You can measure something consistently wrong every single time—that's reliable but invalid. Or your results can scatter widely and land on the true value only by luck—that's unreliable, and an unreliable measure can't be valid either. So reliable but invalid is possible; valid but unreliable is not. Psychology research demands both, and each needs its own check. Raise the confounding strength to see how confounding variables erode internal validity even when measurements stay reliable.
External validity determines whether your findings travel beyond the laboratory walls. Ecological validity asks if your carefully controlled experimental task resembles anything people actually do in real life—does a word-list memory task reflect how students study for tests? Population validity questions whether your sample of 120 undergraduate psychology students tells us anything about decision-making in middle-aged executives, rural farmers, or elderly hospital patients. High internal validity often comes at the cost of external validity because tight experimental control creates artificial situations.