The aim of this study was to examine the repeatability of and relationships among spasticity, co-contraction of agonist-antagonist, and muscle strength in children with cerebral palsy (CP). Eight children with spastic diplegic CP (five males, three females; Gross Motor Function Classification System [GMFCS] Levels I-III; mean age 10y 2mo [SD 2y 9mo], range 6-13y) and nine children in a comparison group (six males, three females; mean age 8y 10mo [SD 2y 4mo], range 6y to 12y 6mo) were assessed twice to examine repeatability of Composite Spasticity Scale, soleus stretch reflexes, electromyography (EMG) co-contraction ratio, and torque recorded during maximal isometric voluntary contraction of ankle dorsiflexors and plantarflexors. Sixty-one children with spastic CP, (54 diplegic, seven hemiplegic; 32 males, 29 females; GMFCS levels I-III; mean age 10y 8mo [SD 2y 9mo], range 6-15y) were then assessed to delineate possible correlations among these measures. Intraclass correlation coefficients (0.78-0.97) showed high data repeatability in both groups. Children with spastic CP demonstrated significantly larger soleus stretch reflex/M-response areas smaller torques, but larger EMG co-contraction ratios during both voluntary dorsiflexion and plantarflexion (all p<0.05). Children with spastic CP who had larger soleus stretch reflex/M-response areas demonstrated larger plantarflexion co-contraction ratio (r = 0.28), and produced smaller plantarflexion and dorsiflexion torques (r = -0.48 and -0.27 respectively). However, no correlation was noted between soleus stretch reflex and clinical spasticity. Our findings demonstrated that hyperactive soleus stretch reflex affected torque production of ankle muscles. Moreover, the severity of spasticity may not be fully described by either stretch reflex or tone measure alone.