In this study, 10 sagittal lower limb range-of-motion measures were conducted in a blinded fashion in 25 children with spastic diplegic cerebral palsy and in 25 age- and sex-matched controls. The participants comprised 22 males, mean age 10 years 8 months and 28 females, with mean age 9 years 8 months; age range 6 to 17 years. One paediatric physical therapist performed duplicate goniometric measures at zero time and 7 days later using the same sequence of measures, location, and time of day. Mean absolute differences for measures within one session ranged from 0.7 to 2.9° in controls and from 1 to 4.2° in children with spastic diplegia. Most intraclass correlation coefficients (ICCs) for intra-sessional measures were more than 0.90 in both groups. Measures between sessions were less reliable. Mean absolute differences between sessions were up to 7.1° for children with spastic diplegia and 8.6° for controls, with most ICCs being less than 0.80. Inter-sessional variation in measures was similar in both groups, suggesting that measurement variability is not influenced by the presence of spasticity. Averaging of two measures did not improve inter-sessional reliability compared with the use of a single measure. Dynamic measures (R1) were as reliable as passive measures (R2), but there were inter-sessional differences in calculations using R1 and R2 measures of up to 30°.