This study aims to examine if and how academic achievement and gender group composition affect the quality of online SGQ and the use patterns of procedural prompts provided to support SGQ activities. Forty-one university sophomores enrolled in an English as a foreign language class participated in a four-week study. All questions generated were categorized based on the revised Bloom's taxonomy for quality evaluation, and a content analysis along the set of integrated online procedural prompts was conducted to reveal usage patterns. Five key findings were obtained: First, the provision of the online procedural prompts served as an efficacious learning scaffold to help the participants at both high- and low-academic achievement levels generate the most questions at high-cognitive levels. Second, based on the results of the Fisher's exact test, no significant relationships were found between academic achievement and the quality of SGQ. Third, the participants in the all-male and mixed-gender groups generated the majority of their questions at high-cognitive levels, whereas the all-female group generated an equal number of questions at both low- and high-cognitive levels. Fourth, no significant relationships between the gender group composition and the quality of SGQ were found according to the chi-square test of independence. Fifth, the results of the content analysis revealed that while some same usage patterns related to online procedural prompts were exhibited by students at both low- and high-academic achievement levels and with different gender group compositions, slightly different usage patterns were observed.