Hello all, I've used perused this site numerous times in the past for assistance with statistical issues, but this is my first post.

I work with fish, and am specifically interested in the environmental effects on these fish. I ran regressions of growth increment (growth experienced in a given year) vs. duration of a specific temperature seen in that same year of growth. My goal is to determine if duration of a specific high temperature may affect fish growth.

These were not long lived fish, so I only have a limited number of years to work with...

For example...

Growth year Duration of temps at C Average growth increment
2011 __ days __ inches
2010 __ days __ inches
2009 __ days __ inches
2008 __ days __ inches
2007 __ days __ inches
2006 __ days __ inches
2005 __ days __ inches

So... I took the average growth increments (averaged from a total of 200 fish), and regressed the average growth that they experienced in a given year against temperature variables seen in those same years. Is there a situation here where some sort of power analysis would allow me to determine if the sample size is sufficient? I can use hundreds of fish to determine the growth increment variables, but the regressions themselves are really only going to have 6-8 data points, representing the 6-8 years of growth that I'm studying. Any assistance with how I can support this sample size, or any information on holes in this study design related to sample size that need to be fixed would be greatly appreciated.