This study was, a Web-enhanced course, designed to investigate the effect of presence and absence of instructor in an online discussion on learner participation, perceived social presence and quality of scientific paper writing. Gender and cognitive style were considered as moderating variables affecting participation, social presence and research writing skills. 52 postgraduate Master of Information and Communication Technology students, from the Faculty of Education-Albaha University in Saudi Arabia, participated in a Scientific Research Method course and were assigned to either a treatment condition (instructor intervention-instructor postings; n=26) or control condition (no instructor intervention-peer postings; n=26) for six weeks online discussions about designing experimental researches. Four discussion forums were designed as web-enhanced tools to accompany with a traditional research method course. Two forums were designed for male students and the other two were for females because mixed-sex discussions were not possible. Amount of participation was measured with the frequency and percentage of postings while the type of participation was analyzed based on a coding scheme rubric designed for this purpose. Learner's perceived social presence was estimated with a 30-item questionnaire type scale. Quality of learner in writing a scientific research paper was measure based on a 10 item-standardized rubric. Classifying participants into dependent and independent learners were done using the GEFT which was a post procedure. The results of the study revealed several significant differences. First, amount of participation in online discussion was affected by instructor intervention and gender type; with the peer facilitated condition produced more responses than instructor intervention condition and gender type was found to be a moderating effect with respect to amount of postings, with the male learners producing more postings. Cognitive style found to affect amount of postings with dependent learners being higher. Second, related and unrelated postings were affected by gender type, with the males posting more off-topic responses whereas for female the reverse was observed. Third, with respect to social presence, a significant effect was obtained with instructor interventions, with the peer facilitated treatment showing higher social presence than instructor presence treatment. Social presence was also found to be affected by gender type, with the females being feeling higher than males. Dependent learners had higher perceived social presence than independent learners. Fourth, quality of paper writing appeared to be affected by interventions and gender, with the no instructor condition increased writing quality and female participants being better. Finally, the study also found two and three way interactions. The results of the study were discussed in terms of their implications for designing online discussions.