To examine the relationship between children's social-cognitive status and the production of persuasive arguments and to determine the effects of age and the channel of communication on these arguments, 16 first grade, 17 third grade, and 16 fifth grade students were asked to produce both oral and written persuasive messages and were administered a measure of cognitive role taking. Significant partial correlations between the role taking measure and the following measures of written persuasives were found: variety of requests, highest level of request, and total number of supports for requests. The results provided weak support for the notion that persuasion reflects children's social-cognitive status. Age effects on all measures of persuasion indicated that children's arguments became longer, more varied, and more complex with age. Age and channel interactions for variety and total number of requests suggested that the written channel was facilitative for only fifth grade students in the sample. (MM)