iiinova — where your digital self keeps evolving

Reposted by Ana Macanovic

The Institute for Analytical Sociology @iasliu.bsky.social · 24/11/2025

Join us on Thursday, 27 November, at 14:30 CET for the International Roundtable on Computational Social Science with @anamacanovic.bsky.social 🔹 Limits to Gender Equity in Science Communication 🔹 More info: liu.se/en/article/s...

liu.se

Seminars and lectures at IAS

Welcome to IAS public lectures and seminars.

031 0

Reposted by Ana Macanovic

ODISSEI @odissei.bsky.social · 03/11/2025

The world’s first Macroscope is coming to the Netherlands 🇳🇱 NWO just awarded €16.8M to build this groundbreaking research infrastructure—co-led by @odissei-soda.bsky.social and CLARIAH, coordinated by @erasmusuniversity.bsky.social #Research #Innovation odissei-data.nl/2025/11/03/t...

1134 0

Reposted by Ana Macanovic

erin pineda @erinrpineda.bsky.social · 25/10/2025

3) For some reason (likely related to ridiculous under-staffing and underpayment of library workers) it seems next to impossible to get specific editions or translations put on reserve, rather than any old copy that is available

41916 0

Reposted by Ana Macanovic

Per Engzell @pengzell.bsky.social · 09/10/2025

Academia rewards resilience, by which I mean the ability to rewrite the same sentence forever

4805 0

Reposted by Ana Macanovic

Mark Wittek @markwittek.bsky.social · 10/05/2025

Thank you for participating in the Measuring Culture workshop funded by the Academy of Sciences in Heidelberg @hadw-bw.bsky.social. It was an incredible experience for me due to all the brilliant people attending. Special thanks to keynote speakers @nehagondal.bsky.social and @olizardo.bsky.social

0224 0

Reposted by Ana Macanovic

Mark Wittek @markwittek.bsky.social · 30/04/2025

I am very grateful for two days of stimulating intellectual exchange at our workshop funded by the Cologne International Forum. Special thanks to our keynotes @isabellevdv.bsky.social, @christian.czymara.com, and the organisation by @anamacanovic.bsky.social, @maxpi.bsky.social, and Christof Nägel.

0145 0

Reposted by Ana Macanovic

Milena Tsvetkova @mtsvetkova.bsky.social · 02/04/2025

I have two postdoc positions for a computer scientist and a social scientist to work on my ERC-funded project HUMANET Human-Machine Social Systems at LSE. Please read more and apply at jobs.lse.ac.uk/Vacancies/W/.... More information about the project is available at humanet.science.

jobs.lse.ac.uk

Research Officer in Computational Social Science (2 Posts)

Research Officer in Computational Social Science (2 Posts), , LSE is committed to building a diverse, equitable and truly inclusive university ...

0147 0

Ana Macanovic @anamacanovic.bsky.social · 17/02/2025

Only two more weeks left to submit an abstract for our IPODA workshop on combining insights from survey methods and online data sources for research on immigration, public opinion, and crime. 📚 Make sure to check out the call below if you work in these domains!

031 0

Ana Macanovic @anamacanovic.bsky.social · 15/01/2025

If you are an early career researcher, please get in touch with us as we can subsidize your travels. We are hosting this workshop with Christof Nägel, @maxpi.bsky.social, and @markwittek.bsky.social with the support from the Cologne International Forum.

011 0

Ana Macanovic @anamacanovic.bsky.social · 15/01/2025

The workshop strives to bring together scholars using online and offline data sources to study topics such as political discourse, hate speech, voting, political radicalization, or crime. It will take place in-person at the University of Cologne. We will provide coffee, lunch and dinner on the 28th.

111 0

Ana Macanovic @anamacanovic.bsky.social · 15/01/2025

If you use online or survey data to study immigration, public opinion, or crime, consider sending an abstract for our workshop on April 28/29 2025 at the University of Cologne. We will have great keynotes by @christian.czymara.com and @isabellevdv.bsky.social! See the call below:

Immigration, Public Opinion, and Crime in the Digital Age - Leveraging online data sources with insights from traditional survey methods
How can social scientists leverage the potential of online data sources in the light of lessons learned from survey research and qualitative interviews to inform societal and scientific debates on immigration and crime? In recent years, scholars from various disciplines have shown a growing interest in using online data sources—such as social media platforms and discussion forums—to study public opinion and political discourse concerning immigration and crime. The rise of right-wing populist political actors and the increasing threats of political polarization around the globe sparked a novel literature that applies computational methods to analyze public opinion in the digital realm. In parallel, social scientists have been studying political attitudes, voting behavior, radicalization, social movements, polarization, and crime with quantitative and qualitative interviews as well as administrative data for nearly a century. Yet, these strands of literature mainly operate independently and follow different methodological approaches and theoretical frameworks. This workshop aims to bring together scholars from various research communities using online and offline data sources to inform shared research questions concerned with immigration, public opinion, and crime. Presentations may cover topics such as the influence of media coverage of events triggering negative outgroup sentiments (e.g., terror attacks or crime) on political discourse, individual hate speech, crime, and voting behavior, political radicalization on online platforms and in offline spaces, or the influence of political actors on public opinion. We encourage scholars who combine online and offline data or would like to learn more about research paradigms outside of their own research area to send an abstract (max. 750 words) by the 1st of February 2025 to project.publ…

1107 0

Reposted by Ana Macanovic

Workshop on Online Abuse and Harms @woahworkshop.bsky.social · 18/12/2024

🚨 Exciting news! The call for papers is now OPEN for the 9th Workshop on Online Abuse and Harms (#WOAH2025) at #ACL2025 @aclmeeting.bsky.social in Vienna. 📆 Submission deadline: March 7, 2025 📄 Check out the CfP: www.workshopononlineabuse.com/cfp.html Looking forward to your submissions! 🤩 #NLProc

workshopononlineabuse.com

Call For Papers

The 9th Workshop on Online Abuse and Harms (WOAH) at ACL 2025.

