×

Status of Operations

This website is currently not being updated due to the suspension of Federal government services. The last update to the site was 1/22/18. Updates to the site will start again when the Federal government resumes operations.

  ETA Home  >  Research & Publication  >   

 


 
Series # :
  RERS 2001
 
Title :
  National Job Corps Study: Methodological Appendixes on the Impact Analysis
 
Release Date :
  2001
 
 
Abstract :
  In a series of appendixes, this report discusses methodological issues related to the 48-month impact analysis for the National Job Corps Study. The appendixes are intended to complement the 48-month impact report (Schochet et al. 2001), which presents impacts of Job Corps on key participant outcomes during the 48 months after random assignment.

This report contains the following five appendixes:

1. "The 12-, 30-, and 48-Month Interviews." The outcome measures for the 48-month impact analysis were constructed using follow-up interview data collected 12, 30, and 48 months after random assignment. This appendix provides a detailed discussion of the design of the follow-up interviews and examines response rates.

2. "The Treatment of Missing Values and Outliers." This appendix describes our procedure for treating missing values and outliers for the outcome measures used in the 48-month impact analysis.

3."The Adjustment for Crossovers." This brief appendix describes procedures that were used to adjust the impact estimates for the small number of control group members who enrolled in Job Corps during their three-year restriction period and afterwards.

4. "The Calculation of Sample Weights and Standard Errors." This appendix discusses the calculation of sample weights used in the 48-month impact analysis to obtain unbiased impact estimates that could be generalized to the study population. The appendix also discusses the calculation of standard errors of the impact estimates.

5. "Regression-Adjusted Impact Estimates." This appendix discusses impact estimates obtained using multivariate regression procedures. These regression-adjusted impact estimates are compared to the simple differences-in-means estimates that are presented in the impact report.



Publication Author(s)
1.
Battelle
2.
Decision Information Resources, Inc.
3.
Mathematica Policy Research, Inc.
 
Author(s):
  • Peter Z. Schochet
Full Text Document(s)
Additional Information

Hard copy available: YES