Hi, I am new to brms and have a question regarding the use of the mi() function in brms for incorporating measurement errors into predictions when fitting a GAMM. I’ve read the related documentation and forum posts on mi(), but I’m still unsure how to apply it correctly in my model.
I am trying to predict the effect of temperature on the movement distance of an animal using the following GAMM formula:
Movement_distance ~ s(Temperature) + s(ID, bs="re")
In my case, the measurement error for the movement distance ranges from 0 to 20 meters and follows a normal distribution. I would like to account for this error in my model.
I was thinking of using the following formula:
Movement_distance | mi(20/2.58) ~ s(Temperature) + s(ID, bs="re")
Since 99.9% of the measurement error falls within 20 meters, I calculated the standard deviation as 20/2.58. Is this approach correct for implementing the measurement error, or is there a better way to handle this in the model?
Thank you very much for your support!