Accounting for error propagation and regression errors.

Hello,

I'm currently working with experimentally observed data of harmonic(ish) oscillations. Using the law of error propagation, I've estimated uncertainties of each measurement. When I fit the data to a cosine using nlinfit, the covariance matrix gives me the uncertainties of the fit on each parameter. I would like to use both of these uncertainties to describe how well my fitted curve represents reality, although I'm on bare grounds as how to do so. I see that nlinfit has an option called 'ErrorModelInfo', although I have not been able to find information on how to use this option, and truth be told, I don't even know if this option is the one I need to use to incorporate the uncertainties from the law of error propagation. Any help would be greatly appreciated.

Thank you for your time.