AI-Powered Analysis

Tutorial

AI Coding vs. Human Coding: Running Your Own Inter-Rater Reliability Check (2026)

Inter-rater reliability - checking how well two independent coders agree - has been standard practice in qualitative research for decades. Extending that same check to an AI coder paired against a human one is more accessible than it sounds, and it's the single most credible way to answer \"is this good enough to trust\" for your own specific data, rather than relying on someone else's published study. This guide walks through how to actually run one.

Tutorial

When to Trust AI Categorization (and When to Double-Check It) (2026)

AI classification of open-ended responses is reliable most of the time, and \"most of the time\" is exactly the phrase that gets people in trouble if they don't know which responses fall outside it. This guide covers what confidence scores actually tell you, the specific kinds of responses AI classifiers most often get wrong, and a practical spot-checking habit that catches problems without re-reading everything by hand.

Tutorial

Writing a Good Classification Goal: Prompting AI to Find the Right Themes (2026)

Tell an AI classifier to \"find the themes\" in a pile of open-ended responses and you'll get themes back - just not necessarily the ones you actually needed. The gap between a vague classification goal and a useful one is almost entirely about how specifically you describe what you're looking for. This guide covers how to write a classification goal that gets you categories worth reporting on the first pass, with examples of vague prompts next to the specific rewrites that fixed them.