I am consulting with another student with a stats background, but am pretty sure he his saying something that contradicts what I have been told about my data.
I have a daily response variable measured on 16 individuals in a GLMM where the random effect is the individual. Over 6000 days there are 6000 values.
I have several co variates only one of which is measured on the same time scale, daily temperature. The other variates are either vary on a monthly scale, or monthly x individual scale.
I have always worked on the presumption that GLMM is powerful enough to handle this, but this guy who is advising me says I have to average my response to a monthly basis, (same for the temperature co variate) so that they vary on same temporal scale. But I am sure that this is unnecessarily throwing away information that is important.
When i search for problems relating to scale in stats I get very different types of answers to do with relative values of variates so even if you might shed some light on what to call my problem and where I could find a reference that resolves my query either way.
