In the courtroom, child witnesses must undergo cross-examination. Prior research has shown that children change their original testimony under cross-examination, but the effect of these changes on the accuracy of children's testimony is unknown. The authors examined the effect of cross-examination on the accuracy of 5- and 6-year-old children's (N=46) reports of a contrived event. Consistent with prior research, children made changes to their original responses during cross-examination. Furthermore, these changes occurred irrespective of original accuracy. Finally, prior exposure to misleading information did not affect children's responses to cross-examination. Even children with no prior exposure to misinformation altered their original responses, decreasing their ultimate levels of accuracy. These findings demonstrate that cross-examination style questioning is inappropriate for young children.