Risk of falling is a major concern of long-term care facilities with residents diagnosed with dementia. Use of a brief cognitive assessment focusing on visual spatial abilities could be one strategy in the prevention of falls in residents with dementia. The objective of this study was to determine if a clock test could predict a risk of falls in residents diagnosed with dementia. Three hundred sixty-four individuals with dementia participated (92 men and 272 women; ages 37 to 100, mean 80.5 years, median 83 years). Each participant was given the Reality Comprehension Clock Test (RCCT) three times, and Mini Mental-Status Examination (MMSE) two times to determine criterion-related validity, test-retest reliability, internal consistency; and to set and evaluate a risk of falls score based on the Visual Spatial Score (VSS) component produced by the RCCT. Significant findings included .72 (p < .01) correlation between the RCCT and the MMSE, .90 (p < .01) correlation between the first administration of the RCCT and the second administration of the RCCT; an alpha of .95 (p, < .001) and an F value of 7.6 (p < .001) producing a risk of falls initial VSS of 5 or lower compared to 9 or greater. Chi-square of 6.3 for 30 days (p,< .01), 11.08 for 60 days (p < .01) and 13.3 for 90 days (p < .01) indicated a significant difference in the number of falls occurring in the high risk group (VSS of 5 or lower) compared to the low/ no risk group (VSS of 9 or higher). A risk factor analysis suggested that residents in the higher risk group were three times more likely to have fallen than residents in the low risk group. Knowing a resident's visual spatial ability offers health care providers an opportunity to implement a resident-specific intervention that addresses their cognitive ability and visual spatial deficit that may reduce the resident's risk of falling.