* * * * T H E P A R K T E S T * * * *

  1. Run an OLS regression and saved the residuals (e)
  2. Square the residuals: ESQ=e2
  3. Take the natural log of ESQ: LNESQ=ln(ESQ)
  4. Take the natural log of variable which has suspected heteroskedasticity: LNX = ln(X)
  5. Run the OLS model: LNESQ = f(constant, LNX)
  6. If LNX is significant then heteroskedasticity is present
SPSS OUTPUT
                
Equation Number 1    Dependent Variable..   LNESQ

********************************************************
NOTE: The dependent variable is the natural log of the 
squared residuals
********************************************************

Block Number  1.  Method:  Enter      LNAGE

********************************************************
NOTE: The independent variable is the natural log of age 
(the variable that appears to be heteroskedastic)
********************************************************

R Square             .05292
Adjusted R Square    .05243
Standard Error      2.37106

Analysis of Variance
                    DF      Sum of Squares      Mean Square
Regression           1           613.13367        613.13367
Residual          1952         10973.96564          5.62191

F =     109.06148       Signif F =  .0000


------------------ Variables in the Equation ------------------

Variable              B        SE B       Beta         T  Sig T

LNAGE         -1.483884     .142090   -.230033   -10.443  .0000
(Constant)    17.451896     .534475               32.652  .0000

*********************************************************************
NOTE: The T-stat on LNAGE is "greater than 2" indicating that 
heteroskedasticity is present.
*********************************************************************