are you attending @colmweb.org 🦙 psst, don't miss our First Workshop on Bridging NLP and Public Opinion Research tomorrow #COLM2025 🛎️ we have a great line up of keynotes, presentation, and panel discussion, followed by a happy hour 📍 room 518B schedule and info: sites.google.com/view/nlpor20...
Stephanie Eckman
@stephnie.bsky.social
I collect high quality data for social science and model training statistics, data science, surveys, cats, biking, urbanism www.stepheckman.com
Looking for tips on how to use the new AI tools to help with a literature review. I was so excited about NotebookLM, but my frustration is growing. I uploaded the papers (~40) I want it to draw from, and then spent several hours trying to get it to draft a reasonable lit review *with citations*
@statstas.bsky.social — your point about totals rather than percents
The actual number of votes that Trump got is less than he lost with in 2020. This was an absolute failure of Democrats to turn out the vote. I don't think any of these places actually became more Republican
Can anyone point me to a paper, WP, blog post that reviews the literature on synthetic samples? Thank you
Has anyone used DC’s school lottery as an instrument to evaluate effect of pre-k on subsequent educational achievement ?
Took us a while, but we finally put another chapter of our book "Data Collection with #Wearables, #Apps, and #Sensors" online. Chapter 4: Errors of Nonobservation: #Sampling and #Coverage. bookdown.org/wasbook_feed... Let us know what you think! @stephnie.bsky.social
Chapter 4 Errors of nonobservation: Sampling and coverage | Data Collection with Wearables, Apps, and Sensors
Chapter 4 Errors of nonobservation: Sampling and coverage | Data Collection with Wearables, Apps, and Sensors
bookdown.org
In case you missed it! "Interviewer Involvement in Respondent Selection Moderates the Relationship between Response Rates and Sample Bias in Cross-National Survey Projects in Europe" by Marta Kołczyńska, Piotr Jabkowski & Stephanie Eckman @stephnie.bsky.social doi.org/10.1093/jssa...
@aapor.bsky.social @floriankeusch.bsky.social @daob.nl
Survey researchers of Blue Sky, this is a new initiative of mine, and right now the queue is wide open meaning any new submissions have a high likelihood of acceptance: docs.google.com/document/d/1... #polisci #polisky
Fantastic thread on data quality issues with inattentive respondents.We can’t weight our way out of these problems if the demographics are provide are wrong
To your points, yes, this is GESIS. The question was "recent research" and I certainly think "forthcoming at JSSAM" is responsive to that question. But it's not the only study to find that inattentive respondents differ in politically-relevant demographics
Searching for the dates of the next AAPOR conference, I typed in "aapor.org conference." None of the top hits were helpful --previous years, archived site I think we could use a little Search Engine Optimization, @aapor.bsky.social In the meantime -- include 2024 in your search (or "79th" 🙄🤪)
New Article Alert! "Interviewer Involvement in Respondent Selection Moderates the Relationship between Response Rates and Sample Bias in Cross-National Survey Projects in Europe" by Marta Kołczyńska, Piotr Jabkowski & Stephanie Eckman @stephnie.bsky.social doi.org/10.1093/jssa...
Please share concrete examples of how you’re using LLMs in your jobs Draft email to boss reminding them they owe you something? Translate code from R to python? I’m sure you all have ideas I haven’t thought of
If we think of political survey participation as itself a form of political behavior, it follows that a way to increase and equalize this participation is to lower the costs of participation. And yet, flagship poli sci surveys take 90 minutes or more to complete
With polling in the headlines, Tori Gorton & I have revised our paper, "On the Internet, No One Knows You’re an Activist: Patterns of Participation and Response in an Online, Opt-in Survey Panel." tl;dr Survey participation is itself a political behavior. papers.ssrn.com/sol3/papers....
I'm interested in software to manage a survey panel. Can those of you who run such panels let me know what you use What are the crucial features? @carinacornesse.bsky.social @ashleykirzinger.bsky.social @jonburton.bsky.social @kwcollins.bsky.social @usociety.bsky.social @aapor.bsky.social
New paper: What social science can teach AI and ML researchers about how to collect training data arxiv.org/abs/2311.14212 forthcoming in EMNLP 2023 FIndings (My first paper thread on Bluesky, we'll see how this goes)
I'm still new to BlueSky and finding my way. So far I'm seeing way more porn than on the other site. All the feeds, regardless of topic, are porn filled -- unfortunate