Forty-four children aged 3 to 11 years with quadriplegic cerebral palsy were prospectively stratified and randomised into four treatment groups. The acquisition of motor skills was assessed in a 2 x 2 factorial design using the Gross Motor Function Measure. The two factors were conventional amounts of physiotherapy vs intensive amounts of physiotherapy, and the use of broad, generalised aims vs the use of specific measurable goals directed at motor skill acquisition. 82% of the children improved. Over the two-week period, intensive physiotherapy produced a slightly greater effect than conventional physiotherapy but the factor more strongly associated with increased motor skill acquisition was the use of specific measurable goals.