We find that ChatGPT is not perfect but accurate in identifying pre-defined arguments in large text corpora. This is a simple and fast way to *deductively* extract opinions and attitudes in detail from unstructured text datasets.
We find that ChatGPT is not perfect but accurate in identifying pre-defined arguments in large text corpora. This is a simple and fast way to *deductively* extract opinions and attitudes in detail from unstructured text datasets.