Given that learner reflections are associated with cognition and emotion elements, an issue has been raised regarding how to organize, classify and evaluate reflection contents. With respect to reflective behaviors, Morgan (1999) reported a hierarchy with four levels of reflective thinking: Not qualified, Fair, Good, and Excellent. Teachers from South Brunswick schools adopted assessment criteria comprising purpose of reflection, supportability, systematicity, sentence structure and vocabulary and writing mechanics, each of which was accompanied with a four-level scheme (King-Shaver, 1999). Cheng (2002), who assessed reflective thinking in a Web-based portfolio setting, discovered that students generally performed medium level and failed to attain higher level of reflection. This study result was supported by several researchers. Lin (2004) indicated although shallow and deep types were found existing, students tended to perform the former type more frequently. Likewise, Wood (2001) concluded students had higher possibility to operate pre-reflective thinking, comparing to quasi-reflective or reflective thinking. A study conducted by Li (2002) suggested lower levels of reflective thinking appeared to be more accessible to high school students. In other words, previous investigations have revealed the tendency of reflective thinkers toward lower levels of reflection.