The sensitivity of the pendulum test to variation in spasticity in persons with spastic cerebral palsy (CP) was tested in 30 participants with CP and 10 participants without CP (controls) (mean age 13.8 years). The participants with CP were classified into three groups, with normal (mean age 15.9 years), mild/moderate (13.0 years), or severe (23.0 years) muscle tone, as assessed clinically using a modified Ashworth scale. Joint motion during the pendulum test was measured with an electrogoniometer. Muscle relaxation was confirmed using surface EMG. Outcome measures from the pendulum test were (1) number of oscillations, (2) duration of oscillations, (3) excursion of the first backward swing, and (4) relaxation index (first swing excursion/difference between the starting and resting angles). Data were assessed using one-way analysis of variance. Outcome measures 1 to 3 differed significantly between control participants and participants with CP (p<0.05). The first swing excursion was the best predictor of the degree of spasticity in persons with CP, being significantly different between all groups (p<0.05). The number of oscillations and their duration differentiated between control participants and all participants with CP (p<0.05) but not between participants with CP who had mild/moderate versus severe spasticity (p>0.05). The relaxation index was not a sensitive measure (p>0.05 between most study groups). We conclude that the pendulum test is a valid tool for assessing spasticity in persons with CP and that the first swing excursion is the most sensitive outcome measure.