Researchers increasingly take advantage of the comparative case design to build theory, but the degree of case dependence is occasionally discussed and theorized. We suggest that the comparative case study design might be subject to an often underappreciated threat--dependence across cases--under certain conditions. Using research on innovation diffusion as an illustration, we explore the role of social linkages across cases when building theory through comparison and contrast between cases. We develop an agent-based simulation, grounded by comparative case research about innovation diffusion, as novel way to study the implications of case dependence in theory building using multiple-case study research. Our simulation results suggest that the degree of case dependence has a nontrivial bearing on innovation diffusion experienced by case entities, specifically when the researcher draws a few case entities operating in a highly interconnected industry. Under these conditions, overlooking the degree of case dependence might weaken newly built theory against commonly held standards of internal validity and external validity in inductive research. We conceptualize the issue of case dependence as a concern about researchers' bounded rationality. Accordingly, we build on our findings to provide actionable advice aiming to alleviate this concern while being amendable to the variety of approaches to build theory from multiple cases in social sciences.