Principals are increasingly expected to be the instructional as well as administrative leaders of their schools. Moreover, their instructional leadership matters for student success (Hallinger & Heck, 1996). However, little is known about how principals reason through the instructional issues that they face. An analysis of principal reasoning in instructional contexts is critical. Drawing on interviews with 36 urban school principals, I examine principal problem solving from three different perspectives. These perspectives are divided into three essays.

In the first essay, I compare the problem solving processes that principals of varying degrees of expertise apply in solving instructional scenarios. This quantitative approach tests the hypothesis that expert principals use more expert processes and typical principals use more typical processes. The results confirm that expert principals relied more on gathering data across all instructional scenarios and spent more time following up on their solutions when solving problems in the mathematics domain. Typical principals were more likely to share anecdotes that were negative or unsuccessful in nature across all domains.

In the second essay, I apply a word-use methodology to look at frequently used words which have been linked to certain emotional, cognitive or social categories (Pennebaker, Chung, Ireland, Gonzales, & Booth, 2007). Specifically, I investigate differences in the words principals use when solving problems across mathematics, literacy and general scenarios. My analysis shows that overall, principals reasoned through literacy problems from a more collaborative perspective than mathematics scenarios. Principals used more words relating to certainty (e.g. always, never) when working on literacy scenarios, and more words relating to money in mathematics scenarios. In a separate series of tests, the responses of expert principals were compared across subject matter to typical principals.

In the third essay, I compare methodologies for transforming the principals' transcribed responses into maps. I analyze these maps with the aim of identifying salient components in the principal's reasoning. The maps show that some components coalesce into what I term "levers" which give the principal leverage to solve the problem at hand. The second part of the paper presents two levers which arise from the data.