Calculating Median Titers for Vaccines Antibodies

[Rehman Najeeb]

I have a data of antibody counts on three different time points (below is attached data).
I have calculated Geometric Means for that now I want to calculate median titer and their CI's.

Should I calculate it as it is or should I transform this into log form or something?
Is there any package to do that?

* Example generated by -dataex-. To install: ssc    install    dataex
clear
input double(b_v10 f_v8 f_v9)
2.008  719.492       .
1.429  813.835       .
1.329  948.829       .
3.969   771.26       .
1.621 1059.113       .
2.821  407.136       .
.528  641.715       .
1.329  816.694       .
2.831 1037.571       .
5.821        . 976.257
1.06        . 762.986
1.167        . 777.379
.616        . 956.089
1.476        . 469.547
.432        . 203.753
.512        . 497.113
.528        . 878.459
.652        . 828.762
4.914  307.819       .
10.104   258.58       .
end
label var b_v10 "Concentration (U/ml)_0day" 
label var f_v8 "Concentration (U/ml)_14Weeks" 
label var f_v9 "Concentration (U/ml)_26Weeks"

[Phil Bromiley]

You didn't get a quick answer. You'll increase your chances of helpful answer by following the FAQ on asking questions - provide Stata code in code delimiters, readable Stata output, and sample data (which you did). Extra stuff like the labels does not help.

You need to remember that we are almost all not from your area of research. So, if you can explain what you really mean in standard statistical terminology, we can probably help.

If you want medians, egen will calculate it and allow you to do it by whatever sub-set of the data you want. Alternatively, summary will give you the 50th percentile (saved as r(p50)). The CI's is trickier. See https://www.stata.com/statalist/arch.../msg00348.html
https://www.statalist.org/forums/for...licate-weights