svy_nbreg.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}
{* 29mar2005}{...}
{cmd:help svy: nbreg}{...}
{right:dialog:  {bf:{dialog svy_nbreg}}{space 16}}
{cmd:help svy: gnbreg}{...}
{right:{bf:{dialog svy_gnbreg}}{space 15}}
{right:also see:  {help "svy: nbreg postestimation"}}
{hline}

{title:Title}

{p2colset 5 25 27 2}{...}
{p2col :{hi:[SVY] svy: nbreg} {hline 2}}Negative
binomial regression for survey data{p_end}
{p2colreset}{...}


{title:Syntax}

{pstd}
Negative binomial regression

{p 8 42 2}
{cmd:svy} [{it:vcetype}]
	[{cmd:,} {it:svy_options}]
{cmd::} {cmd:nbreg} {depvar} [{indepvars}] {ifin}{break}
	[{cmd:,} {it:{help "svy: nbreg##nbreg_options":nbreg_options}}]


{pstd}
Generalized negative binomial regression

{p 8 42 2}
{cmd:svy} [{it:vcetype}]
	[{cmd:,} {it:svy_options}]
{cmd::} {cmd:gnbreg} {depvar} [{indepvars}] {ifin}
	[{cmd:,} {it:{help "svy: nbreg##gnbreg_options":gnbreg_options}}]


{synoptset 25 tabbed}{...}
INCLUDE help svy_optable


{marker nbreg_options}{...}
{synopthdr:nbreg_options}
{synoptline}
{syntab:Model}
{synopt :{opt nocon:stant}}suppress the constant term{p_end}
{synopt :{opt d:ispersion}{cmd:(}{opt m:ean}{cmd:)}}parameterization
	of dispersion; {cmd:dispersion(mean)} is the default{p_end}
{synopt :{opt d:ispersion}{cmd:(}{opt c:onstant}{cmd:)}}constant
	dispersion for all observations{p_end}
{synopt :{opt e:xposure(varname_e)}}include
	ln({it:varname_e}) in model with coefficient constrained to 1{p_end}
{synopt :{opt off:set(varname_o)}}include
	{it:varname_o} in model with coefficient constrained to 1{p_end}
{synopt :{opt const:raints(constraints)}}apply
	specified linear constraints{p_end}

{syntab:Reporting}
{synopt :{opt ir:r}}report incidence-rate ratios{p_end}

{syntab:Max options}
{synopt :{it:{help maximize:maximize_options}}}control
	the maximization process; seldom used{p_end}
{synoptline}


{marker gnbreg_options}{...}
{synopthdr:gnbreg_options}
{synoptline}
{syntab:Model:}
{synopt :{opt nocon:stant}}suppress the constant term{p_end}
{synopt :{opt lna:lpha(varlist)}}dispersion model variables{p_end}
{synopt :{opt e:xposure(varname_e)}}include
	ln({it:varname_e}) in model with coefficient constrained to 1{p_end}
{synopt :{opt off:set(varname_o)}}include
	{it:varname_o} in model with coefficient constrained to 1{p_end}
{synopt :{opt const:raints(constraints)}}apply
	specified linear constraints{p_end}

{syntab:Reporting}
{synopt :{opt ir:r}}report incidence-rate ratios{p_end}

{syntab:Max options}
{synopt :{it:{help maximize:maximize_options}}}control
	the maximization process; seldom used{p_end}
{synoptline}
{p2colreset}{...}


{title:Description}

{pstd}
{cmd:svy: nbreg} fits negative binomial (Poisson with overdispersion)
regression models for complex survey data;
see {helpb nbreg} for a description of this model involving nonsurvey data,
and {helpb "svy: poisson"} for Poisson regression.

{pstd}
{cmd:svy: gnbreg} fits generalized negative binomial regression models for
complex survey data;
see {helpb gnbreg} for this model involving nonsurvey data.

{pstd}
{cmd:svy: gnbreg} allows the shape parameter alpha to be estimated as a
function of other variables.  The parameterization of alpha is the same as
that of the default parameterization of {cmd:svy: nbreg}.


{title:Options for svy}

{phang}
{it:svy_options}; see {helpb svy}.


{title:Options for nbreg}

{dlgtab:Model}

{phang}
{cmd:dispersion(mean}|{cmd:constant)}
specifies the parameterization for dispersion in the model.
The default is {cmd:dispersion(mean}}.

{phang2}
{cmd:dispersion(mean)}
specifies the model with dispersion observation equal to 1+alpha exp(xb +
offset); that is, the dispersion is a function of the expected mean of the
count for a given observation: exp(xb + offset).

{phang2}
{cmd:dispersion(constant)}
has dispersion equal to 1 + delta; that is, it is a constant for all
observations.

{pmore}
For the default model, alpha = 0 corresponds to dispersion = 1, and, thus, it
is simply a Poisson model.  Likewise, for the alternative parameterization,
delta = 0 corresponds to dispersion = 1, and it is simply a Poisson model.

{phang}
{opt noconstant},
{opt exposure(varname_e)},
{opt offset(varname_o)},
{opt constraints(constraints)};
see {help estimation options}.

{dlgtab:Reporting}

{phang}
{opt irr}; see {it:{help eform_option}}.

INCLUDE help svy_ml_maximize_options


{title:Options for gnbreg}

{dlgtab:Model}

{phang}
{opt noconstant}; see {help estimation options}.

{phang}
{opt lnalpha(varlist)} allows specifying a linear equation for ln(alpha).
Specifying {cmd:lnalpha(male old)} would specify ln(alpha) = a0 + a1*male +
a2*old where a0, a1, and a2 are parameters to be estimated along with the
other model coefficients.
If this option is
not specified, {cmd:svy: gnbreg} and {cmd:svy: nbreg} will produce the same
results because the shape parameter will be parameterized as a constant.

{phang}
{opt exposure(varname_e)},
{opt offset(varname_o)},
{opt constraints(constraints)};
see {help estimation options}.

{dlgtab:Reporting}

{phang}
{opt irr}; see {it:{help eform_option}}.

{dlgtab:Max options}

{phang}
{it:maximize_options} control the maximization process; see
{help maximize}.
If the maximization is taking a long time, you may wish to
specify the {cmd:log} option to view the iteration log.
The {cmd:log} option implies {cmd:svy}'s {cmd:noisily} option.
You should never have to specify any of the other {it:maximize_options}.

{title:Examples}

{phang}
{cmd:. webuse alq99_00}
{p_end}
{phang}
{cmd:. gen drinkdays = alq120q if alq120q <= 365}
{p_end}
{phang}
{cmd:. xi: svy: nbreg drinkdays ridageyr i.ridreth1, irr}
{p_end}


{title:Also see}

{psee}
Manual:  {bf:[SVY] svy: nbreg}

{psee}
Online:  {help "svy: nbreg postestimation"};{break}
{helpb nbreg},
{helpb "svy: poisson"},
{helpb "svy: regress"}
{p_end}