🔊 New paper just accepted in JPSP 🥳 In "Inference from social evaluation", we explore how people use social evaluations, such as judgments of blame or praise, to figure out what happened. 📜 osf.io/preprints/ps... 📎 github.com/cicl-stanfor... 1/6 - ThreadSky

tobigerstenberg.bsky.social • 35 days ago

🔊 New paper just accepted in JPSP 🥳

In "Inference from social evaluation", we explore how people use social evaluations, such as judgments of blame or praise, to figure out what happened.

📜 https://osf.io/preprints/psyarxiv/h9c5w

📎 https://github.com/cicl-stanford/inference_from_social_evaluation

1/6

Comments

drlendall.bsky.social•34 days ago

Fascinating! Hoping I can gain access to this article and share it with students.

tobigerstenberg.bsky.social•34 days ago

Thank you! You can access the preprint of the paper here: https://osf.io/preprints/psyarxiv/h9c5w

tobigerstenberg.bsky.social•35 days ago

Several factors influence how people assign responsibility. For example, when a person put in a lot of effort, we praise them more for success and blame them less for failure. Because our intuitions for who is deserving of blame or praise are shared, these judgments provide rich information.

2/6

tobigerstenberg.bsky.social•35 days ago

In Experiment 1, we demonstrate that people can infer ability, effort, knowledge, intention, and social roles from blame judgments in short scenarios. For example, when George was blamed more than Amanda, and the question is who put in less effort, it must have been George.

3/6

tobigerstenberg.bsky.social•35 days ago

In Experiment 2, we develop a computational model that maps from effort onto praise. Going out of one's way to get a co-worker a coffee is deserving of more praise. This means that learning how much praise someone received allows us to make inferences about what must have happened.

4/6

tobigerstenberg.bsky.social•35 days ago

In Experiment 3, we look at situations where a group fails to optimally coordinate their actions. We develop a model of blame that considers what a person should have done, and how much it mattered that they didn't do it. Knowing how much each agent was blamed reveals info about what happened.

5/6

tobigerstenberg.bsky.social•35 days ago

Social evaluations are commonplace in everyday conversations. While they don't fully reveal what happened, they provide valuable clues. We combine generative models with Bayesian inference to capture this.

Joint work with Zach Davis, Kelsey Allen, @maxkw.bsky.social and @julianje.bsky.social 🙏

🔚

Comments

Posting Rules

Reply