WIA Net Impacts: Preliminary Estimates ADARE Project Briefing

1
WIA Net Impacts Preliminary EstimatesADARE
Project Briefing 3
Welcome to the Marshall Center PowerPoint
Template. To add text, click on the visible text,
delete the template titles, and enter your
data. To add additional slides, use Ctrl M. Yo
u can modify the slide layout (i.e. select a
pre-set layout for tables instead of a layout for
text) by right-clicking on the slide background
and selecting Slide Layout. Commonly used sample
layouts are included. To delete slides you dont
want to use, choose Delete Slide from the Edit
menu. You can also Duplicate Slides through the
Insert menu. Pre-set slide transitions are includ
ed. They can be changed through the Slide Show
menu. Finally, to hide comment fields (like this
box), select Comments from the View menu.
Please email Sarah Looney at looneys_at_mail.utexas.e
du if you have any questions.

• Kevin Hollenbeck, W. E. Upjohn Institute
• Christopher T. King Daniel Schroeder,
• Ray Marshall Center, University of Texas-Austin
• May 7, 2003

W.E.Upjohn Institute for Employment Research
2
Overview
Use this page to create numbered lists.

• Background WIA Context
• Approach
• Preliminary Impact Estimates
• Discussion
• Next Steps

3
Background Context
This page creates bulleted lists but it also
includes animation (click F5 to view).
To add animation to other slides, choose Slide
Show from the menu bar. You can then choose from
Preset Animation features or design a custom
animation style.

• Key to assessing program impacts answer two
  questions
  
• What happened to those who participated in the
  program (the treatment group)?
  
• What would have happened to them if they had not
  participated?
  
• First question is straightforward, but the second
  is more complicated

4
Background Context
This page creates bulleted lists but it also
includes animation (click F5 to view).
To add animation to other slides, choose Slide
Show from the menu bar. You can then choose from
Preset Animation features or design a custom
animation style.

• The second question is answered by creating a
  group of people who are as similar as possible to
  the treatment group, except that they do not
  receive the treatment.
• Outcomes for the treatment group and this
  comparison group are tracked over time.
  
• Program impact is the difference between the
  outcomes for the two groups.
  

5
Background Context
This page creates bulleted lists but it also
includes animation (click F5 to view).
To add animation to other slides, choose Slide
Show from the menu bar. You can then choose from
Preset Animation features or design a custom
animation style.

• Ideal Approach Experimental design, which uses
  random assignment of participants to groups to
  ensure the similarity of the treatment and
  control groups.
• This is the gold standard, but it can be
  expensive to implement, and is not always
  feasible.

6
Background Context
This page creates bulleted lists but it also
includes animation (click F5 to view).
To add animation to other slides, choose Slide
Show from the menu bar. You can then choose from
Preset Animation features or design a custom
animation style.

• An experiment is not possible for WIA impact
  estimation because eligible clients cannot be
  denied service.
  
• Next-Best Approach? Quasi-experimental design,
  with a comparison group selected to ensure
  similarity to the treatment group.
  
• But, must be carefully done to get at true
  program impacts, and not just differences that
  already existed between the groups.

7
Background ...
This page creates bulleted lists but it also
includes animation (click F5 to view).
To add animation to other slides, choose Slide
Show from the menu bar. You can then choose from
Preset Animation features or design a custom
animation style.
History of quasi-experiments Multiple, complicat
ed answers to the impact question are not useful
to policymakers. Examples CETA CLMS, 1st Job C
orps Evaluation 1985 JTLS Advisory Committee
National JTPA Study, 2nd Job Corps Evaluation,
state evaluations (e.g., TX, UT, WA)
8
Background ...
This page creates bulleted lists but it also
includes animation (click F5 to view).
To add animation to other slides, choose Slide
Show from the menu bar. You can then choose from
Preset Animation features or design a custom
animation style.

