In view of contribution-based pedagogy and observational learning theory, students' perceived uses, preferences, usage, and selection considerations with regard to citing peers' work were examined in an online learning environment targeting student-constructed tests. Data were collected from 84 fifth-grade students who participated in online student-constructed tests with and without citing in an 11-week study. Quantitative and qualitative data in response to an end-of-session questionnaire and actual online citing behaviour were analyzed. Several major findings were obtained. First, significantly more participants supported and preferred "citing" over "no citing" for online student-constructed tests. Second, data with regard to perceived uses, preferences, and reported usage all supported the potential of citing for providing an observational learning space. Third, citing allowed the participants to attend to areas pinpointed by their peers but initially ignored by them, thus making social construction of knowledge possible. Fourth, the quality and the author of the item are the two determining factors affecting citing decisions. Fifth, a statistically significant positive correlation between students' academic achievement and their generated questions cited by peers was confirmed. Finally, actual online citing behaviour varied greatly among participants, with the majority using the citing function during online test-construction to various extents.