Lazar et al. have gathered a nice collection of central methods in HCI, and also in analysing both quantitative and qualitative data. The part of analysis starts from the very beginning, so it is a very good and practical book for beginners in statistical analysis.
For me, maybe the most revealing thing was the scientific approach and reminder of the fact that no scientist or researcher is good in every phase of empirical study. Each discipline that affects to HCI tend to focus on one phase of the process. In their model, computer science and design tend to focus on implications of the research and study, whereas psychologists focus on research methods and sociologists on test subjects and test context. I found from myself a big potion of computer scientist and a small psychologist