iptATE.sthlp

{smcl}
{* *! version 1.0.0  24may2011}{...}
{cmd:help iptATE}
{hline}

{title:Title}

{phang}
{bf: iptATE} {hline 2} Inverse probability tilting estimate of the average treatment effect (ATE)



{title:Syntax}

{p 8 17 2}
{cmdab:iptATE}
[{varlist}]
{ifin}
{weight}
[{cmd:,}
{it:options}]

{synoptset 20 tabbed}{...}
{synopthdr}
{synoptline}
{syntab:Main}
{synopt:{opt cluster(clustvar)}}Compute standard errors which allow for arbitrary patterns of dependence within sampling units defined 
							by {it:clustvar.} NOTE: no finite sample degrees of freedom correction is implemented.{p_end}
{synopt:{opt optroutine(optchoice)}}Specifies optimization routine use in the computation of the inverse probability tilts. If {it:optchoice} 
								is set equal to {it:e1} then the Broyden-Fletcher-Goldfarb-Shanno (bfgs) quasi-Newton method with analytic 
								first derivatives is used (this is the default). If {it:optchoice} is set equal to {it:e2} a modified Newton-Raphson 
								procedure based on analytic first and second derivatives is used. This is the method detailed in the Appendix of 
								Graham, Pinto and Egel (forthcoming).{p_end}	
{synopt:{opt bal:ance}}Produces a table means of {it: controlvars} by treatment status.{p_end}
{synoptline}
{p2colreset}{...}
{p 4 6 2}
{cmd:pweight}s are allowed; see {help weight}.

{title:Description}
{pstd}
{cmd:iptATE} calculates the inverse probability tilting (IPT) estimate of the average treatment effect (ATE) as described in Graham, Pinto and Egel 
	(forthcoming). The {it: varlist} should be of the form {it: outcomevar} {it: treatmentvar} {it: controlvars}; {it: outcomevar} equals
	the variable name of the scalar outcome of interest; {it: treatmentvar} the variable name of the binary treatment indicator, and {it: controlvars} 
	a variable list which defines the t(X) vector as described in Graham, Pinto and Egel (forthcoming).

{title:Examples}
{phang}{cmd:. iptATE Y D X}
{phang}{cmd:. iptATE Y D X, cluster(village)}	

{title:Reference}

{phang}
Graham, Bryan S., Cristine Pinto and Daniel Egel. (forthcoming) 
{browse "https://files.nyu.edu/bsg1/public/": Inverse probability tilting for moment condition models with missing data, }
{it:Review of Economic Studies}