I've calculated genetic and environmental cross-trait correlations in a bivariate model, based on the correlational approach for scalar and non-scalar sex limitations (Neale et al., 2006, Twin Research and Human Genetics).
I've tested their significance by looking at the CIs and by checking what happens when I drop each parameter.
Now I have the situation where zero lies within the CIs - but when I drop the parameter, the model changes significantly. How is that possible (--> output attached!)?
(For a different phenotype, this didn't happen.)
Any help would be very much appreciated!