For individuals with neuromuscular disorders, the assessment of walking energy efficiency is useful as a clinical outcome measure. Issues surrounding data collection methodology, normalization of the data, and variability and clinical utility of energy efficiency data preclude universal application. This study examined the variability and the clinical utility of velocity, energy efficiency index (EEI), gross cost, and net nondimensional cost (NNcost) in children and adolescents with spastic diplegic cerebral palsy (CP) in Gross Motor Function Classification System (GMFCS) levels I to III. The energy efficiency of walking was evaluated in 23 children and adolescents (12 males, 11 females, mean age 11y 3mo [SD 3y 5mo]; range 7-17y). Day-to-day variability was similar for all energy efficiency variables, with no significant differences in magnitude of variability between GMFCS levels. Correlations between EEI and gross cost and EEI and NNcost were fairly good (r=0.65, p<0.001, and r=0.74, p<0.001 respectively). However, only gross cost and NNcost discriminated between GMFCS levels in children with CP. Gross cost required the greatest amount of change to be considered clinically significant, whereas NNcost and EEI required a similar amount of change. For cohorts of children with CP who are evaluated over time, NNcost is the best normalization method as it reduces the variability between participants of different ages, height, and weight while evaluating only the amount of energy used to ambulate.