Community-driven Question Answering services are gaining increasing attention with tens of millions of users and hundreds of millions of posts in recent years. Due to its size, there is a need for users to be able to search these large question answer archives and retrieve high quality content. Research work shows that user reputation modeling makes a contribution when incorporated with relevance models. However, the effectiveness of different link analysis approaches and how to embed topical information---as a user may have different expertise in various areas---are still open questions. In this work, we address these two research questions by first reviewing different link analysis schemes---especially discussing the use of PageRank-based methods since they are less commonly utilized in user reputation modeling. We also introduce Topical PageRank analysis for modeling user reputation on different topics. Comparative experimental results on data from Yahoo! Answers show that PageRank-based approaches are more effective than HITS-like schemes and other heuristics, and that topical link analysis can improve performance.