Confidence Interval graphs Logit models

[Tamanna Maqbool Shah]

I am trying to create confidence interval graphs using logit and melogit commands. The melogit command gives me correct graphs but the logit command does not. Pictures are attached after each command. Can someone please help me figure out what I am doing wrong? Thank you.

** Fixed effects ***

logit conflict rel_heg rel_heg2 i.num_cntry
predict double x_out3, xb
predict double stdp3, stdp
generate double lb3 = x_out3 + stdp * invnormal(0.025)
generate double ub3 = x_out3 + stdp * invnormal(0.975)
graph twoway ///
line x_out3 rel_heg, sort lcolor(black) lpattern(solid) || ///
line lb3 rel_heg, sort lcolor(black) lpattern(dash) || ///
line ub3 rel_heg, sort lcolor(black) lpattern(dash) ///
ylabel( , angle(horizontal) nogrid) legend(off)

** Logistic "fixed effects model" *

logit conflict r_pol2 r_pol2_sq i.num_cntry
predict double x2_out, xb
predict double stdp2, stdp
generate double lb2 = x2_out + stdp2 * invnormal(0.025)
generate double ub2 = x2_out + stdp2 * invnormal(0.975)
graph twoway ///
line x2_out r_pol2, sort lcolor(black) lpattern(solid) || ///
line lb2 r_pol2, sort lcolor(black) lpattern(dash) || ///
line ub2 r_pol2, sort lcolor(black) lpattern(dash) ///
ylabel( , angle(horizontal) nogrid) legend(off) saving(polfixed)

Command I use to create confidence interval using melogit

** Mixed Effects **

melogit conflict rel_heg rel_heg2 || country:
predict double x_out4, xb
predict double stdp5, stdp
generate double lb4 = x_out4 + stdp * invnormal(0.025)
generate double ub4 = x_out4 + stdp * invnormal(0.975)
graph twoway ///
line x_out4 rel_heg, sort lcolor(black) lpattern(solid) || ///
line lb4 rel_heg, sort lcolor(black) lpattern(dash) || ///
line ub4 rel_heg, sort lcolor(black) lpattern(dash) ///
ylabel( , angle(horizontal) nogrid) legend(off) saving(relhegmixed)