{smcl}
{* 28 October 2004}{...}
{hline}
help for {hi:margin}{right:(This version: 05 October 2004)}
{hline}
{title:Average marginal effects for categorical and limited dependent variable models}
{p 8 17 2}
{cmdab:margin}
[{cmd:,}
{cmdab:c:ount}
{cmdab:d:ummies(}{it:Varlist_1} [ \ {it:Varlist_2} \ ... \ {it:Varlist_N}]{cmd:)}
{cmdab:e:form}
{cmdab:hascons:}
{cmdab:mean}
{cmdab:m:odel(}{it:stata_cmd}{cmd:)}
{cmdab:p:ercent}
{cmdab:r:eplace}
{cmdab:t:able} ]
{p 4 4 2} where stata_cmd is one of
{col 10}[class 1:] {col 20}{help probit} {col 30}{help logit} {col 40}{help logistic} {col 50}{help cloglog} {col 60}{help xtprobit} {col 70}{help xtlogit}
{col 10}[class 2:] {col 20}{help oprobit} {col 30}{help ologit} {col 40}{help gologit} {col 50}{help mlogit} {col 60}{help biprobit}
{col 10}[class 3:] {col 20}{help poisson} {col 30}{help nbreg} {col 40}{help zip} {col 50}{help zinb}
{col 10}[class 4:] {col 20}{help tobit} {col 30}{help cnreg} {col 40}{help intreg} {col 50}{help heckman} {col 60}{help heckprob}
{p 4 4 2} and {it:Varlist_K} includes dummy variables where all of the dummies indicate
different categories of the same underlying categorical variable.
{title:Description}
{p 4 4 2}
{cmd:margin} calculates marginal effects, and standard errors for marginal effects
using the delta method. By default, {cmd:margin} calculates the average of partial and discrete
changes over the observations. {cmd:margin} can also compute marginal effects at the sample means of the variables if
the {cmd:mean} option is specified.
In either cases, calculations are restricted to the estimation sample.
{p 4 4 2}
By default, the marginal effects calculated are as follows:
{p 8 19 4}[class 1:] changes in the probability of positive outcome {p_end}
{p 8 19 4}[class 2:] changes in the probabilities of all possible outcomes defined by the dependent variable(s){p_end}
{p 8 19 4}[class 3:] changes in the expected number of counts or in the incidence rate (for {cmd:zip} and {cmd:zinb}, changes in {it:E(y|y>0)} and {it:Pr(y>0)} also displayed) {p_end}
{p 8 19 4}[class 4:] changes in the expected value of the dependent variable {it:conditional} on the dependent variable being observed (not censored) {p_end}
{p 4 4 2}
However, if the eform option is specified, {cmd:margin} calculates the changes in the value
of {it:exp(xb)}, where {it:xb} is the linear prediction.
{p 4 4 2}
{cmd:margin} can be also be used after commands that are not on the list above. For
example, users may wish to obtain marginal effects after the survey version of
{it:stata_cmd}, or after their own program. This is made possible by the {cmd:model()}
option.
{p 4 4 2}
{cmd:margin} can be used both as an estimation command and as a post-estimation command;
see help {help est} and {help postest}. By default, {cmd:margin} behaves as a post-estimation command,
i.e. it does not affect the estimation results. However, option {cmd:replace} forces {cmd:margin}
to behave as an estimation command. This enables the use of post-estimation commands like
{cmd:lincom} or {cmd:test} after {cmd:margin}.
{p 4 4 2}
Typed without arguments, {cmd:margin} redisplays the marginal effects.
{title:Options}
{p 4 8 2}
{cmd:count} modifies the calculation of marginal effects for count variables, i.e.
variables that take more than two values and all of the values are integers.
By default, {cmd:margin} treates count variables as continuous variables, thus
marginal effects correspont to small changes in the independent variables.
If the count option is specified, the marginal effects are changes in
probabilities when the count variables increases by unity.
{p 4 8 2}
{cmd:dummies(}{it:Varlist_1} [ \ {it:Varlist_2} \ ... \ {it:Varlist_N}]{cmd:)}
modifies the calculation of marginal effects for dummy variables. Let {it:xvar} be a categorical variables
with {it:K} ({it:K}>2) categories. In this case, not xvar, but ({it:K}-1) dummies are included
in the regression model. The estimated marginal effects for these {it:K}-1 dummies
may be misleading (see the example below). The correct result is obtained if
one specifies the dummy option by including the {it:K}-1 variables in {it:varlist_1}.
