Stat Computing > SAS > Code

SAS Code Fragments
Getting Robust Standard Errors for OLS regression parameters

One way of getting robust standard errors for OLS regression parameter estimates in SAS is via proc surveyreg. Here are two examples using hsb2.sas7bdat.

proc reg data = hsb2;
   model write = female math;
run;
quit;                        Parameter Estimates

                     Parameter       Standard
Variable     DF       Estimate          Error    t Value    Pr > |t|

Intercept     1       16.61374        2.90896       5.71      <.0001
FEMALE        1        5.21838        0.99751       5.23      <.0001
MATH          1        0.63287        0.05315      11.91      <.0001

Example 1

proc surveyreg data = hsb2;
   cluster ses;
   model write = female math;
run;
quit;

Estimated Regression Coefficients

                             Standard
Parameter      Estimate         Error    t Value    Pr > |t|

Intercept    16.6137389    1.06324657      15.63      0.0041
FEMALE        5.2183771    0.78520124       6.65      0.0219
MATH          0.6328663    0.03540689      17.87      0.0031

NOTE: The denominator degrees of freedom for the t tests is 2.

Example 2

    If we only want robust standard error, we can specify the cluster variable to be the identifier variable.

proc surveyreg data = hsb2;
   cluster id;
   model write = female math;
run;
quit;

             Estimated Regression Coefficients

                             Standard
Parameter      Estimate         Error    t Value    Pr > |t|

Intercept    16.6137389    2.69631975       6.16      <.0001
FEMALE        5.2183771    1.02236809       5.10      <.0001
MATH          0.6328663    0.04627457      13.68      <.0001

NOTE: The denominator degrees of freedom for the t tests is 199

UCLA Researchers are invited to our Statistical Consulting Services
We recommend others to our list of Other Resources for Statistical Computing Help
These pages are Copyrighted (c) by UCLA Academic Technology Services