Below are some general data analysis tips. For more in-depth tips and tutorials, look here.
- Back up your data! A given, but very important. And don't throw out unsuccessful results of old analyses until you're absolutely sure you don't need them anymore
- Budget in plenty of time - real analyses are much more messy than tutorials
- Be prepared to not see the results you're looking for right away - it's just part of the process. There are so many factors to consider, you won't get them all your first go. Errors can happen at different levels, and you'll have to troubleshoot each rigorously:
- your study design*
- execution of your study*
- your model
- your analysis methods
- your code
- dataset idiosyncrasies, e.g. motion or magnet irregularities in the first few TRs for fMRI, data that needs to be demeaned to remove collinearity, etc.
- You may have to analyze your data from many different angles, depending on the complexity of your dataset; be patient, and don't get too attached to one analysis strategy
- Ask around - talking to fellow lab members as well as non-lab members can give you ideas for new ways to analyze your data, and can help you figure out what's going wrong when you've exhausted all your own ideas
- Always stay tuned in to what's happening on the cutting edge of data analyses - make notes at conferences, when you come across new papers, etc. This can also give you new ideas