While multi-method research is experiencing a surge of popularity, there are reasons to worry about whether multi-method applications are producing more grounded, justified, and persuasive inferences than single-method studies. These concerns arise from the'triangulation’ framework as prototypical research design for multi-method social science.Simply put, triangulation designs involve asking the same causal inference question using two different methods, checking same substantive conclusions. The metaphor is to the geometric technique of estimating distance by measuring the angle of sight toward an object from two different vantage points. Triangulation in social sciences has major flaws. One is well-known and widely discussed: what conclusion should be drawn when the two methods produce different findings? Unfortunately, the list of intellectually plausible responses to such an outcome is all but unbounded. For example, a scholar could reasonably conclude that both findings are correct but capture different aspects of the phenomenon of interest; that one method or the other displayed fundamental limitations in the analysis; that the divergence provides evidence against the credibility of the assumptions involved in both methods; that the outcome involves a contrast in terms of scope or relevant populations for the two methods; or that the failure to triangulate simply leaves the inference in a state of uncertainty. In decades of writing about triangulation, no definitive guidance has yet emerged about how to respond to divergent findings between the multiple methods in a study.