*==================================.
* Obtaining White's Standard Errors in SPSS.
* Gwilym Pryce,  18th May 2005. 
*==================================.


* 1. Open up your data file and save it under a new name since the following procedure will alter it. 
*You might find it easiest to create a directory in your C: drive called REGRESSION & copy you data file and syntax files into it .
*so that the following syntax will need minimal alterations.
*(you can obtain the correct syntax & directory for opening your file by clicking on File, Save As etc and selecting Paste).

GET  FILE='C:\REGRESSION\miimortalitydata2005.sav'.

* 2.  Run you OLS regression and save UNSTANDARDISED residuals as ehat:.

REGRESSION  /DEPENDENT smr_0_64  /METHOD=ENTER fem_1664 househol singlpar unemploy ythunemp claim24 highered  
  / SAVE RESID(ehat).

* 3.   Filter out missing values and save the file UNDER A DIFFERENT NAME:.
*(you can obtain the correct syntax & directory for saving your file by clicking on File, Save As etc and selecting Paste).

SELECT IF(MISSING(ehat)) = 0.
EXECUTE.

SAVE OUTFILE='C:\REGRESSION\miimortalitydata2005_TEMP1.sav'.


*4. Open the file you have just saved and run the same regression again using your original syntax to check the results are the same: .

GET FILE='C:\REGRESSION\miimortalitydata2005_TEMP1.sav'.

REGRESSION  /DEPENDENT smr_0_64  /METHOD=ENTER fem_1664 househol singlpar unemploy ythunemp claim24 highered . 

*5. Now open and run the hcreg.sps syntax :.
*(open the file, block off the whole macro by pressing Ctrl+A, and run it by pressing Ctrl+R).
*NB you only need to do this step once unless you close and reopen SPSS. i.e. the macro program will stay active so long as you don't close SPSS.

*Now close the hcreg.sps syntax  file.


*6. Now run the heteroskedasticity macro (by Andrew Hayes, Ohio State University). 
* Basic syntax of the Hayes heteroskedasticity macro: "HCREG method = m dv = yvar iv = ivlist ."
* where yvar = dependent variable, and ivlist = list of independent variables .
* and where m is an integer between 0 and 3 which tells SPSS whether to run HC0, HC1, HC2 or HC3. 

* For example, to obtain HC0 type and run the following:.

HCREG method = 0 dv = smr_0_64    iv = fem_1664 househol singlpar unemploy ythunemp claim24 highered  .


* to obtain HC1 type and run the following:.

HCREG method = 1 dv = smr_0_64    iv = fem_1664 househol singlpar unemploy ythunemp claim24 highered  .


* to obtain HC2 type and run the following:.

HCREG method = 2 dv = smr_0_64    iv = fem_1664 househol singlpar unemploy ythunemp claim24 highered  .


* to obtain HC3 type and run the following:.

HCREG method = 3 dv = smr_0_64    iv = fem_1664 househol singlpar unemploy ythunemp claim24 highered  .