LaLonde

This data set gives the outcomes a well as treatment assignments and covariates for the National Supported Work Study, as analyzed in LaLonde (1986).

datasets

The heterogeneous treatment effect estimation procedure
proposed by Imai and Ratkovic (2013).
The proposed method is applicable, for
example, when selecting a small number of most (or least)
efficacious treatments from a large number of alternative
treatments as well as when identifying subsets of the
population who benefit (or are harmed by) a treatment of
interest. The method adapts the Support Vector Machine
classifier by placing separate LASSO constraints over the
pre-treatment parameters and causal heterogeneity parameters of
interest. This allows for the qualitative distinction between
causal and other parameters, thereby making the variable
selection suitable for the exploration of causal heterogeneity.
The package also contains the function, INT, which estimates
the average marginal treatment effect, the average treatment
combination effect, and the average marginal treatment interaction
effect proposed by Egami and Imai (2015).

Naoki Egami

FindIt

Finding Heterogeneous Treatment Effects

LaLonde function

A data frame consisting of 12 columns (including a treatment assignment vector) and 2787 observations. <table><tr><td>
	  outcome </td>
<td> integer </td>
<td> whether earnings in 1978 are larger than in 1975 </td>
</tr><tr><td> 0,1 </td>
<td>
	  treat </td>
<td> integer </td>
</tr><tr><td> whether the individual received the treatment </td>
<td> 0,1 </td>
<td>
	  age </td>
</tr><tr><td> numeric </td>
<td> age in years </td>
<td>  </td>
</tr><tr><td>
	  educ </td>
<td> numeric </td>
<td> education in years </td>
</tr><tr><td>  </td>
<td>
	  black </td>
<td> factor </td>
</tr><tr><td> black or not </td>
<td> 0,1 </td>
<td> 
	  hisp </td>
</tr><tr><td> factor </td>
<td> hispanic or not </td>
<td> 0,1 </td>
</tr><tr><td> 
	  white </td>
<td> factor </td>
<td> white or not </td>
</tr><tr><td> 0,1 </td>
<td> 
	  marr </td>
<td> factor </td>
</tr><tr><td> married or not  </td>
<td> 0,1 </td>
<td>
	  nodegr </td>
</tr><tr><td> factor </td>
<td>  an indicator for no high school degree </td>
<td> 0,1 </td>
</tr><tr><td> 
	  log.re75 </td>
<td> numeric </td>
<td> log of earnings in 1975 </td>
</tr><tr><td> </td>
<td> 
	  u75 </td>
<td> factor </td>
</tr><tr><td> unemployed in 1975 </td>
<td> 0,1 </td>
<td>
	  outcome </td>
</tr></table> 

Format

Data from the National Supported Work Study.  A benchmark matching dataset. 1975 earnings are pre-treatment.

LaLonde: National Supported Work Study Experimental Data

Description

Usage

Arguments

Format

Source

References

outcome	integer	whether earnings in 1978 are larger than in 1975
0,1	treat	integer
whether the individual received the treatment	0,1	age
numeric	age in years
educ	numeric	education in years
	black	factor
black or not	0,1	hisp
factor	hispanic or not	0,1
white	factor	white or not
0,1	marr	factor
married or not	0,1	nodegr
factor	an indicator for no high school degree	0,1
log.re75	numeric	log of earnings in 1975
	u75	factor
unemployed in 1975	0,1	outcome