angela - Network

A few fun stats/computation things that I've been playing with lately:💡 Hugely grateful for @vincentab's {marginaleffects} R...

A few fun stats/computation things that I've been playing with lately:

💡 Hugely grateful for @vincentab's {marginaleffects} R package and incredible new book! It really makes going from model ➡️ insight more straightforward - no more putting regression tables in presentations, just use this package!

👩‍✈️ Asking Github Copilot Chat explain #opensource code/implementations to me, making it more straightforward to 1) understand and 2) incorporate new statistical techniques into my own work

6) Gratitude for the mentoring relationship and the ability to shape a new scholar's path: my student's enthusiasm, hard work, and...

6) Gratitude for the mentoring relationship and the ability to shape a new scholar's path: my student's enthusiasm, hard work, and eager questions made me remember one of the reasons I'm here in the first place.

Especially as I continue in a phase where I focus on my own research agenda as a primary contributor, it was good to have this #mentoring experience and confirm that I want to do this in the long term.

#Grateful for the opportunity, and big thanks to the ReMatch+ staff this summer! 👏✨

5) Seeing mentoring as a skill to practice, and improve in community with other mentors: I experimented with #comentoring with amazing...

5) Seeing mentoring as a skill to practice, and improve in community with other mentors:

I experimented with #comentoring with amazing scholar Christina Pao, and we shared resources and tips with each other over the summer. I'd team up with another person to co-mentor future students since it allows for knowledge transfer and reduced burden.

The cohort structure of the program also helped - I heard ideas from engineers & scientists about their lab structures/practices that supported research.

4) Realizing that my own mentors were human too: since I was juggling my own research and deadlines with mentoring, there were times where I...

4) Realizing that my own mentors were human too: since I was juggling my own research and deadlines with mentoring, there were times where I was busy, distracted, or tired -- that had absolutely nothing to do with the quality of the work that my mentee was doing.

I reflected on all the times that I thought my mentors might secretly hate my research bc they were busy or distracted. But they likely just had a lot going on at the time!

Being on the other side helped me gain perspective on this.

3) Benefiting from the different worldviews of my mentee to deepen our joint questioning.We worked together on research that explored...

3) Benefiting from the different worldviews of my mentee to deepen our joint questioning.

We worked together on research that explored contextual factors in an administrative dataset. In this case, our different backgrounds really helped us approach the work from multiple angles.

One example: I'm from the East Coast (w/stints in the Midwest), but my mentee was from the West Coast. When they used their home in California as a case study, I learned a lot about diff social service/gov contexts.

2) Making my own knowledge / resources explicit: I realized that I've developed my own "toolkit" over time for doing research:...

2) Making my own knowledge / resources explicit: I realized that I've developed my own "toolkit" over time for doing research: rules of thumb to think about data, #opensource materials to learn new methods, preferences when it comes to the types of theoretical / modeling frameworks I use (explanation over prediction, etc).

Working with a new researcher helped me put that down in words - we had an ongoing shared doc where I'd link various resources. I plan to use this with future students!

My summer mentoring experience is coming to an end, and I wanted to reflect on my learnings! It was the first time I'd worked with an...

My summer mentoring experience is coming to an end, and I wanted to reflect on my learnings!

It was the first time I'd worked with an undergrad research intern, and overall I felt lucky to work with such a smart and motivated student.

1) Non-linearity of training time: my initial expectation was that I'd do training, and then I'd be "done". But I learned that it ebbs and flows - that every time I introduced something new (ie., modeling, #scicomm) I would need to invest time.

#AcademicChatter

I recently started applying for grants for my research at my institution.One thing that has really helped is asking other students who...

I recently started applying for grants for my research at my institution.

One thing that has really helped is asking other students who successfully applied for similar grants for materials.

Before I start writing, I've been *reverse-outlining their materials* so I can focus on the underlying structure. It also gets my brain going re: thinking about where my own citations go.

Super useful for getting comfortable with this new type of writing!

@AcademicChatter @Sociology #AcademicWriting

Finally, in the early matching process, I allowed for "re-entry points". In case a student had ignored an ask, I included them on...

Finally, in the early matching process, I allowed for "re-entry points".

In case a student had ignored an ask, I included them on the next email in case they'd missed something in the middle of their semester. If they ignored that email, then I didn't bother them again.

This allowed students who had other things going on to jump back in. Sometimes students have life things pop up, and I wanted to give them a chance to move past those things without having to explain themselves.

I shared these things across multiple avenues for getting to know me: whole-group meetings, 1-on-1s, and emails. And I also allowed mentees...

I shared these things across multiple avenues for getting to know me: whole-group meetings, 1-on-1s, and emails. And I also allowed mentees to engage w me across multiple formats.

While a single email/Zoom meeting only gives you one data point, you start to get a better sense of folks (and more data points!) when you meet in person and receive emails back - or not.

Informed by qual methods, I know that depth of your person-data matters! More *exposure*/contact is good.

https://www.ucpress.edu/blog/60122/the-secret-to-assessing-qualitative-research/

Another idea from the match process: I think it's important to provide ample info upfront so not only do you select them, they select...

