Action Research

Action research and quality improvement (e.g., Plan-Do-Study-Act cycles) are closely related and share cyclical approaches. QI is more narrowly focused on improving specific metrics or processes; AR is broader, aiming to understand a problem and generate knowledge alongside practical improvement. AR emphasizes collaboration and democratization more explicitly; QI may be more top-down. Both involve cycles of action and learning. Some consider QI a form of action research; others view them as related but distinct. Distinctions matter less than ensuring your project is rigorous: has clear aims, involves relevant stakeholders, collects data systematically, reflects regularly, and documents learning.

There is no fixed number. Typically, 3–5 cycles over 6–18 months allow sufficient learning and refinement. Early cycles test initial ideas and generate unexpected insights; later cycles refine based on learning. The process is complete when the solution is adequately refined, outcomes are stable or improving, and insights are sufficient to answer research questions or guide ongoing practice. Document when and why cycles end.

True collaboration means: (1) Shared decision-making—practitioners and researchers jointly set research questions and decisions about action and change; (2) Equal voice—mechanisms for practitioners' perspectives to influence research direction; (3) Capacity building—researchers support practitioners' research skills; (4) Transparent process—clear communication about research aims, methods, findings; (5) Benefit sharing—findings and learning belong to all parties, not just researchers. Build in regular reflection on collaboration itself: Is it genuine? Who is driving decisions? Are power imbalances present? Adjust as needed.

Collect data addressing your research questions and improvement goals. Typically this includes: (1) Process data—documentation of what was implemented, how, when (field notes, logs); (2) Outcome data—metrics related to your improvement goals (e.g., communication quality, patient satisfaction, error rates); (3) Stakeholder perspectives—interviews or surveys with those involved about what worked, barriers, unintended effects; (4) Implementation fidelity—was the intervention implemented as planned? Be intentional about measurement; use mixed methods (quantitative and qualitative) when possible.

Ideally, action research involves multiple stakeholders collaborating. A single researcher can conduct action research if they are investigating their own practice (e.g., a teacher testing new instructional methods in their own classroom) and are in genuine dialogue with colleagues or peers for reflection and input. However, true collaboration—with practitioners outside the researcher's immediate role—is preferable because it broadens perspective and strengthens implementation of findings. If alone, build in external advisory or peer debrief structures.