A standardized test development committee pilots two new questions. On Question X, students from historically high-scoring demographic groups outperform students from historically low-scoring groups. On Question Y, the opposite occurs. The committee decides to include Question X and discard Question Y, arguing that a question is 'reliable' only if its results are consistent with the test's historical performance patterns. What is the most significant flaw in the committee's reasoning?