Objective: The main objective was to assess the relative frequency and types of errors on the clock-drawing test (CDT) in normal elderly subjects, and the dependence of their CDT performance upon their age and education. Method: The CDT performance of 242 normal elderly subjects, stratified according to age and education, was analyzed using both an absolute error count and a modified scoring sheet derived from 2 scales. Results: Normal elderly subjects (average age: 73.4 ± 8.4 years; mean Mini-Mental State Examination score: 27.7 ± 1.6) often have problems placing figures on the clock face and differentiating the clock hands correctly. The absolute number of errors increases with age and decreases with formal school education. A cumulative effect of both high age and a low level of school education yields substantially inferior scores, on the modified scale used, in comparison with younger subjects and those with higher levels of education. Conclusion: The frequent errors made by normal elderly subjects, and a combined negative influence of high age and low education level, suggest caution when interpreting the significance of CDT performance in the elderly.