NSW {DOS2}

National Supported Work Randomized Experiment

Description

National Supported Work (NSW) randomized experiment in 185 matched pairs. Used as an example in Chapter 2 of Design of Observational Studies.

Usage

data("NSW")

Format

A data frame with 370 observations on the following 11 variables.

id

Matched pair id, 1, 1, 2, 2, ..., 185, 185.

z

z=1 for treated, z=0 for control

age

Age in years

edu

Education in years

black

1=black, 0=other

hisp

1=Hispanic, 0=other

married

1=married, 0=other

nodegree

1=no High School degree, 0=other

re74

Earnings in 1974, a covariate

re75

Earnings in 1975, a covariate

re78

Earnings in 1978, an outcome

Details

Compare with Table 2.2 of Design of Observational Studies.

Source

Dehejia and Wahba (1999).

References

Couch, K.A. (1992) <doi:10.1086/298292> "New evidence on the long-term effects of employment training programs". Journal of Labor Economics 10, 380-388.

Dehejia, R.H., Wahba, W. (1999) <doi:10.1080/01621459.1999.10473858> "Causal effects in nonexperimental studies: Reevaluating the evaluation of training programs". Journal of the American Statistical Association 94, 1053-1062.

LaLonde, R.J. (1986) <https://www.jstor.org/stable/1806062> "Evaluating the econometric evaluations of training programs with experimental data". American Economic Review 76, 604-620.

Examples

data(NSW)
fivepairs<-c(15,37,46,151,181)
# Table 2.2 in Design of Observational Studies (DOS)
NSW[is.element(NSW$id,fivepairs),]
# Pair differences in earnings after treatment
dif<-NSW$re78[NSW$z==1]-NSW$re78[NSW$z==0]
# Chapter 2, footnote 7 of
stats::wilcox.test(dif,conf.int=TRUE)