This follow-up experiment will explore the effects of altering relative task difficulty. Previous follow-ups have suggested that when presented with the opportunity to use both short-term change detection information with long-term recognition, subjects either cannot or do not make use of both sources of information at once. They seem to default to one task, namely the 6AFC memory task. This may be because the recognition task results in higher accuracy, or because it is easier. This follow-up will explore whether this default-task is selected to maximize accuracy or minimize effort. To do so, we will alter the change detection tasks from the previous experiments to have a much longer ontime for the pre-change array. More encoding time increases change detection performance (Brady, Konkle, Oliva, & Alvarez, 2015), and by increasing pre-change ontime we can raise change detection accuracy above that of 6AFC accuracy. This will make change detection significantly easier. Now, when faced with a hybrid memory/change detection task, subjects can perform pure change detection (and thus have accuracy higher than 6AFC), perform 6AFC memory recognition (and thus have lower accuracy than unstudied change detection), or perhaps combine them both now that the change detection task is easier. We will also change the order in which tasks are presented, so that subjects will complete unstudied change detection and 6AFC memory recognition first, and then complete the hybrid task. It may be that subjects need prior exposure to the component tasks before they can combine the strategies from both. The experimental procedure and data analysis will be similar to that of the other experiments. Subjects will study a series of objects and then complete three tasks: change detection with unfamiliar objects at a long pre-change exposure, with instructions simply to find the change; the 6AFC recognition memory test; and change detection with familiar objects, with instructions to attempt to find the change, and that if they see a familiar item, it is the one that changed. If it is the case that subjects do not or cannot spontaneously task-switch and instead default to one or the other, results from this experiment will reveal whether this is done in order to maximize accuracy (in which case subjects should persist in change detection) or mimimize effort (in which case subjects should not bother to encode the six objects in the pre-change array and instead just look for the familiar item in the post-change array). Or it may be that with change detection made sufficiently easy, subjects can combine both sources of information; in that case, we should see an additive increase in accuracy beyond 6AFC or unstudied change detection performance. The data analysis procedure will examine the following comparisons: * Unstudied change detection accuracy vs. hybrid task accuracy * Unstudied change detection accuracy vs. 6AFC memory accuracy * Hybrid task accuracy vs. 6AFC accuracy