To determine whether parent and child visual analogue scale (VAS) scores for the pain associated with acute conditions in the child agree sufficiently for these methods of measurement to be considered interchangeable in pain and analgesia research. This was a prospective, two-group, repeated measures, blinded study in an urban pediatric emergency department. Children aged 8 to 15 years seeking treatment for painful conditions and the parents of these children were asked to rate the child's pain independently using a VAS on as many as four occasions at 20-minute intervals. Both participants were blinded to their previous rating and the rating of the other participant. The main outcome measure was the correlation of child and parent VAS pain scores by Pearson correlation and bias plot (Bland-Altman) analysis of agreement between tests. Seventy-eight child-parent sets participated, yielding 289 VAS pain score comparison pairs for evaluation. The correlation between child and parent VAS pain scores was 0.63 (95% CI, 0.56-0.70). Bias plot analysis revealed a bias of 5% and 95% limits of agreement from -38 to +47 mm. The degree of difference between child and parent scores was variable, but there was an increasing tendency for parents to underestimate the child's pain when the child recorded VAS pain scores at the higher end of the scale. Parents' VAS score ratings of their children's pain correlate only moderately with the children's VAS pain scores and show poor levels of agreement. The difference between the measures is variable and appears to be more marked when the child reports a higher VAS score. This research raises doubt about whether parental rating of a child's pain is an appropriate surrogate marker in pediatric pain and analgesia research.