This study examined the use of discovery learning in science and how it affects students' academic performance as well as their self-efficacy in science. It also used a diagnostic tool to identify students' misconceptions about processes in the water cycle and where the misconceptions originated. While the study showed that the treatment group had a statistically significant greater academic gain from the pre-test to the post- test than did the no-treatment comparison group, from a teachers view point the gain would not be enough to benefit a student's performance on high stakes tests. Because the study was able to identify eight common misconceptions, it suggests that the misconceptions that students possess are difficult to uproot even using teaching methods that have been proven successful. / CITE/Mathematics and Science Education