• WIA Context
• Service Sequencing Core, Intensive Training
  Services
  
• Young, highly devolved program

9
Approach
This page creates bulleted lists but it also
includes animation (click F5 to view).
To add animation to other slides, choose Slide
Show from the menu bar. You can then choose from
Preset Animation features or design a custom
animation style.
Data WIASRD files for FL, GA, IL, MD, MO, TX,
WA ES registrant data for two states only and
linked UI wage records for all states.
Impacts estimated for WIA Adults Dislocated Wo
rkers, not Youth.
10
Approach ...
This page creates bulleted lists but it also
includes animation (click F5 to view).
To add animation to other slides, choose Slide
Show from the menu bar. You can then choose from
Preset Animation features or design a custom
animation style.
Treatment group those who received intensive
or training services and exited WIA in PY2000.
Comparison group those who received core serv
ices only, and exited in PY 2000.
Thus, this design estimates the incremental impa
ct of intensive/training, above and beyond the
receipt of core services.
11
Approach ...
This page creates bulleted lists but it also
includes animation (click F5 to view).
To add animation to other slides, choose Slide
Show from the menu bar. You can then choose from
Preset Animation features or design a custom
animation style.

• Impact estimates will be presented for these
  Labor Market Outcomes
  
• Employment in 4th post-exit quarter
• Earnings in 4th post-exit quarter
• Note States de-identified until state clearance
  is granted. Criterion for employment is a
  minimum of 100 quarterly earnings.

12
Preliminary Impact Estimates
This page creates bulleted lists but it also
includes animation (click F5 to view).
To add animation to other slides, choose Slide
Show from the menu bar. You can then choose from
Preset Animation features or design a custom
animation style.

• Outcomes for 6 states employment
  (conditional) earnings in 4th post-exit quarter.
  
• Employment, 4th Post-Exit Quarter
• Adults participation positive impacts in 3
  states, negative impacts in 3 states.
  
• Dislocated Workers participation positive
  impacts in 4 states, negative in 2 states.

13
Impact Estimates ...
This page creates bulleted lists but it also
includes animation (click F5 to view).
To add animation to other slides, choose Slide
Show from the menu bar. You can then choose from
Preset Animation features or design a custom
animation style.
14
Impact Estimates ...
This page creates bulleted lists but it also
includes animation (click F5 to view).
To add animation to other slides, choose Slide
Show from the menu bar. You can then choose from
Preset Animation features or design a custom
animation style.

• Earnings, 4th Post-Exit Quarter
• Adults participation positive impacts in 4
  states, zero impact in 1 state, negative impact
  in 1 state.
  
• Dislocated Workers participation positive
  impacts in 2 states, negative in 4 states.
  

15
Impact Estimates ...
This page creates bulleted lists but it also
includes animation (click F5 to view).
To add animation to other slides, choose Slide
Show from the menu bar. You can then choose from
Preset Animation features or design a custom
animation style.
16
Discussion
This page creates bulleted lists but it also
includes animation (click F5 to view).
To add animation to other slides, choose Slide
Show from the menu bar. You can then choose from
Preset Animation features or design a custom
animation style.
Still editing from here on. -DGS
Considerable impact variation across subgroups
(i.e., adults, dislocated workers) and across
states Small sample sizes, especially for WIA c
ore services Alternative matching techniques, e
.g., caliper matching and others
Local stratification
17
Next Steps
This page creates bulleted lists but it also
includes animation (click F5 to view).
To add animation to other slides, choose Slide
Show from the menu bar. You can then choose from
Preset Animation features or design a custom
animation style.
Use ES registrant data to construct comparison
groups for all 7 current ADARE states (and add
new states) Further refine and test alternative
statistical matching techniques
Estimate impacts using PY2001 WIA participant da
ta longer-term UI wage records
Vet methods and results more widely
18
Use this page as the final slide in your
presentation.
For More Information
Kevin Hollenbeck 269/385-0431 hollenbeck_at_upjohnin
stitute.org Chris King 512/471-2186 ctking_at_u
ts.cc.utexas.edu Daniel Schroeder 512/471-2196
schroed_at_uts.cc.utexas.edu