The exploration phase uses an inductive epistemic orientation for understanding the problem, meaning that we start with our own a priori assumptions (i.e., current beliefs and schema), and begin our process of understanding the problem and the contextual variables that surround it. This is typically done through: observation of the problem, reviews of academic literature, and discussions with those closest to the problem itself. This is an immersive and distinctively qualitative phase. As our schema deepens, causal drivers of problems will begin to emerge. Identification of these drivers allows us to separate symptoms from causes.  

Once an initial sense of the problem is developed, we quickly shift into a deductive phase, meaning that we be begin hypothesizing and testing solutions (i.e., interventions) to the identified drivers. We are not attempting to address the entire problem, just one or two causal drivers at a time. In this way, we can begin gathering empirical data to see which interventions work and which fall short. This is typically a quantitative phase. The timely feedback gained from short intervention testing cycles allows us to gain initial traction on the problem, incrementally working to better analyze, understand, and ultimately address it. Understanding deepens with engagement. Working within the medium of the problem leads to a more robust understanding of it.

We've identified a problem and our understanding of it is deepening. Members of our team have reviewed and synthesized relevant academic literature; others have interviewed the people who are most familiar with the problem, and have summarized their findings; and others have conducted observations, and written-up their results. All this information has been shared, a handful of candidate causal drivers have been identified, and we've updated our priors. Now it's time to create and test some solutions using our updated schema. This can be done using a variation of the PDSA sequence tailored for short-cycle improvement studies (Figure 2).