Hello,

I have made a multilevel model for which I want to plot the effects of one of the levels, which has 675 observations through a standard error bar chart (serrbar command). I have ordered the effects with the rank command, in addition to obtaining the standard errors. The problem is that I tried unsuccessfully to apply the "in" option so that the graph only shows the 20 best observations, but the 675 appear making the graph impossible to read. This is the code without the "in" option:



Code:
generate labheight = v0 + 1.96*v0se + 0.05
serrbar v0 v0se v0rank if pickone_Driver==1, scale(1.96) yline(0) ///
addplot(scatter labheight v0rank, ///
msymbol(none) mlabel(Driver) ///
mlabposition(1) mlabangle(vertical) mlabcolor(navy)) ///
ytitle("Predicted Driver effect") xtitle("Rank") ///
legend(off)
How should I include the option to achieve my goal? Thanks for your help