Hello, and welcome! It's impossible to say without knowing more about the data structure. That said...there may be no degrees of freedom left over for error. i.e., you are trying to estimate n-1 effects with n points. Can you provide more detail and context about your data and what you hope to get out of it?