The questions of the relevance of the speed of the test takers to the ability being tested was examined via a factor analysis of 3 timed sections of the GRE--a 25 minute 55-item verbal reasoning section, on which 46.9% of the candidates attempted the last item; a 50 minute, 40-item passage comprehension section on which 70.9% attempted the last item, and a 75 minute 55-item quantitative section, on which 23.8% of the candidates attempted the last item. Responses of 8,000 candidates were analyzed. It was concluded that speed was relevant to quantitative ability, irrelevant to comprehension of connected discourse, and weakly correlated to science-related passages and reading comprehension. Limitations on interpretations of Stafford's speededness quotient were noted. Other practical considerations and problems in determining when speededness is relevant include paying particular attention to the omitting group on standard item analyses, examining the relationship of ability as estimated by the first part of the test to ability as measured by the final section of the test, and inability to determine the relevance of speededness if subjects scan items first and skip the difficult ones.