How do you approach analyzing experimental results when you suspect there is a strong seasonal bias in the data?