Attentional functions are essential to safely and smoothly perform activities of daily living. A range of different symptoms is observed in patients with impairments of attentional functions. Thus, the same patient may show differences in symptoms depending on the environment in which he or she is situated. In this study, we sought a means to characterize attention deficit with higher sensitivity than conventional neuropsychological tests, such as the Trail Making Test (TMT) and the Clinical Assessment for Attention (CAT), do. We developed two reaction time (RT) tasks: a simple task and a speed change task. We then enrolled 10 right-handed healthy older adults with no history of brain injury and 10 right-handed patients with stroke to perform the RT tasks. In addition, the stroke group underwent the TMT, the CAT and behavioral observation by an occupational therapist to identify symptoms characteristic of attention deficits. Results for findings of attention deficits on standard neuropsychological tests had a sensitivity of 25% in static situations (e.g., desk activities) and 33% in dynamic situations (e.g., walking or moving). In contrast, when applied to static situations, the simple and speed change RT tasks achieved a sensitivity of 75% and 100%, respectively. The sensitivity in dynamic situations was 33% to 44% for the simple task and 100% for the speed change task. These results suggest that the RT tasks developed in this study are capable of identifying attentional deficits in patients with stroke, and may be more sensitive than neuropsychological testing combined with behavioral observations.