Topics within small-group communication have been explored in many contexts, such as work group, organizational meeting, or online network. This area of discipline is considered crucial because this type of communication assimilates interpersonal relations within a social setting. Two elements that largely affect small-group communication dynamics are anonymity and social identity. This research invokes previous research in anonymity and social identity within small-group communication pertaining to the level of agreement and the level of group attraction through a series of experiments. Anonymity in small-group communication context is defined as a condition where the group members are not identifiable. To create anonymity among group members, this study utilized the benefit of a chat room in computer-mediated communication (CMC), which allows group members to participate in group discussion anonymously without the fear of being judged. It is argued that groups communicating synchronously via CMC would have a higher agreement than those communicating face-to-face (FtF) because the anonymity in CMC eliminates all of visual cues and therefore, unites all group members. It is also argued that members in groups in FtF are more likely to be interpersonally attracted than those in CMC. Thus, members communicating via FtF would have larger cumulative group attraction than those in CMC. Meanwhile, social identity in small-group communication context is defined as the tendency of a group member to associate with fellow members who share similarities with him or her and hold prejudice against members who are different than him or her. The element of social identity that was being activated in this study was the gender identity. This was done through using a gender-related case, an opinion scale, and distributing participants into groups of different gender compositions. It is argued that single-gender groups would have higher level of agreement and group attraction than mixed-gender groups. The experiment assigned participants into six different groups. The groups communicated via FtF or via CMC. In each setting, there were male-only groups, female-only groups, and mixed-gender groups. The only statistically significant result from the experiments suggested that in CMC, female-only groups had a higher level of agreement than mixed-gender groups. However, there were also differences of mean agreement between female-only groups in FtF and female only groups in CMC. Those communicating via CMC had higher agreement. In terms of level of group attraction, there was not any significant result in any condition. This finding suggests that in CMC, groups that are exclusively females are more conducive than other gender compositions in reaching agreement. Meanwhile, the lack of significance in group attraction between FtF and CMC suggests that people have become more familiar with anonymous CMC settings allowing them to substitute the available textual cues for visual cues.