Little information is available about how formal, social science theory has been used in outcome evaluations. This gap exists in debates about what role theory has in evaluation. To help address this, we need to understand variation in formal theory use in real-world practice. This research applied systematic qualitative coding to identify and classify patterns of theory use, followed by qualitative comparative analysis. The terms describing theory are defined and formal theory is differentiated from other kinds of theory, such as program theory. The study found instances of borrowing and repurposing theoretical material in a cohesive sample of 17 outcome evaluation reports covering programs addressing complex social problems, drawn from cross-cultural contexts. Theory was mostly used for post hoc explanation and less often used upfront in framing, design and conduct of evaluation. The concrete approach in the literature of applying formal theory to measure and explain causal pathways to behavioural outcomes was found, as was layered use of a range of theoretical material. This article offers insights that may assist evaluators to undertaken more sophisticated and reflective approaches to using theory. Thinking about ways we use theory could form part of the tool-kit of techniques available to evaluation practice.