Find normal transform from ladder of powers
-------------------------------------------

	^ladder^ varname [^if^ exp] [^in^ range] [^, g^enerate^(^newvar^)^]

searches a subset of the ladder of powers (See Tukey, 1977 or Hamilton,
1990) for the transform that converts varname into a normally-distributed 
variable.  ^sktest^ is used to test for normality.

If ^generate()^ is specified, a new variable is created that corresponds to the 
minimum chi-square value from the table.  ^generate()^ is not, in general,
recommended since it is quite literal in its interpretation of smallest, 
thus ignoring nearly equal but perhaps simpler interpretations.

References:

Tukey, John W. (1977).  ^Exploratory Data Analysis^.  Reading, MA:  Addison-
Wesley.

Hamilton, Lawrence C. (1990).  ^Statistics with Stata^.  Pacific Grove, CA:
Brooks/Cole.

Example
-------

	. ^ladder mpg, gen(mpgx)^

	Transformation         formula             Chi-sq(2)     P(Chi-sq)
	------------------------------------------------------------------
	cube                   mpg^^3                 238.53        0.000
	square                 mpg^^2                  80.88        0.000
	raw                    mpg                    18.38        0.000
	sqare-root             sqrt(mpg)               6.70        0.035
	log                    log(mpg)                1.57        0.456
	reciprocal root        1/sqrt(mpg)             0.27        0.874
	reciprocal             1/mpg                   1.83        0.400
	reciprocal square      1/(mpg^^2)              17.11        0.000
	reciprocal cube        1/(mpg^^3)              65.20        0.000

	(mpgx = 1/sqrt(mpg)   generated)

In this case, the automatic ^gen()^ option chose 1/sqrt(mpg) whereas, based on 
interpretation, we would have been tempted to select 1/mpg.

Also see ^help gladder^.