{smcl}
{hline}
help for {hi:film}
{hline}

{title:Fully interacted linear matching}

{p 8 21 2}{cmdab:film}
{it:outcomevar}
{it:treatment_dummy}
{it:indepvars}
[{cmd:if} {it:exp}]
[{cmd:in} {it:range}]
[{cmd:,}
    {cmdab:ate}
	{cmdab:com:mon}
    {cmdab:lin:ear}{cmd:(}{it:indepvars2}{cmd:)}
	{cmdab:rob:ust}
    {cmdab:cl:uster}{cmd:(}{it:varname}{cmd:)}
    {cmdab:l:evel}{cmd:(}{it:real}{cmd:)}
    {cmd:show}]

{title:Description}

{p 4 4 2}{cmd:film} first runs a simple linear regression (OLS) of {it:outcomevar} on
{it:treatment_dummy} and {it:indepvars}, then implements a fully interacted linear regression model
in which {it:treatment_dummy} is interacted with {it:each}
variable in {it:indepvars}. The overall effect - averaged for the treated
and optionally also for the non-treated and the population - is then displayed,
together with analytical standard errors. 
In an evaluation framework this amounts to allowing the treatment effect to vary according
to each observable variable in {it:indepvars}.
An F-test of no heterogeneous effects is also displayed.

{p 4 4 2}{cmd:film} allows the user to perform various steps to increase the flexibility
of the regression model until it resembles a matching estimator (see {help psmatch2}) --
from simple dummy-variable OLS regression, to full interactions to imposition of the common
support -- while retaining analytical standard errors and related saving in bootstrapping time.

{p 4 4 2}{cmd:film} returns in {it:r(ols)} the coefficient from the OLS regression, and
in {it:r(att)} and optionally in {it:r(atu)} and {it:r(ate)} the estimated treatment effects
from the fully interacted model for the treated (Average effect of Treatment on the Treated)
and optionally untreated (Average effect of Treatment on the Untreated) and population
(Average Treatment Effect). Corresponding standard errors and p-values are also returned in
{it:r(name_se)} and {it:r(name_p)}.

{title:Options}

{p 4 8 2}{cmdab:ate} with this option the average treatment effect (ate) and average treatment
effect on the untreated (atu) are reported in addition to the average treatment effect on the treated (att).
The estimates are returned in {it:r(ate)}, {it:r(atu)} and {it:r(att)} respectively, see above.

{p 4 8 2}{cmdab:com:mon} performs all estimation after having imposed common support
(A) in the absence of the option {cmdab:ate}, on the treated for OLS and att and
(B) with the option {cmdab:ate}, on the treated for att, on the untreated for atu and on both
for ate and OLS.
Common support is imposed at the boundaries by using the propensity score.

{p 4 8 2}{cmdab:lin:ear}{cmd:(}{it:indepvars2}{cmd:)} specifies that these independent
variables are to be controlled only linearly in the outcome regression, i.e. they shall
not be interacted with the treatment dummy. Note though that they will be used to
impose common support under the option {cmd:common}.

{p 4 8 2}{cmdab:robust} specifies that the Huber/White/sandwich estimator of variance is
to be used in place of the traditional calculation. {cmd:robust} combined with
{cmd:cluster()} further allows observations which are not independent within cluster
(although they must be independent between clusters). 

{p 4 8 2}{cmdab:cl:uster}{cmd:(}{it:varname}{cmd:)} specifies that the observations are
independent across groups (clusters) but not necessarily independent within groups.
{it:varname} specifies to which group each observation belongs; e.g., {cmd:cluster(personid)}
in data with repeated observations on individuals. Specifying {cmd:cluster()} implies
{cmd:robust}.

{p 4 8 2}{cmdab:l:evel}{cmd:(}{it:real}{cmd:)} specifies the confidence level, in percent, for confidence
intervals of the coefficients. Default is 95. 

{p 4 8 2}{cmdab:show} requests all the interaction effects (with significance level)
and the R2 of the interacted model to be displayed. 


{title:Examples}

    {inp: . film logwage training age gender exper, robust}
    {inp: . film logwage training age gender exper, robust ate common}
    {inp: . film logwage training age gender, ate linear(exper) show}

{title:Also see}

{p 4 4 2}The commands {help psmatch2}.

{title:Disclaimer}

{p 4 4 2}THIS SOFTWARE IS PROVIDED "AS IS" WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESSED OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. THE ENTIRE RISK AS TO THE QUALITY AND PERFORMANCE OF THE PROGRAM IS WITH YOU. SHOULD THE PROGRAM PROVE DEFECTIVE, YOU ASSUME THE COST OF ALL NECESSARY SERVICING, REPAIR OR CORRECTION.

{p 4 4 2}IN NO EVENT WILL THE COPYRIGHT HOLDERS OR THEIR EMPLOYERS, OR ANY OTHER PARTY WHO MAY MODIFY AND/OR REDISTRIBUTE THIS SOFTWARE, BE LIABLE TO YOU FOR DAMAGES, INCLUDING ANY GENERAL, SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING OUT OF THE USE OR INABILITY TO USE THE PROGRAM (INCLUDING BUT NOT LIMITED TO LOSS OF DATA OR DATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY YOU OR THIRD PARTIES OR A FAILURE OF THE PROGRAM TO OPERATE WITH ANY OTHER PROGRAMS), EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES.


{title:Authors}

{p 4 4 2}Edwin Leuven, Department of Economics, University of Amsterdam. If
you observe any problems {browse "mailto:e.leuven@uva.nl"}.

{p 4 4 2}Barbara Sianesi, Institute for Fiscal Studies, London, UK.