Imagine 10 people.
In year 2016, the first 5 people have a response $Y$ measured on days randomly allocated throughout the months of January to July. The days were random for each individual.
In year 2017, the other 5 people were again measured on different individual-specific random days throughout the same period.
I have similar data for years 2015, 2014, .... etc, (involving 5 other people each time).
I want to analyse the response using regression, but my main issue is, how do I account for temporal autocorrelation?
In fact, my primary issue is, how do I even model the "time" variable?
What would you do?
