Hi SAS community, I'm looking to find resources on how SAS takes into account survey weights within proc severity and also how weights can be accommodated with other methods. For instance, I refer to Dr Wicklin's blog post to estimate pareto distributions with nlmixed (see below), but it is unclear to me how this could be adjusted to factor in survey weights in the correct way. Thanks in advance for any help
proc sql;
select Min(x) into :minX from Pareto; /* store x_m in macro */
run;
proc nlmixed data=Pareto;
parms x_m 1 a 2; * initial values of parameters;
bounds 0 < x_m < &minX, a > 0; * bounds on parameters;
LL = log(a) + a*log(x_m) - (a+1)*log(x); * or use logpdf("Pareto", x, a, x_m);
model x ~ general(LL);
run;
