The purposes of this study were to compare age, static balance performance, and step-width variables between elderly noninstitutionalized women with and without a history of falls and to determine the relationship between balance performance and step width. Each subject performed a maximum of three time trials on the sharpened Romberg and one-legged stance tests with eyes open and with eyes closed. The first and best trial measurements were used for analysis. Each subject walked on paper walkways making ink prints for step-width measurements. The mean and the variability of each subject's step-width measurements were used for analysis. Data from 110 women, aged 60 to 89 years, were analyzed. The fallers (n = 26) had sinificantly lower values than the nonfallers (n = 84) on the best trial of the sharpened Romberg test in the eyes-open condition (t = 1.98, df = 108, p < .05). No significant differences between fallers and nonfallers were revealed in age, the mean and variability of step width, the first trials of the balance tests, and the best trials on the other balance tests. For the total group, the mean measurements on the first trials were significantly lower than those on the best trials for each balance test. Small, but statistically significant (p < .05), negative relationships existed between balance performance and the mean and variability of step width. The results of this study indicate that the methods of measuring balance and step width are clinically applicable, and the data of patients from a similar population sample may be compared with the data established in this study.