symmetry.hlp

是一个经济学管理应用软件 很难找的 但是经济学学生又必须用到

{smcl}
{* 27jan2005}{...}
{cmd:help symmetry}, {cmd:help symmi}{right:dialogs:  {bf:{dialog symmetry}}  {bf:{dialog symmi}}}
{hline}

{title:Title}

{p2colset 5 21 23 2}{...}
{p2col:{hi:[R] symmetry} {hline 2}}Symmetry and marginal homogeneity tests
{p_end}
{p2colreset}{...}


{title:Syntax}

{pstd}
Symmetry and marginal homogeneity tests

{p 8 17 2}
{cmd:symmetry} {it:casesvar} {it:controlvar}
{ifin} {weight} [{cmd:,} {it:options}]

{pstd}
Immediate form of symmetry and marginal homogeneity tests

{p 8 17 2}
{cmd:symmi} {it:#}11 {it:#}12 [{it:...}]{cmd:\} {it:#}21 {it:#}22
	[{it:...}] [{cmd:\}{it:...}]
{ifin} [{cmd:,} {it:options}]

{synoptset 11 tabbed}{...}
{synopthdr}
{synoptline}
{syntab:Main}
{synopt:{opt not:able}}suppress output of contingency table
{p_end}
{synopt:{opt con:trib}}report contribution of each off-diagonal cell pair
{p_end}
{synopt:{opt ex:act}}perform exact test of table symmetry
{p_end}
{synopt:{opt mh}}perform two marginal homogeneity tests
{p_end}
{synopt:{opt tr:end}}perform a test for linear trend in the (log)
relative risk (RR)
{p_end}
{synopt:{opt cc}}use continuity correction when calculating test for
linear trend
{p_end}
{synoptline}
{p2colreset}{...}
{p 4 6 2}
{opt fweight}s are allowed; see {help weight}.
{p_end}


{title:Description}

{pstd}
{opt symmetry} performs asymptotic symmetry and marginal homogeneity tests,
and an exact symmetry test on KxK tables where there is a 1-to-1 matching of
cases and controls (nonindependence). This is used to analyze
matched-pair case-control data with multiple discrete levels of the exposure
(outcome) variable. In genetics, the test is known as the
Transmission/Disequilibrium test (TDT) and is used to test the association
between transmitted and nontransmitted parental marker alleles to an affected
child. In the case of 2x2 tables, the asymptotic test statistics reduce to the
McNemar test statistic, and the exact symmetry test produces an exact McNemar
test; see {help epitab}.  For numerous exposure variables, {opt symmetry}
can optionally perform a test for linear trend in the log relative risk.

{pstd}
{opt symmetry} expects the data to be in wide format; that is, each
observation contains the matched case and control values in variables
{it:casesvar} and {it:controlvar}.  Variables can be numeric or string.

{pstd}
{opt symmi} is the immediate form of {opt symmetry}.  The {opt symmi} command
uses the values specified on the command line; rows are separated by '\', and
options are the same as for {opt symmetry}.
See {help immed} for a general introduction to immediate commands.

{title:Options}

{dlgtab:Main}

{phang}
{opt notable} suppresses the output of the contingency table.
By default, {opt symmetry} displays the n x n contingency table at the top of
the output.

{phang}
{opt contrib} reports the contribution of each off-diagonal cell pair
to the overall symmetry chi-squared.

{phang}
{opt exact} performs an exact test of table symmetry.
This option is recommended for sparse tables.  CAUTION:  the
exact test requires substantial amounts of time and memory for large tables.

{phang}
{opt mh} performs two marginal homogeneity tests that do not require
the inversion of the variance-covariance matrix.

{pmore}
By default, {opt symmetry} produces the Stuart-Maxwell test statistic, which
requires the inversion of the nondiagonal variance-covariance test matrix, V.
When the table is sparse, the matrix may not be full rank, and, in that case,
the command substitutes a generalized inverse V* for V^(-1).
{opt mh} calculates optional marginal homogeneity statistics that do not require
the inversion of the variance-covariance matrix.  These tests may be preferred
in certain situations.

{phang}
{opt trend} performs a test for linear trend in the (log) relative risk
(RR).  This option is only allowed for numeric exposure (outcome) variables,
and its use should be restricted to measurements on the ordinal or interval
scales.

{phang}
{opt cc} specifies that the continuity correction be used when
calculating the test for linear trend. This correction should only be
specified when the levels of the exposure variable are equally spaced.


{title:Examples}

{phang}{cmd:. symmetry cases controls}{p_end}
{phang}{cmd:. symmetry cases controls, not}{p_end}
{phang}{cmd:. symmetry cases controls, contrib mh}{p_end}
{phang}{cmd:. symmetry cases controls, mh ex}{p_end}
{phang}{cmd:. symmetry cases controls, ex  contrib}{p_end}
{phang}{cmd:. symmetry cases controls, trend}{p_end}
{phang}{cmd:. symmetry cases controls, trend cc}

{p 4 8 2}{cmd:. symmi 47 56 38 \ 28 61 31 \ 26 47 10 , contrib}


{title:Also see}

{psee}
Manual:  {bf:[R] symmetry}

{psee}
Online:  {help epitab}
{p_end}