This is not necessarily the case because in order to construct a norm-referenced test where some test-takers score above average while the other score below, test makers must make small differences among people appear large. Since item content differs from one exam to the next, even tests that claim to measure the same thing often produce very different results.