Variability is commonly observed in complex behavior, such as the maintenance of upright posture. The current study examines the value added by using nonlinear measures of variability to identify dynamic stability instead of linear measures that reflect average fluctuations about a mean state. The largest Lyapunov exponent (λ1) and SD were calculated on mediolateral movement as participants performed a sit-to-stand task on a stable and unstable platform. Both measures identified changes in movement across postures, but results diverged when participants stood on the unstable platform. Large SD indicated an increase in movement variability, but small λ1 identified those movements as stable and controlled. The results suggest that a combination of linear and nonlinear analyses is useful in identifying the proportion of observed variability that may be attributed to structured, controlled sources. Nonlinear measures of variability, like λ1, can further be used to make predictions about transitions between stable postures and to identify a system’s resistance to disruption from external perturbations. Those features make nonlinear analyses highly applicable to both human movement research and clinical practice.