heckman_postestimation.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}
{* 31mar2005}{...}
{cmd:heckman postestimation} {right:dialogs:  {bf:{dialog heckma_p:predict}}}
                    {right:also see:  {bf:{help heckman}}}
{hline}

{title:Title}

{p2colset 5 35 37 2}{...}
{p2col :{hi:[R] heckman postestimation} {hline 2}}Postestimation tools for heckman{p_end}
{p2colreset}{...}


{title:Description}

{pstd}
The following postestimation commands are available for {cmd:heckman}:

{synoptset 13 tabbed}{...}
{p2coldent :command}description{p_end}
{synoptline}
INCLUDE help post_adjust1star
{p2coldent:+ {helpb estat}}AIC, BIC, VCE, and estimation sample summary{p_end}
INCLUDE help post_estimates
INCLUDE help post_lincom
INCLUDE help post_lrtest
INCLUDE help post_mfx
INCLUDE help post_nlcom
{synopt :{helpb heckman postestimation##predict:predict}}predictions, residuals, influence statistics, and other diagnostic measures{p_end}
INCLUDE help post_predictnl
{p2coldent:+ {helpb suest}}seemingly unrelated estimation{p_end}
INCLUDE help post_test
INCLUDE help post_testnl
{synoptline}
{p2colreset}{...}
{p 4 6 2}
* {cmd:adjust} does not work with time-series operators.
{p_end}
{p 4 6 2}
+ {cmd:estat ic} and {cmd:suest} do not work after {cmd:heckman, twostep}.
{p_end}



{marker predict}{...}
{title:Syntax for predict}

{phang}
After ML or twostep

{p 8 16 2}
{cmd:predict}
{dtype}
{newvar} {ifin} [{cmd:,}
{it:statistic} {opt nooff:set}]

{phang}
After ML

{p 8 16 2}
{cmd:predict}
{dtype} {c -(}{it:stub*}{c |}{it:newvar_reg}
{it:newvar_sel}
{it:newvar_athrho}
{it:newvar_lnsigma}{c )-}
{ifin}
{cmd:,} {opt sc:ores}
 
{synoptset 21 tabbed}{...}
{synopthdr :statistic}
{synoptline}
{syntab :Main}
{synopt :{opt xb}}xb, fitted values (the default){p_end}
{synopt :{opt yc:ond}}E(y {c |} y observed){p_end}
{synopt :{opt ye:xpected}}E(y*), y taken to be 0 where unobserved{p_end}
{synopt :{opt ns:hazard} or {opt m:ills}}nonselection hazard (also called inverse of Mills' ratio{p_end}
{synopt :{opt ps:el}}P(y observed){p_end}
{synopt :{opt xbs:el}}linear prediction for selection equation{p_end}
{synopt :{opt stdps:el}}standard error of the linear prediction for selection equation{p_end}
{synopt :{opt p:r(a,b)}}Pr(y {c |} {it:a} < y < {it:b}){p_end}
{synopt :{opt e(a,b)}}E(y {c |} {it:a} < y < {it:b}){p_end}
{synopt :{opt ys:tar(a,b)}}E(y*), y* = max{c -(}{it:a},min(y,{it:b}){c )-}{p_end}
{synopt :{opt stdp}}standard error of prediction{p_end}
{synopt :{opt stdf}}standard error of forecast{p_end}
{synoptline}
{p2colreset}{...}
INCLUDE help esample

INCLUDE help whereab


{title:Options for predict}

{phang}
{opt xb}, the default, calculates the linear prediction.

{phang}
{opt ycond} calculates the expected value of the dependent variable conditional on the dependent variable being observed, i.e., selected.

{phang}
{opt yexpected} calculates the expected value of the dependent variable
(y*), where that value is taken to be 0 when it is expected to be unobserved.

{pmore}
The assumption of 0 is valid for many cases where nonselection implies
nonparticipation (e.g., unobserved wage levels, insurance claims from those
who are uninsured, etc.) but may be inappropriate for some problems (e.g.,
unobserved disease incidence).

{phang}
{opt nshazard} and {cmd:mills} are synonyms; both calculate the
nonselection hazard{hline 2}what Heckman referred to as the inverse of
the Mills' ratio{hline 2}from the selection equation.

{phang}
{opt psel} calculates the probability of selection (or being observed).

{phang}
{opt xbsel} calculates the linear prediction for the selection
equation.

{phang}
{opt stdpsel} calculates the standard error of the linear prediction
for the selection equation.

{phang}
{opt pr(a,b)} calculates the {bind:Pr({it:a} < xb + u < {it:b})}, the
probability that y|x would be observed in the interval ({it:a},{it:b}).

