When is an ‚area‘ of valid truth, rather than a single point, the more appropriate assumption for annotating texts? In our new preprint, led by @ahrabhikat.bsky.social with @janabernhard.bsky.social and Hajo Boomgaarden, we study coder bias when annotating constructs of marginalization.
Reposted from
Ahrabhi Kathirgamalingam
Preprint alert! 🚨Happy to share our work on a key challenge in content analysis: coder bias – the systematic variation in annotations driven by individual characteristics when coding with human or LLMs. Let’s break it down… osf.io/preprints/so... #Bias #Research #Methods #AI
Comments