The aim of this study was to establish the reliability and validity of visual gait assessment in children with spastic diplegia, who were community or household ambulators, using a modified version of the Physicians Rating Scale, known as the Observational Gait Scale (OGS). Two clinicians viewed edited split-screen video recordings of 20 children/adolescents (11 males, 9 females; mean age 12 years, range 6 to 21 years) made at the time of three-dimensional gait analysis (3-DGA). Walking ability in each child was scored at initial assessment and reassessed from the same videos three months later using the first seven sections of the OGS. Validity of the OGS score was determined by comparison with 3-DGA. The OGS was found to have acceptable interrater and intrarater reliability for knee and foot position in mid-stance, initial foot contact, and heel rise with weighted kappas (wk) ranging from 0.53 to 0.91 (intrarater) and 0.43 to 0.86 (interrater). Comparison with 3-DGA suggests that these sections might also have high validity(wk range 0.38-0.94). Base of support and hind foot position had lower interrater and intrarater reliabilities (wk 0.29 to 0.71 and wk 0.30 to 0.78 respectively) and were not easily validated by 3-DGA.