Two studies are reported here that investigated elementary school chil
dren's text revision. In the first experiment, both semantic and surfa
ce flaws were inserted in texts that varied in reading difficulty. Sec
ond-grade through fifth-grade students revised these experimenter-gene
rated passages, presented as examples of submissions to a class newspa
per. Differences in text reading difficulty did not affect revision ef
fectiveness, nor were the semantic flaws especially difficult to detec
t and revise. An age effect showed growth in the revision of both sema
ntic and surface errors from grades 2 to 4 with 2nd-graders revising o
ne-third of the inserted errors, and 4th- and 5th-graders revising thr
ee-quarters of them. Revision and doze reading comprehension skills we
re correlated. A second study compared students' revision of their own
as well as another's text flaws. Fifth-graders wrote a narrative for
a classroom anthology, and they revised both their own and inserted fl
aws. Their writing was evaluated holistically. Rates of both semantic
and surface revision were somewhat lower for their own as opposed to a
nother's text errors, but revision rates were nevertheless relatively
high, and they correlated with writing quality; that is, children who
wrote high-quality texts also revised more errors, especially experime
nter-inserted flaws. These data confirm that children respond positive
ly to writing challenges in the area of revision, a skill in process o
f development, which is amenable to inspection and appears ripe for fa
cilitation.