{p 4 8 2}
{cmd:eform} forces {cmd:margin} to define marginal effects as changes in the value of {it:exp(xb)},
where {it:xb} is the linear prediction.
{p 4 8 2}
{cmd:hascons} forces {cmd:margin} to display the expected value that corresponds to the constant term.
Note that this is not marginal effect (that is, change in expected value). For instance, after a {cmd:logit}
model the {cmd:hascons} option adds a new line to the output showing the predicted probability of the event
conditional on all explanatory variables are zero.
{p 4 8 2}
{cmd:mean} forces {cmd:margin} to compute the marginal effects at the sample means of explanatory variables.
{p 4 8 2}
{cmd:model(}{it:stata_cmd}{cmd:)} allows one to run {cmd:margin} after any command. {it:stata_cmd} must be
one of the commands listed above. {cmd:margin} will produce correct results only if
the estimation commannd issued before {cmd:margin} and {it:stata_cmd} have exactly the same
link functions, i.e. the function which links the linear combination of variables
and parameters to the outcome under study. It is the user's responsibility to ensure
that this condition holds.
{p 4 8 2}
{cmd:percent} causes {cmd:margin} to display the results in a percentage form.
{p 4 8 2}
{cmd:replace} causes {cmd:margin} to overwrite the estimation results left behind.
{p 4 8 2}
{cmd:table} causes {cmd:margin} to display a table containing the descriptive statistics of
individual marginal effects.
{title:Remarks}
{p 4 4 2}
{cmd:margin} is a work-in-progress; comments, suggestions, bug reports are welcome!
Please direct correspondence to the adress described at the end of the help file. {p_end}
{p 4 4 2}
To keep {cmd:margin} up-to-date, visit the website {browse "http://www.bkae.hu/bartus/stata"} or type {p_end}
{p 8 4 2} {cmd: net from http://www.bkae.hu/bartus/stata} {p_end}
{p 8 4 2} {cmd: net install margin}{p_end}
{title:Warning}
{p 4 4 2}
Before using {cmd:margin}, make sure that you coded your dummy variables as 0/1 variables.
Otherwise your dummies will be considered as being continuous variables, and you run
the risk of obtaining misleading results.
{title:Examples}
{p 4 4 2}
Illustrating the importance of the dummies( varlist_1 \ ... ) option
{p_end}
{p 4 4 2}Type the following commands:{p_end}
{p 8 8 2}{cmd:. [save mydata, replace]}{p_end}
{p 8 8 2}{cmd:. tabi 60 30 10 \ 20 60 20 \ 10 10 80 , replace }{p_end}
{p 8 8 2}{cmd:. xi: mlogit col i.row [fw=pop] }{p_end}
{p 8 8 2}{cmd:. margin }{p_end}
{p 4 4 2}You will see that the marginal effects do not correspond to changes in probabilities in
the 3X3 table. Only the changes in the probabilities of the third outcome approximate the true
changes, namely 0.1 and 0.7. The correct results are obtained after specifying the dummies(..)
option: {p_end}
{p 8 8 2}{cmd:. margin , dummies(_I*)}{p_end}
{title:Acknowledgements}
{p 4 4 2}
Some parts of the code relies on margfx (version 30 Jul 1999 for Stata 5)
written by Jonah B. Gelbach, Dept of Economics, Univ of MD at College Park.
{title:Also see}
{p 4 13 2}
Online: help for {help est}, {help postest}; {help mfx}
{title:Author}
{p 4 8 2} {browse "http://www.bkae.hu/bartus":Tamas Bartus} {p_end}
{p 4 8 2} Department of Sociology and Social Policy,{p_end}
{p 4 8 2} Corvinus University Budapest (former University of Economics),{p_end}
{p 4 8 2} Budapest, Hungary {p_end}
{p 4 8 2} URL: {browse "http://www.uni-corvinus.hu/bartus":http://www.uni-corvinus.hu/bartus}{p_end}
{p 4 8 2} Email: {browse "mailto:tamas.bartus@uni-corvinus.hu":tamas.bartus@uni-corvinus.hu}{p_end}