We report a spatial-memory scanning experiment that was used to measure age differences in entropy. A target grid consisting of four adjacent letters followed by the presentation of a single probe letter was presented on each trial. Half of the trials presented the probe stimulus in the same spatial position was the target letter (i.e., the probe letter was always a member of the positive set), and half of the trials transposed the target letter one, two, or three spaces of the right or left of the original target display position (i.e., different trials). The experiment involved blocks of primary-memory and secondary-memory tasks. Reaction-time and error-rate data, as well as entropy analyses and the fitting of an entropy model (based on Allen, Kaufman, Smith, and Propper, in press) to the empirical data indicated that older adults showed higher entropy levels than young adults. These results are interpreted in a "computational temperature" framework in which older adults' higher computational temperatures result in less efficient spatial, episodic memory functioning.