I am investigating diversity-stability relationships and I am trying to test for the portfolio effect and insurance effect with my data. I understand that overyielding indicates the portfolio effect is occurring and that overyielding occurs when 'mean-variance scaling' is > 1.
1) how do I calculate the mean-variance scaling? is it just the slope of the correlation of log(mean) : log(variance) for all data points in my repeated measures?
2) How do I measure test the insurance effect?