to do it again because of how rapidly things have become so much amore competitive. excited to one day annoy the youths with my tales of “just cast a wide net and youll get in somewhere”

wrapped up the last meeting this semester with my undergrad RAs by preregistering their research projects, using my dissertation data that they helped collect! it is truly through the shared excitement and discussions over the years (with 14 RAs in total now..) that my research stays alive 😅 #PhDSky

sometimes I would be deep in data analysis or teaching prep and wonder why I have decided against the academic path, but then always face the same rude awakening when the time comes to actually do some ✨writing✨

I felt adequately prepared for much of dissertation life, but definitely not for explaining that most of our participants were "quite young, in the 25-34 age group" and have one research assistant immediately say oh that's not that young ☠️ #PhDSky

Stats people online: "You buffoon, you complete moron, how can you still be using method X? Method Y is better!" Me, inevitably, after spending 5 hours reading and getting method Y to work: Coefficients and SEs change by 0.02, conclusions identical.

in less than a month, 55 romantic couples have signed up for my dissertation study, commiting to a week-long ESM process AND a relationship narrative interview 🤯 there are no-shows and all that jazz, but definitely more confident now about that data collection target #PhDSky

going from a full-day employee recognition celebration at my industry gig yesterday back to the classic grad-student combo of teaching + learning + dissertating today has been … interesting. might actually miss all this chaos when i graduate #PhDSky

As I said elsewhere: What's supposed to be surprising about arriving at different results under different models making different assumptions, using different methods that also make different assumptions? It's almost as if our inference depends on the model & method. www.nature.com/articles/d41...

Reproducibility trial: 246 biologists get different results from same data sets

Wide distribution of findings shows how analytical choices drive conclusions. Wide distribution of findings shows how analytical choices drive conclusions.

nature.com

#rstats What are your favourite examples when teaching students to think critically about data/figures and interpretations? For example, Simpson's Paradox and the importance of considering important subpopulations - is brilliantly displayed in the Palmer Penguins dataset.

A negative correlation between bill length and bill depthA positive correlation between bill depth and bill length when appropriate sub-populations are considered