The greatest source of variance was the test format—effects from timed tests were larger than from untimed tests, and they argued that this was because timed tests were more likely to assess a student's capacity to automatically apply word recognition and decoding skills.
