Later: the effect diminishes, returning to near-zero bias levels.

The clearest evidence appears with a continuous tone 𝐨𝐮𝐭-𝐠𝐫𝐨𝐮𝐩 𝐟𝐚𝐯𝐨𝐫𝐢𝐭𝐢𝐬𝐦 difference. The findings align with out-group favoritism rather than in-group discrimination.

𝐖𝐡𝐚𝐭 𝐰𝐞 𝐟𝐢𝐧𝐝

Baseline: As expected, 𝐧𝐨 𝐬𝐲𝐬𝐭𝐞𝐦𝐚𝐭𝐢𝐜 racial bias exists 𝐛𝐞𝐟𝐨𝐫𝐞 heightened awareness.

After attention: 𝐨𝐯𝐞𝐫𝐜𝐨𝐫𝐫𝐞𝐜𝐭𝐢𝐨𝐧, where players receive 𝐟𝐞𝐰𝐞𝐫 fouls from crews of more “opposite-race/dissimilar-tone” officials.

We examine the #WNBA, which is known for its clear stance against all forms of discrimination.

Using 11 seasons of data, quasi-random referee assignments, and Machine Learning to predict race and skin tone, we explore how increased public attention influences officiating.

📢 New working paper!
“𝐑𝐚𝐜𝐢𝐚𝐥 𝐁𝐢𝐚𝐬, 𝐂𝐨𝐥𝐨𝐫𝐢𝐬𝐦, 𝐚𝐧𝐝 𝐎𝐯𝐞𝐫𝐜𝐨𝐫𝐫𝐞𝐜𝐭𝐢𝐨𝐧” (with
Alex Krumer, Rosa Lavelle-Hill, and Tim Pawlowski)

🧵 below 👇

Inspired by @dynarski.bsky.social bakes...revived the starter for my first bake in Germany. 71% hydration, 100% roggen levain, rest dinkel. Over night cold proof. crumb will get better! 🫧 🍞

Hey @deutsche-bank.bsky.social you would probably have a fair case against Grok and Elon Musk for copyright infringement

Mit vielen Grüßen aus dem Magdeburger Dom. Friede sei mit euch 🎄 🕊️ 🙏