I am trying to use bootstrapping to get a 95% CI for Cohen's d, but the command produces coef. only. No bootstrapping SE, z, p-value, and 95% CI. And I don't get any error messages!
This is very strange because the same codes actually worked without any problem a few weeks ago. I had to revisit the analysis recently so just ran what I wrote before to review, but now, the codes are not working.
FYI, the codes below are a part of what I wrote. And the data is not real data. I used frame since I actually need to handle multiple data sets for this. But it produces the same issue.
Please advise!
Code:
frame reset frame create sample3 frame sample3 { * Example generated by -dataex-. For more info, type help dataex clear input byte id double(group outcome_t0 outcome_t1 outcome_t2) 44 2 84.860979 114.87961250000001 41.4992725 45 1 49.8758465 55.0360915 50.1319545 46 1 80.74328550000001 77.3736495 81.3353915 48 1 42.024735500000006 54.970216 57.763051000000004 49 2 52.415797 56.568540999999996 50.503508499999995 50 2 48.055269 48.4297725 46.5440855 58 2 39.3368345 37.401369 37.568878 60 2 50.3642245 45.998895000000005 46.2884515 61 2 71.5496255 71.77641600000001 70.61160749999999 62 2 49.8083935 44.5434755 47.758653499999994 63 1 57.545792000000006 65.44662199999999 66.445455 64 2 40.995180000000005 46.6467495 44.6673595 65 2 52.914714000000004 46.365172 50.470093 66 2 60.423458499999995 . . 71 2 50.232740500000006 46.5209385 43.5583465 end label values group group_l label def group_l 1 "S", modify label def group_l 2 "C", modify } capture program drop boot_test program boot_test, rclass version 17 frame sample3 { * getting mean and variance of the outcome at each time point for each group sum `1'_t0 if group==1 local n_grp1_t0=r(N) local dep_grp1_t0=r(mean) local sd_grp1_t0=r(sd) local var_grp1_t0=r(Var) sum `1'_t1 if group==1 local n_grp1_t1=r(N) local dep_grp1_t1=r(mean) local sd_grp1_t1=r(sd) local var_grp1_t1=r(Var) * difference between two time points local dif_grp1_t1t0=`dep_grp1_t1'-`dep_grp1_t0' * pooled sd local sd_pooled_grp1_t1t0=sqrt((`var_grp1_t1'+`var_grp1_t0')/2) * Cohen's d return scalar cohend_grp1_t1t0=`dif_grp1_t1t0'/`sd_pooled_grp1_t1t0' } end bootstrap r(cohend_grp1_t1t0), reps(1010) seed(11271980): boot_test outcome
0 Response to Bootstrapping produces coef only
Post a Comment