{pmore}
{it:a} and {it:b} may be specified as numbers or variable names; lb and 
ub are variable names;{break}
{cmd:pr(20,30)} calculates {bind:Pr(20 < xb + u < 30)};{break}
{cmd:pr(lb,ub)} calculates {bind:Pr(lb < xb + u < ub)}; and{break}
{cmd:pr(20,ub)} calculates {bind:Pr(20 < xb + u < ub)}.

{pmore}
{it:a} missing {bind:({it:a} {ul:>} .)} means minus infinity;
{cmd:pr(.,30)} calculates {bind:Pr(xb + u < 30)};{break}
{cmd:pr(lb,30)} calculates {bind:Pr(xb + u < 30)} in
observations for which {bind:lb {ul:>} .}{break}
(and calculates {bind:Pr(lb < xb + u < 30)} elsewhere).

{pmore}
{it:b} missing {bind:({it:b} {ul:>} .)} means plus infinity; {cmd:pr(20,.)}
calculates {bind:Pr(xb + u > 20)}; {break}
{cmd:pr(20,ub)} calculates {bind:Pr(xb + u > 20)} in
observations for which {bind:ub {ul:>} .}{break} (and calculates
{bind:Pr(20 < xb + u < ub)} elsewhere).

{phang}
{opt e(a,b)} calculates
{bind:E(xb + u | {it:a} < xb + u < {it:b})}, the expected value of
y|x conditional on y|x being in the interval ({it:a},{it:b}), which is to
say, y|x is censored.  {it:a} and {it:b} are specified as they are for
{opt pr()}.

{phang}
{opt ystar(a,b)} calculates E(y*),
where {bind:y* = {it:a}} if {bind:xb + u {ul:<} {it:a}}, {bind:y* = {it:b}}
if {bind:xb + u {ul:>} {it:b}}, and {bind:y* = xb + u} otherwise, meaning
that y* is truncated.  {it:a} and {it:b} are specified as they are for
{opt pr()}.

{phang}
{opt stdp} calculates the standard error of the prediction, which can be
thought of as the standard error of the predicted expected value or mean for
the observation's covariate pattern.  This is also referred to as the standard
error of the fitted value.

{phang}
{opt stdf} calculates the standard error of the forecast, which is the
standard error of the point prediction for a single observation.  It is
commonly referred to as the standard error of the future or forecast value.
By construction, the standard errors produced by {opt stdf} are always larger
than those produced by {opt stdp}.

{phang}
{opt nooffset} is relevant when you specify {opth offset(varname)} for
{cmd:heckman}.  It modifies the calculations made by {cmd:predict} so that
they ignore the offset variable; the linear prediction is treated as xb rather
than xb + offset.

{phang}
{opt scores}, not available with {opt twostep}, calculates equation-level score variables.

{pmore}
The first new variable will contain the derivative of the log likelihood with
respect to the regression equation.

{pmore}
The second new variable will contain the derivative of the log likelihood with
respect to the selection equation.

{pmore}
The third new variable will contain the derivative of the log likelihood with
respect to the third equation ({hi:athrho}).

{pmore}
The fourth new variable will contain the derivative of the log likelihood with
respect to the fourth equation ({hi:lnsigma}).


{title:Examples}

{phang}{cmd:. heckman wage educ age, select(married children educ age)}

{phang}{cmd:. predict yhat}{p_end}
{phang}{cmd:. predict yhat, xb}

{phang}{cmd:. predict mystdp, stdp}{p_end}
{phang}{cmd:. predict mystdf, stdf}

{phang}{cmd:. predict ycond, ycond}{p_end}
{phang}{cmd:. predict ystar, yexpected}{p_end}
{phang}{cmd:. predict probseen, psel}{p_end}
{phang}{cmd:. predict selindex, xbsel}

{phang}{cmd:. predict mymill, mills}

{phang}{cmd:. predict p0to20, pr(0,20)}{p_end}
{phang}{cmd:. predict less15, pr(.,15)}{p_end}
{phang}{cmd:. predict ey0to20, e(0,20)}{p_end}
{phang}{cmd:. predict ys0to20, ystar(0,20)}


{title:Also see}

{psee}
Manual:  {bf:[R] heckman postestimation}

{psee}
Online:  {helpb heckman};{break}
{helpb adjust}, 
{helpb estimates}, {helpb lincom}, {helpb lrtest}, {helpb mfx},
{helpb nlcom}, {helpb predictnl}, 
{helpb suest}, {helpb test}, {helpb testnl}
{p_end}