.-
help for ^ellip^                                     (STB-46: gr32)
.-

Confidence ellipses 
-------------------

    ^ellip^ indepvar1 [indepvar2] [^if^ exp] [^in^ range] [^,^ ^l^evel^(^#^)^
           ^g^enerate^(^newyvar newxvar^)^ [^a^dd^(^oldyvar oldxvar^)^ 
          [^p^ool^(^#^)^]] ^fy^var^(^fmt^)^ ^fx^var^(^fmt^)^ graph_options] 


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

^ellip^ graphs confidence ellipse(s) around two estimated regression coefficients
estimates from ^fit^. If indepvar2 is not specified, it is the constant. 


Options
-------

^l^evel^(^#^)^ specifies the confidence level in percent; default is 95. 

^g^enerate^(^newyvar newxvar^)^ generates two new variables which define the 
    confidence ellipse. If the dataset has less than 400 observations, the
    gen() option increases the length of the dataset to 400 observations
    by adding missing values to the original variables.

^a^dd^(^oldyvar oldxvar^)^ must be used together with gen(). It adds the
    new ellipse to the old ellipse, as defined by oldvar1 and oldvar2,
    and displays the result which is an overlaying graph.

^p^ool^(^#^)^ must be used together with if/in, gen() and add(). It
    constructs a "locus curve" of # weighted regressions (and a table),
    where the weights are fractional importance weights on the data not
    specified by if/in; see Bartels (1996).

^fy^var^(^fmt^)^ specifies a display format, such as %1.0f or %3.2g, for yvar. 

^fx^var^(^fmt^)^ specifies a display format, such as %1.0f or %3.2g, for xvar.  

graph_options are any options allowed with graph, twoway. Defaults are: c(l)
	s(.) (or c(ll) 	s(..) if add() is not missing) t1(" ") t2 (" ")  
	l1(Estimated indepvar1)  b2(Estimated indepvar2). 


Remarks
-------

The ^ellip^ command works only immediately after ^fit^. For the pool(#) option 
you must have the gphdt programs from gr20 of STB-36 installed. If you have 
installed the parsoptp program from ip22 of STB-40, you can have titles with 
embedded parentheses. However, then the option names cannot be abbreviated 
and ellip is slower. 


Examples
--------

       . input depvar indepvar 
   
            depvar indepvar 
        1. 346 26427
        2. 293 15671
        3. 151 700
        4. end

        . fit depvar indepvar, level(50)	
        . ellip indepvar, level(50) gen(y50 x50)
        . discard
        . fit depvar indepvar
        . ellip indepvar, gen(y95 x95) add(y99 x99)


Author
------

      Anders Alexandersson
      Mississippi State University, MS
      email: andersa@@rocketmail.com


References
----------

Bartels, Larry M. 1996. "Pooling Disparate Observations." American Journal of 
     Political Science. 40:905-942.

Douglas, J. B. 1993. "Confidence Regions for Parameter Pairs." The American 
     Statistician. 47:43-45.

Montgomery, Douglas C., and Elizabeth A. Peck. 1982. Introduction to Linear   
	Regression Analysis. New York: 	John Wiley & Sons.

Press, William H., Saul A. Teukolsky, William T. Vetterling, and Brian P.    
    Flannery. 1992. 2nd ed. Numerical Recipes in C: The Art of Scientific 
    Computing. Cambridge University Press.

Theil, Henry. 1971. Principles of Econometrics. New York: John Wiley and Sons.
 

Also see
--------

    STB: gr20 (STB-34), ip22 (STB-40), gr32 (STB-46)
 Manual: ^[R] fit^
On-line: help for @fit@, @graph@, @gphdt@, @gphsave@, @level@