I am using Stata 15.1. I ran a GLM in svy, and then used parmest to generate an output dataset with a p-value variable (I installed the Stata 11 version, as the latest parmest program is for Stata 16.0). I then used qqvalue to obtain p-values adjusted for multiple comparisons using the Simes method. However, my q-value for my independent variable (lgb) is equal to its p-value. I've ran multiple GLMs, both adjusted and unadjusted, on different outcome variables and I keep having the same issue. Does anyone have an idea why this might happen?
Here is my code and output:
Code:
svy: glm q1_bin2 i.lgb pscore, family(binomial) link(identity) parmest, list(,)saving("C:\filepath") use "C:\filepath" qqvalue p, method(simes) qvalue(simes) list +--------------------------------------------------------------------------------------------------------------+ | eq parm estimate stderr dof t p min95 max95 simes | |--------------------------------------------------------------------------------------------------------------| 1. | q1_bin2 0b.lgb 0 0 8480 . . 0 0 . | 2. | q1_bin2 1.lgb .01474893 .0208141 8480 .70860288 .47859043 -.02605178 .05554964 .47859043 | 3. | q1_bin2 pscore .10122122 .12771139 8480 .79257784 .428046 -.14912424 .35156667 .47859043 | 4. | q1_bin2 _cons .17528199 .00853499 8480 20.53687 1.656e-91 .15855133 .19201265 4.967e-91 | +--------------------------------------------------------------------------------------------------------------+
clear
input str7 eq str6 parm double(estimate stderr) int dof double(t p min95 max95 simes)
"q1_bin2" "0b.lgb" 0 0 8480 . . 0 0 .
"q1_bin2" "1.lgb" .014748931015701737 .02081409982986773 8480 .7086028767161661 .47859042879152147 -.026051778561407606 .05554964059281108 .47859042879152147
"q1_bin2" "pscore" .10122121527813703 .12771138765846995 8480 .7925778361192517 .4280459983926721 -.14912423708995476 .35156666764622885 .47859042879152147
"q1_bin2" "_cons" .17528199292272936 .008534990445583212 8480 20.536870432401756 1.6556059648229733e-91 .15855133105120625 .19201265479425247 4.96681789446892e-91
end
Thank you.
0 Response to Qqvalue: q-values the same as p-values
Post a Comment