Circular Regression with grouped data and two explanatory variables

I am looking for the right test and software to do a regression test with circular data. I have one dependent variable consisting of circular data (degrees) and two independent variables which are also circular (degrees). Furthermore, the data is grouped: There are multiple trials, locations and days. I want to test how well the independent variables explain the dependent variable. Do you have any clue which test could be appropriate to test this?


Fortran must die
I am not sure what you mean by circular data. If you mean data where X influences Y and Y influences X that violates the rules of standard regression. Some forms of structural equation models and time series will address this, will permit it. Neither are simple.

Please explain what you mean by circular data.


Active Member
id bet the circularity in circular data is compass direction. Im not big on the subject though ive seen it mentioned here and there. I'd be interested in teh answer myself.