Closed amstilp closed 2 years ago
I am still testing this, but pushed the branch and started a pull review for commenting.
After running Score.SPA
tests using one of my glm models with the new code, it looks much better - assoc test statistics have a much better QQ plot.
In
SPA_pval
, the fitted.values for glm models already have the expit applied, so taking the expit again causes incorrect association test statistics.