Alvin Zhou

Standing room only at @zeningduan.bsky.social and @alvinyxz.bsky.social 's @icacm.bsky.social panel on LLMs #ica25

Amidst all the posts about how NSF funds the basic science that (eventually) leads to marketable products, its role in funding data infrastructure is getting lost.

The GSS (1972), ANES (late 1960s), and PSID (1968) are rounding error in the discretionary budget but vital national resources.

1/6

[Extended Abstract Deadline: January 15, 2025]

We’ve pushed the deadline back two weeks to give everyone a little extra breathing room during the holidays. Take your time, and we look forward to your submissions! #GenAI #AI #Computational

Excited to share my new #JEPS paper with @xiaoxiaoshen.bsky.social introducing ReChat, a research tool for live-text interaction studies! Makes it easy to embed live chats in surveys, opening new possibilities for research on social interaction. Check it out! 🧵👇
#AcademicSky #PoliSciSky #PoliSky

🚨Abstract for the special issue on GenAI at Computational Communication Research @computationalcommunication.org due Dec 31!

🚨We (@alvinyxz.bsky.social, @ewam.bsky.social) are editing a special issue for Computational Communication Research on GenAI! Submissions on GenAI as comm phenomena or research tools are welcome: z.umn.edu/ccrgenai

Abstracts due: Dec 31 '24
Full papers: Apr 30 '25

@computationalcommunication.org

We @hubbardschoolumn.bsky.social are hiring a postdoc in journalism to work with Benjamin Toff and the Minnesota Journalism Center on projects studying the effectiveness of newsroom strategies around building trust and engaging with news avoiders. April 22 is the priority deadline t.co/OLdwKEfF9f

This finding has implications for how we think about misinformation interventions.

If those who are exposed to misinformation "already know better", maybe our solution should look past psychological processes or individual skills.

Our panel tracked N ≈ 140,000 individuals in the United States for 12 months, so we can also do intra-person analyses (fixed-effects).

The puzzle holds: if my news diet becomes more ideologically diverse next month, I am more likely to get exposed to unreliable content.

In fact, the ideological diversity of news diet predicts misinformation exposure as strongly as "age 55+" and partisan leaning (@andyguess.com)

We quantify ideological diversity using "ideological distance" which measures the average pairwise distances between news pages.

Instead of focusing on the prevalence/reach of misinformation overall, we think it is important to know how that exposure is distributed among news consumers, and what kinds of news those misinformation-prone people also consume.

Gist of our findings:

With Tian Yang and @sgonzalezbailon.bsky.social, our new paper at NMS shows that ideologically diverse news diets are *positively* associated with misinformation exposure.