Different designs answer different questions and should be judged by how well they match the question asked. Cross-sectional networks provide meaningful evidence when interpreted for what they actually estimate, and criticism should focus on unsupported inferences rather than the design itself.
