help ladder, help gladder, dialogs: ladder gladder
help qladder qladder
-------------------------------------------------------------------------------
Title
[R] ladder -- Ladder of powers
Syntax
Ladder of powers
ladder varname [if] [in] [, generate(newvar) noadjust]
Ladder-of-powers histograms
gladder varname [if] [in] [, histogram_options combine_options]
Ladder-of-powers quantile-normal plots
qladder varname [if] [in] [, qnorm_options combine_options]
by is allowed with ladder; see [D] by.
Menu
ladder
Statistics > Summaries, tables, and tests > Distributional plots and
tests > Ladder of powers
gladder
Statistics > Summaries, tables, and tests > Distributional plots and
tests > Ladder-of-powers histograms
qladder
Statistics > Summaries, tables, and tests > Distributional plots and
tests > Ladder-of-powers quantile-normal plots
Description
ladder searches a subset of the ladder of powers (Tukey 1977) for a
transform that converts varname into a normally distributed variable.
sktest tests for normality; see [R] sktest. Also see [R] boxcox.
gladder displays nine histograms of transforms of varname according to
the ladder of powers. gladder is useful pedagogically, but we do not
advise looking at histograms for research work; ladder or qnorm (see [R]
diagnostic plots) is preferred.
qladder displays the quantiles of transforms of varname according to the
ladder of powers against the quantiles of a normal distribution.
Options for ladder
+------+
----+ Main +-------------------------------------------------------------
generate(newvar) saves the transformed values corresponding to the
minimum chi-squared value from the table. We do not recommend using
generate() because it is literal in interpreting the minimum, thus
ignoring nearly equal but perhaps more interpretable transforms.
noadjust is the noadjust option to sktest; see [R] sktest.
Options for gladder
histogram_options affect the rendition of the histograms across all
relevant transformations; see [R] histogram. Here the normal option
is assumed, so you must supply the nonormal option to suppress the
overlaid normal density. Also, gladder does not allow the width(#)
option of histogram.
combine_options are any of the options documented in [G] graph combine.
These include options for titling the graph (see [G] title_options)
and for saving the graph to disk (see [G] saving_option).
Options for qladder
qnorm_options affect the rendition of the quantile-normal plots across
all relevant transformations. See options2 in [R] diagnostic plots.
combine_options are any of the options documented in [G] graph combine.
These include options for titling the graph (see [G] title_options)
and for saving the graph to disk (see [G] saving_option).
Examples
. sysuse citytemp
. ladder tempjuly
. gladder tempjuly, l1title("") ylabel(none) xlabel(none)
. qladder tempjuly, ylabel(none) xlabel(none)
Saved results
ladder saves the following in r():
Scalars
r(N) number of observations
r(invcube) chi-squared for inverse-cubic transformation
r(P_invcube) significance level for inverse-cubic transformation
r(invsq) chi-squared for inverse-square transformation
r(P_invsq) significance level for inverse-square
transformation
r(inv) chi-squared for inverse transformation
r(P_inv) significance level for inverse transformation
r(invsqrt) chi-squared for inverse-root transformation
r(P_invsqrt) significance level for inverse-root transformation
r(log) chi-squared for log transformation
r(P_log) significance level for log transformation
r(sqrt) chi-squared for square-root transformation
r(P_sqrt) significance level for square-root transformation
r(ident) chi-squared for untransformed data
r(P_ident) significance level for untransformed data
r(square) chi-squared for square transformation
r(P_square) significance level for square transformation
r(cube) chi-squared for cubic transformation
r(P_cube) significance level for cubic transformation
Reference
Tukey, J. W. 1977. Exploratory Data Analysis. Reading, MA:
Addison-Wesley.
Also see
Manual: [R] ladder
Help: [R] diagnostic plots, [R] lnskew0, [R] lv, [R] sktest
