.
* if B>25%, R outside [0.5; 2]
Matched 0.003 2.19 0.823 5.7 6.0 13.2 0.82 20
Unmatched 0.156 195.15 0.000 40.6 28.5 104.2* 0.93 0
Sample Ps R2 LR chi2 p>chi2 MeanBias MedBias B R %Var
* if variance ratio outside [0.78; 1.28] for U and [0.78; 1.28] for M
M -3.0106 -2.7974 -7.2 67.9 -0.86 0.391 1.39*
x5 U -3.0106 -2.3455 -22.5 -3.21 0.001 1.06
M -1.6722 -1.4857 -9.2 67.8 -1.02 0.310 0.89
x4 U -1.6722 -1.0928 -28.5 -3.98 0.000 0.93
M 1.9071 2.0272 -6.0 86.3 -0.66 0.510 1.04
x3 U 1.9071 1.0323 43.6 6.27 0.000 1.12
M -1.153 -1.2031 3.5 86.7 0.37 0.710 0.83
x2 U -1.153 -1.5301 26.0 3.65 0.000 0.96
M .01912 -.04552 2.6 96.8 0.30 0.762 1.11
x1 U .01912 2.0581 -82.2 -11.52 0.000 0.94
Variable Matched Treated Control %bias |bias| t p>|t| V(C)
Unmatched Mean %reduct t-test V(T)/
. pstest x1 x2 x3 x4 x5, treated (treat) both
The output from -pstest- lists bias in
the unmatched and matched
covariates. The standardized
difference in the means of x5, for
example, decreased from -22.5% in
the unmatched sample to -7.2% in
the matched sample
The mean and median
standardized difference for all of
the covariates are summarized at
the end of the output of this
command.