Primary school students' relative preference, perceived effects for promoting learning, and perceived difficulty regarding online student-constructed tests (SCT) and student-generated questions (SGQ) learning activities were examined. Three fifth-grade classes (N=82) participated as part of their science learning activities for eleven weeks. An online system was adopted to support the associated learning activities. Five important findings were obtained. First, the majority of the participants did not associate SGQ or SCT with the feeling of difficult or very difficult, after being exposed to both tasks for an extended period of time. Second, a substantial proportion of students regarded SGQ and SCT as at the optimal challenging level. Third, X2 tests indicated that participants' preferences toward SGQ and SCT were not statistically significant (X2 =0.7), but were significant in perceived effects for promoting learning and perceived difficulty, with more participants selecting SGQ over SCT. Fourth, SGQ's facilitating effects for better promoting learning were well perceived despite that predominate percentage of the participants regarded SGQ as more difficult than SCT. Fifth, SCT's potential for knowledge integration and elaboration was rarely recognized by the participants. Implications for instruction are offered.