0148 0

Reposted by Ana Macanovic

ODISSEI @odissei.bsky.social · 10/12/2024

Now let’s peek into the session Science of Science👀 @anamacanovic.bsky.social discovered gender differences in media representation of scientific work of Dutch professors. She shows that women professors are underrepresented in printed media, but not in online media outlets. #ODISSEI2024

053 0

Ana Macanovic @anamacanovic.bsky.social · 15/11/2024

Could you please add me to one of the sociology packs as well? Thank you!

000 0

Reposted by Ana Macanovic

Marc Keuschnigg @marckeuschnigg.bsky.social · 25/10/2024

We are hosting the 11th International Conference on Computational Social Science in Sweden 🚀The IC2S2'25 website is LIVE, and submissions are OPEN! 📍Norrköping | July 21-24, 2025 Call for Abstracts (until Feb 24) Call for Tutorials (until Jan 17) 🔗Explore details & submit: ic2s2-2025.org

07447 0

Reposted by Ana Macanovic

Rense Corten 🟥 @rensec.bsky.social · 23/10/2024

A first starter pack on computational social science, that is, researchers who use computational methods (broadly defined) to answer social science questions. It's a vaguely defined field with blurry borders, so it's certainly incomplete; if you want to be on it, let me know! go.bsky.app/PCckf3C

173919 0

Reposted by Ana Macanovic

Ozan Aksoy @aksoyundan.bsky.social · 19/10/2023

As part of NCRM, I am running a 3-day intensive course on Structural Equation Modelling and Causal Inference which takes place online on 8-10 Nov 2023. Bursaries are available for fees. Details: www.ncrm.ac.uk/training/sho... May be relevant for those in sociology, psychology, social sciences

ncrm.ac.uk

01313 0

Reposted by Ana Macanovic

Chris Bail @chrisbail.bsky.social · 08/10/2023

Repost if you’ve participated in a Summer Institute in Computational Social Science. Let’s get #SICSS Bluesky going!

05163 0

Reposted by Ana Macanovic

Kathleen Weldon @kathleenweldon.bsky.social · 06/10/2023

The most frustrating thing about working for the oldest social science data archive in the world is knowing how much cool stuff we have that goes unused because people don't know about it. In some cases, making data usable requires more resources than we have (THREAD).

16339157 0

Ana Macanovic @anamacanovic.bsky.social · 05/10/2023

Thank you! It can be so hard to navigate the recent literature, so we hope this will help us all think together about the potential of these models for annotation.

010 0

Ana Macanovic @anamacanovic.bsky.social · 05/10/2023

Thank you for sharing! We will soon update our overview to include your newest results as well.

010 0

Reposted by Ana Macanovic

Ted Underwood @tedunderwood.com · 04/10/2023

Quote posting this with the preprint link to make sure people can find it: osf.io/preprints/so...

0103 0

Ana Macanovic @anamacanovic.bsky.social · 04/10/2023

Find our preprint here and let us know what you think! This is very much an ongoing discussion. osf.io/preprints/so...

osf.io

162 0

Ana Macanovic @anamacanovic.bsky.social · 04/10/2023

We are enthusiastic about what AI can bring to the humanities and to the social sciences. But in order to harness its power, we need to avoid falling for the hype.

000 0

Ana Macanovic @anamacanovic.bsky.social · 04/10/2023

At the moment, the models perform better on English-language, and even better if they are fed English prompts about another language.

021 0

Ana Macanovic @anamacanovic.bsky.social · 04/10/2023

Not all data can be fed to an API. Sensitive, personal, copyrighted information cannot leave your computer, yet they do constitute a significant portion of what we work on as social scientists.

001 0

Ana Macanovic @anamacanovic.bsky.social · 04/10/2023

Not all models are open-source or archived, and as they get updated their results change. This is a major problem for replication, pointed out in this Nature paper: www.nature.com/articles/d41...

nature.com

000 0

Ana Macanovic @anamacanovic.bsky.social · 04/10/2023

Another result is that the validation criteria vary between studies. This is sometimes warranted, but can be limiting too. To improve science, we should probably harmonize our standards, as evoked in a recent paper arxiv.org/abs/2308.07832.

arxiv.org

020 0

Ana Macanovic @anamacanovic.bsky.social · 04/10/2023

There are many other results from this, which we summarized in a large table. One is that the recall seems often higher than the precision. This means, again, that if these models could be used to extract relevant information from massive corpora, they should then be validated by humans.

161 0

Ana Macanovic @anamacanovic.bsky.social · 04/10/2023

The performance of these few-shot learners vary widely, and is sometimes very low. This means that we just cannot trust them blindly, and that when using them, we need to make sure to validate the results thoroughly.

010 0

Ana Macanovic @anamacanovic.bsky.social · 04/10/2023

Yet they are often not the best tool available. On average, transformers (like BERT models) fine-tuned with human annotations, yield better results. This is no surprise, but it should guide our choice of tools: do we want a precise classifier, or a quick one to test some ideas?

060 0

Ana Macanovic @anamacanovic.bsky.social · 04/10/2023

Few-shot classifiers can annotate text with success. On a series of tasks, they can save a huge amount of time and spare researchers the repetitive aspect of annotating thousands of tasks.

010 0

Ana Macanovic @anamacanovic.bsky.social · 04/10/2023

Would you like to use GPT-4 and other few-shot learning algorithms to annotate massive corpora? So do we! But before we did so, we reviewed the evidence. Here are the main takeaways from our systematic review:

104218 0