Another idea from the match process:

I think it's important to provide ample info upfront so not only do you select them, they select you!

A few things I highlighted in meetings: 1) my research values - (some) self-sufficiency, consistency, and ability to learn 2) my research story, and how my work is different from other research 3) my personality - I tell them ideas that inspire me and things that make me laugh.

More data on the mentor = better selection mechanisms on the mentee side!

https://fediscience.org/@angela/111653362125373472Student tip: When I have academic books I want to read, my default is to ask my...

https://fediscience.org/@angela/111653362125373472

Student tip: When I have academic books I want to read, my default is to ask my institution to buy it first.

This is heavily influenced by when I worked at the university libraries as an undergrad. I had no idea how much they did for researchers until I was the one fulfilling all sorts of requests (scanning pages, finding sources, wheeling book carts).

It left me with a deep appreciation for what university #libraries do + #libraryWorkers!

@AcademicChatter @sociology

Also, I answered the written qs later in writing. Even if I didn't get to all qs in the meeting, I was able to address student questions...

Also, I answered the written qs later in writing. Even if I didn't get to all qs in the meeting, I was able to address student questions and concerns (about doing research with me). It helped me also in terms of spelling out my expectations for students, as well as my own research values.

Re: written engagement: on a recent Zoom, I provided a separate notes Gdoc along with the meeting and allowed students to log questions...

Re: written engagement: on a recent Zoom, I provided a separate notes Gdoc along with the meeting and allowed students to log questions there as well as ask them by unmuting. Some students are more comfortable asking qs verbally in front of others, but others prefer writing their qs.

When I looked at the Gdoc version history later, I realized *every student* had asked a question, even quieter 1st year student attendees! Multiple modes to interact w/me allowed for more engagement and inclusion.

Recently started participating in a undergrad research mentorship match program. One reflection I've had on it that we rarely train...

Recently started participating in a undergrad research mentorship match program. One reflection I've had on it that we rarely train academics *how* to select and work with undergrads effectively. And additionally, we may not always frame selection processes in an *equity-minded way* (ie. selection can reinforce inequality)

So far, trying to give flexibility for engagement: multiple Zooms, written engagement, clear asks.

Will try to log learnings from this process!

@sociology @AcademicChatter

Second tigeRs meetup done! Thanks to @zbleemer for presenting methods to digitize historical ed records w #rstats. 38 folks from Soc, Econ,...

Second tigeRs meetup done! Thanks to @zbleemer for presenting methods to digitize historical ed records w #rstats.

38 folks from Soc, Econ, Pub Pol, Demog, Psych, Ed, Libraries, History and even Ecology showed up to the talk. Glad to see that we’re really an interdisciplinary group using R on campus.

📌 Zach’s code can be found here: zacharybleemer.com/software

The event was sponsored by Princeton’s Data-Driven Social Science initiative + coorganized by me and DDSS fellow Kim Kreiss!

We've put up the slides + code from this year's Princeton #Sociology Methods Camp: https://pusocmethodscamp.org/This set our...

We've put up the slides + code from this year's Princeton #Sociology Methods Camp: https://pusocmethodscamp.org/

This set our students up to use #rstats in our intro grad stats sequence.

I'm most proud about:
👏 Teaching students to use #chatGPT as a research tool
👏 Incorporating #Quarto and the visual editor for new R users
👏 Teaching w live coding (and my exercises to explicitly teach debugging!)
👏 Generating fake data for releasing materials openly

Please use & share these materials openly!

Finally, one thing that a lot of these convos overlook is that you need to know HOW to ask the #genAI tool in order to get useful output...

Finally, one thing that a lot of these convos overlook is that you need to know HOW to ask the #genAI tool in order to get useful output ("prompt engineering"). But framing the right q to get what you want is exactly where the work happens in research!

➡️ Teaching students to phrase their question to #ChatGPT to get what they want is an important part of problem solving!

➡️ Getting students to where *they knew enough to ask useful questions to #genAI to solve problems* >> #rstats code syntax

Theorizing publicly, my hunches about using #ChatGPT + #genAI in grad #rstats education are that:1) you need to have a baseline comfort with...

Theorizing publicly, my hunches about using #ChatGPT + #genAI in grad #rstats education are that:

1) you need to have a baseline comfort with a specific programming language - and computing in general - to feel comfortable using an assistant like #genAI

2) #Sociology grad students are aware that the tool is often wrong, and come in with prior skepticism of #genAI

3) instructors will prepare their students (and help some a lot!) by introducing the tool - it's worth teaching in structured ways!

On the unsure side:-"I still feel like I'm not fully experienced enough to know when chatgpt is giving me a wrong answer...chatgpt...

On the unsure side:

-"I still feel like I'm not fully experienced enough to know when chatgpt is giving me a wrong answer...chatgpt still intimidates me a little bit, but the introduction to it as a tool was helpful."
-"I still don't fully understand how to use it, but I believe it benefitted others."
-"I think it was interesting but I am not as convinced about the utility it provides."

Folks with *more* prior programming experience found #ChatGPT *most* helpful, novices weren't as sure.