As online platforms become more popular in higher education, students may use these platforms more frequently since they remove social and geographic barriers. However, many students don’t seek help online effectively, a crucial skill for learning and developing programming knowledge. Data captured on students’ help-seeking behaviors could serve as valuable information to direct feedback and just-in-time interventions for instructors. In this paper, we examined online help-seeking behaviors (the frequency and popularity of posts) of students (n=240) in a CS1 course. In addition, we analyzed the linguistic features of the students’ posts on an online Q&A platform. Our study uncovered two categories of online personas: askers (students who asked questions, n=136) and lurkers (students who never posted on the online platform, n=104). Furthermore, askers were divided into two sub-categories: answerers (askers who also answered other students’ questions, n=29) and non-answerers (askers who never answered other students’ questions, n=107). We assessed the association between these personas and course performance and found that askers performed significantly better than lurkers. However, there was no statistically significant difference in performance between answerers and non-answerers. When investigating the linguistic features of askers’ posts, and their relationship with learning outcomes, we found that students with more readable posts (above the median readability score) performed significantly better than students who posted less readable content. Finally, we found that students with posts that received a high number of votes (above the median), and/or a high number of views (and/or unique views) performed better than the ones that did not.