mp-bc

  USEVARIABLES =
       FAM_SUPT ISEL_ALL
       ECEA6 PIEA17 SEEA8
       sash_langnew SASH_SOC
       screen
       ! covariates
       HCE2 MHEA26
       AGE US_BORN
       CENTER2
       CENTER3
       CENTER4
       BK0 BK1 BK2
       BK4 BK5 BK6
     ;
     NOMINAL are
       screen;
     WEIGHT = WEIGHT_FINAL_NORM_OVERALL;
     STRATIFICATION = STRAT;
     CLUSTER = PSU_ID;
     SUBPOPULATION = flag EQ 1;

  ANALYSIS:

     TYPE = COMPLEX;
     ESTIMATOR = MLR;
     ALGORITHM = INTEGRATION;
     INTEGRATION = MONTECARLO(1000);
     PROCESSORS = 4;

  MODEL:

     Accull BY sash_langnew@1 SASH_SOC;
     Accull@1;
     sash_langnew@.001;

     FAM_SUPT ON
       ECEA6 PIEA17 SEEA8 (sesa1-sesa3)
       Accull (acca);
     ISEL_ALL ON
       ECEA6 PIEA17 SEEA8 (sesb1-sesb3)
       Accull (accb);
     FAM_SUPT WITH ISEL_ALL*;

     screen ON
       FAM_SUPT !(supb1)
       ISEL_ALL !(supb2)
       ECEA6
       PIEA17
       SEEA8
       Accull;

     ! covariates
     screen ON
       HCE2 MHEA26
       AGE US_BORN
       CENTER2
       CENTER3
       CENTER4
       BK0 BK1 BK2
       BK4 BK5 BK6
  ;

     ! deal with missing data
     [MHEA26* US_BORN* ECEA6* PIEA17* SEEA8*];

    !MODEL CONSTRAINT:
    !  new(iECEFAM iPIEFAM iSEEFAM iACCFAM
    !      iECEISE iPIEISE iSEEISE iACCISE);
    !  iECEFAM = sesa1 * supb1;
    !  iPIEFAM = sesa2 * supb1;
    !  iSEEFAM = sesa3 * supb1;
    !  iACCFAM = acca * supb1;

    !  iECEISE = sesb1 * supb2;
    !  iPIEISE = sesb2 * supb2;
    !  iSEEISE = sesb3 * supb2;
    !  iACCISE = accb * supb2;


  OUTPUT:
  STDYX; CINTERVAL;