Two experiments reproduced in a simulated computerized test-taking situation the effect of two of the main determinants in answering an item in a test: the difficulty of the item and the time available to answer it. A model is proposed for the time to respond or abandon an item and for the probability of abandoning it or answering it correctly. In the first experiment, the 47 undergraduate participants were instructed to answer before a certain number of seconds. Results of this experiment suggest that the process to cope with an item is a two-stage decision process in which subjects check to see if the demands of an item match the available time resources. If so, they continue looking for a solution, and if not, they abandon the item. The 48 subjects in the second experiment had to meet an overall deadline of 12 seconds when "tests" or triplets of 3 items were presented. In most subjects, the probability of abandoning the item was still a function of the item's difficulty, but fewer subjects showed an effect of the item's deadline. These results show that under a more realistic test-taking situation, the subject's perception of a deadline to produce an answer is not clearly affected by the external indicators of time availability. This research contributes to the understanding of test-taking behavior in the context of computer-based tests. (Contains 13 figures and 30 references.) (SLD)