Standardized Difference with PROC PSMATCH

Elisa97 — Thu, 07 Mar 2024 11:26:40 GMT

Hi everyone,

I would like to calculate Standardized Difference.

I found that I can calculate Standardized Mean Difference (SMD) with PROC PSMATCH, but it's also calculate SMD for categorical variables and I don't understand how.

I have a sample with 155 subjects, 77 in GROUPE 1 and 78 in GROUPE 2.

proc format; 
value sexe
1  = "Masculin"
2  = "Féminin"; 
value groupe
1  = "Groupe 1"
2  = "Groupe 2";
run;

proc psmatch data=table;
   class groupe sexe;
   psmodel groupe(Treated='Groupe 2')= age taille poids sexe;
   assess lps var=(age sexe taille poids) / plots=all weight=none;
run;

OUTPUT of PROC PSMATCH :

Description of SEXE's variable:

I have Standardized Mean Difference for quantitave variable (age, taille and poids) in the row ALL. But how SAS find 0.32697 for the SEXE's variable ?

Thank you.

Re: Standardized Difference with PROC PSMATCH

Mike_N — Fri, 08 Mar 2024 20:03:54 GMT

Take a look at the example here: https://go.documentation.sas.com/doc/en/pgmsascdc/v_048/statug/statug_psmatch_examples04.htm . For binary classification variables, in your case, for the SEXE variable, the computed value is the difference in the proportion of the first ordered level of the variable between treated and control groups.

topic Standardized Difference with PROC PSMATCH in Statistical Procedures

Standardized Difference with PROC PSMATCH

Re: Standardized Difference with PROC PSMATCH