Educational Intervention

Choose how far along the chain to measure

A widely used way to think about training outcomes is the Kirkpatrick model, which describes four levels. Each level is harder to measure and more meaningful than the one before it.

Level	Question it answers	Typical measures	Time needed to see change
Reaction	Did participants find it useful?	Satisfaction or feedback survey	Immediately
Learning	Did they gain knowledge, skill, or confidence?	Knowledge test, skills check, confidence scale	Immediately to weeks
Behavior	Do they practice differently?	Observation, chart audit, documentation rates	Weeks to months
Results	Did patient or organizational outcomes improve?	Clinical outcomes, rates, costs	Months or longer

Most educational projects can measure learning. Fewer can measure behavior, and results are often out of reach in a DNP timeline. That’s fine if your claims match the level you measured. Read the guide

Measurement issues specific to education projects

Confidence isn’t competence. Self-rated confidence often rises after training even when skills don’t. If you measure confidence, say so, and don’t describe it as skill.

Tests need room to improve. If most people score near the top before the session, the test can’t show improvement. Check your pre-test scores early.

Self-made tests need evidence. A test you wrote should be checked by content experts against what you taught, and piloted before use. Read the guide

Timing matters. A test right after the session measures short-term recall. A delayed follow-up shows whether learning lasted.

Linking matters. Paired analyses need each person’s pre and post responses linked. If responses must be anonymous, a self-generated ID code can keep them linkable. Read the guide

Before you start

• You’ve decided which outcome levels you’ll measure, and your claims will match them.
• Your instrument covers what you’ll teach and has room to show improvement.
• You’ll use the same instrument, the same way, before and after.
• You can link each person’s pre and post responses, or you’ve planned your analysis for unlinked data.
• You’ve checked that your program accepts education-focused projects with the outcomes you plan. Some programs don’t accept projects whose only outcome is participants’ knowledge.