The reliability and validity of force platform measures used to evaluate steadiness of stance were investigated in a group of 28 healthy subjects in four basic stance positions: two-legged, step, tandem, and one-legged stance. In each stance position five indices of steadiness were obtained by sampling the three orthogonal force signals and the two horizontal center of pressure (CP) signals for 15 seconds and computing the standard deviation for each signal. Correlations between these five indices derived from the force platform showed that the relationship between force and CP measures was generally weak. In fact, approximately 40% of the correlations were nonsignificant (p greater than .05). There was a strong trend for the retest reliability of force measures to be higher than the retest reliability of CP measures. The difference was statistically significant in three stances (p less than .05). Force measures were more sensitive than CP measures in discriminating the changes in steadiness which resulted from alterations to the base of support in the four stance positions. Factor analysis showed that force measures were the best predictors of steadiness in each stance, but the axis varied according to the particular stance condition. Although it may be appropriate in some cases to justify the choice of measure according to the nature of the clinical condition being studied, these results provide a rationale for choosing force measures in preference to CP measures on the fundamental principles of reliability and validity.