Text this: Comparing Zero-Shot Large Language Model Prompting with Human Coding of Theory Concepts in Student Essays