.-
help for ^dfregcv^
.-

DF Analyses with Double-Entry Twin Data: Asymptotic Standard
Errors and Efficient Estimation
-------------------------------------------------------------

      ^dfregcv^ depvar1 depvar2 [^if^ exp] [^in^ range], 
		^twpairv(^varname^)^ ^twentry(^varname^)^ 
                ^mzdummy(^varname^)^ ^rdz(^#^)
               [^dfmeth(^string^)^ ^covars(^varlist^)^ keepvar]

^dfregcv^ shares the features of all estimation commands; see help
@est@ (except, @predict@ is not implemented).



Description
-----------

^dfregcv^ performs DF analyses (DeFries and Fulker, 1985) with double
entry twin data, and (a) estimates the correct asymptotic standard
errors, (b) allows the combination of DF analyses with variables
describing observed differences in non-shared environments, and (c)
implements the efficient DF analysis via GMM estimation.

The data set needs to be in a "wide" form so that the information for
a twin pair is in one line of the dataset. Moreover, variables that
contain individual-specific information (e.g., education, weight,
etc.) need to be indicated with a suffix 1 or 2 for twin 1 and 2 in a
pair. Moreover, the data need to be double-entry, and each twin pair
thus needs to be included a second time with the twin assignment
reversed.

The variables ^depvar1^ and ^depvar2^ contain the phenotype of
interest.

See @http://user.demogr.mpg.de/kohler@ for further examples.

Options for ^dfregcv^
--------------------

^twpairv(^varname^)^ specifies an unique identifier for each twin pair.

^twentry(^varname^)^ specifies whether a specific line in a dataset is 
        the first or second entry of a twin pair (i.e., the variable
        specified by ^twentry(^varname^)^ must either equal 1 or 2).

^mzdummy(^varname^)^ specifies a dummy for MZ twins, i.e., it
        specifies a variable that equals 1 for MZ and 0 for DZ twins.

^rdz(^#^) specifies the genetic relatedness of DZ twins (which equals
        .5 in standard models with additive genetic heritability).

^dfmeth(^string^)^ specifies the method of estimation. ^varest^
        specifies that the asymptotically correct variance-covariance
        matrix of DF-analysis is estimated (Result 1 in Kohler and
        Rodgers, 2001). ^gmm^ specifies efficient DF estimation
        (Result 2 in Kohler and Rodgers, 2001). ^times2^ specifies a
        degrees of freedom adjustment (i.e., multiplication of
        standard errors with sqrt(2)). ^noadjust^ specifies that no
        adjustment for double entry twin data is taken.

^keepvar^ specifies that various variables generated by the program
        dfregcv (all these variables are named DF*) are not deleted at
        the end of the program. This is useful for parameter tests or
        for further explorations of the data.


Examples
--------

     * estimation of correct standard errors in DF analysis
     .^dfregcv lbmis1 lbmis2, twpairv(twpair)^ 
		            ^twentry(twentry) mzdummy(mono)^ 
		            ^rdz(.5)^
		            ^dfmeth(varest)^

     * inclusion of additional variables describing differences
     * in non-shared environment
     .^dfregcv lbmis1 lbmis2, twpairv(twpair)^ 
		            ^twentry(twentry) mzdummy(mono)^ 
		            ^rdz(.5)^
		            ^dfmeth(varest)^
                            ^covars(ed_evsm ed_amsm)^

     * efficient estimation of the above model
     .^dfregcv lbmis1 lbmis2, twpairv(twpair)^ 
		            ^twentry(twentry) mzdummy(mono)^ 
		            ^rdz(.5)^
		            ^dfmeth(gmm)^
                            ^covars(ed_evsm ed_amsm)^


     Note: in the above example, the variables ^lbmis1^ and ^lbmis2^ are
     the logarithm of the body-mass-index for twin 1 and 2,
     ^twentry^ is an identifier of twin pairs, ^twentry^ indicates the
     first and second entry of a twin pair and ^ed_evsm^ and ^ed_amsm^
     describe differences in the smoking behavior between twins (see
     @http://user.demogr.mpg.de/kohler@ for further details). 


References
----------

DeFries, J. C. and Fulker, D. W. (1985). Multiple regression analysis
of twin data. Behavior Genetics 15(5):467-73. 

Kohler, H.-P. and J. L. Rodgers (2001) DF Analyses with Double-Entry
Twin Data: Asymptotic Standard Errors and Efficient
Estimation. Behavior Genetics, forthcoming.



Author
------

    Hans-Peter Kohler
    Max Planck Institute for Demographic Research
    email: kohler@@demogr.mpg.de
    @http://user.demogr.mpg.de/kohler@