We develop an ordinal approach to comparing the equilibria of economic models. The main advantages of this approach, compared to the traditional approach based on signing derivatives, are that (1) it utilizes only a subset of the assumptions traditionally made, resulting in a simpler theory, (2) it applies to discrete changes and even when there are multiple equilibria and when some equilibria do not vary smoothly with the parameters, and (3) it incorporates a formal theory of the robustness of conclusions to assumptions, which helps modelers distinguish which assumptions are “critical” to their comparative statics conclusions