A study compared students' ability to read their stories written with invented spellings with their ability to reread their stories rewritten with conventional spellings. Fourteen first graders wrote stories using invented spelling and read the text aloud. The stories were then transcribed into conventional text by an adult and reread by the students. Readings were evaluated, using the Goodman-Burke Miscue Analysis, to determine which of the two versions produced the fewest miscue errors. Results indicated that most categories of miscue errors were very close in number, giving no strong indication that either invented spelling or conventional text is an easier format for children to read. When children were divided according to those who performed best in each format, the results showed that nine had fewer miscues on their invented spelling readings, while five had fewer miscues on the conventional spelling versions of their texts. The results suggest that children pass through various stages when learning to read and write, one of which is invented spelling. Use of invented spelling did not appear to hinder children from moving to conventional spellings when ready. (HTH)