To examine the potential role for a placebo cream in reducing reported needle pain severity in children, and the impact of age-related factors on pain self-report, a convenience sample of 117 children scheduled for venipuncture were randomly assigned to one of three treatments: (a) placebo cream with the suggestion that it might help reduce needle pain, (b) placebo cream with no indication as to the cream's purpose, and (c) no cream (control group). In allocation to treatment, children were stratified by age group, (3-7, 8-11, 12-17 years). They rated their needle pain severity (both predicted and reported) using the Faces Pain Scale, and rated their anxiety about the procedure using the Children's Anxiety and Pain Scale. Children in the cream groups were also asked whether they thought the cream had helped. Using video-tapes, an independent observer, blind to the placebo manipulation, rated each child's reaction to the needle. For the two groups receiving cream, 83% of those children told it might help stated that they believed it did, as compared with only 33% of children who received the cream but were told nothing of its purpose; These beliefs, however, were not reflected in self-report ratings of pain which showed no statistically significant treatment effect. Similarly, children who gave higher preprocedural anxiety ratings were no more likely to report less pain as a result of receiving the cream. There was, however, a treatment effect on the observer's ratings: children receiving cream plus suggestion were assigned significantly lower ratings of pain-related behaviour than those children who received the cream alone. While venipuncture was associated with only mild levels of pain, younger children, irrespective of treatment group, did report more pain than older children. Hierarchical regression analysis indicated that 60% of the variance in self-reported pain severity scores could be accounted for by how much the child thought the needle would hurt, how anxious the child was about receiving the needle, gender (higher pain ratings associated with girls), and estimated body surface area (higher pain ratings associated with smaller bodies). We conclude that the efficacy of placebo treatments for needle pain in children may depend on the suggestion of a possible benefit rather than upon treatment application per se.