52006200 Families and Social Policy September 5 2006

1
5200/6200Families and Social
PolicySeptember 5 2006
2
This week.

• Policy Evaluation and Politics
• Ellwood, D.  1988.  "Values the Helping
  Conundrums." Poor Support, C 2. New York Basic
  Books, Pp. 14-44. WEBSITE
• Urban Institute, Federalism after Katrina how
  social programs respond http//www.urban.org/Uplo
  adedPDF/311344_after_katrina.pdf Internet
• Galster, G.  1996.  "The Challenges for Policy
  Research in a Changing Environment." The Future
  of the Public Sector Series No. 7. Urban
  Institute.
• http//www.urban.org/UploadedPDF/pub_07.pdf
• Dye, Thomas R. (2005) Policy Evaluation
  Finding Out What Happens After A Law Is Passed.
  READER
• Book Reportspassing list around
• Case for marriage, linda waite
• Promises to keep, kathryn Edin
• Would be week 11
• Thursday
• Politics, values, and why and how we should we
  help
• Today
• Basics of Policy Evaluation
• Types of policy evaluation
• What are the implications of federalism for
  evaluation
• An Example/What is selection bias?
• Application to Marriage policies

3

• TENTATIVE SCHEDULE
• All book reports on Tuesday of the week assigned
• WEEK SIX 9/26-9/28 HISTORY OF SOCIAL WELFARE
  IN THE U.S.
• WEEK SEVEN 10/3-10/5 WELFARE REFORM AND
  REAUTHORIJZATION
• Melissa Napier/nickel and dimed Tuesday 10/3
• Down and out in America the origins of
  homelessness 10/3who?
• Work over welfare / Jared Bunnell
• WEEK NINE 10/17-10/19 WOMEN, FAMILIES, AND
  POLICY CARING AND COPING
• The time bind, AMY FOX
• The high cost of motherhood, Krystal Razee
• Balancing Act, Mary Ann Leimbech
• WEEK TEN 10/24-10/26 WOMEN, TEEN PREGNANCY
  THE POLICY PROCESS
• When children want children, marissa Martin
• WEEK ELEVEN 10/31-11/2 WOMEN, MARRIAGE
  POLICY
• Promises to keep, dieu
• WEEK TWELVE 11/7-11/9 CHILDREN, INCOME AND
  POLICY
• It takes a village, Who?
• WEEK THIRTEEN 11/14-11/16 CHILDREN, NUTRITION,
  FOOD INSECURITY, AND POLICY
• Sweet charity, Karen gilbert

4
Evaluationdoes the government know what they are
doing?

• Yes
• No

5
What do they know?

• How much money has been spent
• How many are using the program
• The impact of a policy change
• How programs are organized

6
Welfare caseload data
7
Outcome of welfare reform?
8
What they dont often know

• What the benefits are of program for society
• Welfare reform
• If there are effectsshort term or long term?
• Positive or negative effects?
• Are the benefits greater than the costs?
• Expanding medicaid but shrinking caseloads

9
The basics

• Example Goals 2000
• Consideration of the effects of a policy
• Impact of policy change for the target group
• Short and long term benefits
• Indirect and symbolic costs and benefits
• Calculate the costs and benefits?

10
Standard types of evaluation

• Hearings by legislators
• Breaux welfare hearing focuses on hard-to-employ
  Americans
• Site visits
• Hatch and child care
• Program measures
• Welfare caseloads
• Citizen complaints
• AARP

11
Systematic evaluations

• Some sort of comparison or counterfactual
• Before and after comparisons
• Ex Health status and Health insurance coverage
  of kids in Utah before and after the
  implementation of the CHIP program
• Comparisons between areas or groups with and
  without programs
• Compare individuals who participate in jobs
  programs, before and after as well as comparing
  individuals who didnt participate in job
  training program, before and after.

12
Systematic evaluations

• Comparisons of control and experimental groups
  before and after
• classic research design Control and experimental
  must be the same in every way
• New Hope program in WI
• could get 4 benefits
• job search assistance
• placement in a community service job
• earnings supplement
• subsidized health insurance
• subsidized child care.

13
PROBS W/ EXPERIMENTAL DESIGN

• Hawthorne
• Generalizing to nation
• Ethical issues
• Expense!

14
How to Study?
15
WIC and Infant Birth Weight

• investigate the effects of prenatal WIC
  participation on relevant child outcomes with a
  nationally representative sample
• use fixed effect sibling models to minimize
  potential for selection bias

16
Data

• 1990 to 1996 survey rounds of the merged
  child-mother sample of the National Longitudinal
  Survey of Youth (NLSY)
• Overall N 1984 children
• (Sibship sample N 980)

17

• SELECTION
• selection is a question of bias
• positive-if families who have more motivation are
  more likely to participate and then more likely
  to have better outcomes
• negative if families who are doing poorly are
  more likely to get into the program and more
  likely to do worse on outcome.

18

• Regular methods
• Y ? ?1P ?2x ?
• Problem unmeasured characteristics of mothers
  and children can lead to bias in estimating
  relationships
• Sibling fixed effects Methods
• ?yij yj) ?(Pij Pi) ?(xij xi) (?ij
  ?j)
• Advantage removes the biasing effects of
  unmeasured mother and child characteristics

19
Prenatal WIC Participation among NLSY Sibling
Pairs
20
Key Outcomes

• Infant Birth Weight
• Measured in ounces
• Difficult Temperament
• Age-appropriate items pertaining to aspects of
  predictability, fearfulness, positive affect, and
  friendliness

21

• Findings before selection
• WIC is linked with lower infant weight
• WIC is linked with increases in difficult
  temperament
• Findings after selection
• WIC is linked with higher infant weight
• WIC is associated with decreases in difficult
  temperament

22
Problems with evaluation

• Gov..
• gov wants to satisfy many different groups with
  the same policy.
• programs have symbolic value tarnished by
  scathing eval
• Good policy eval takes a lot of time and
• Academic/Gov
• experimental research is unfair so it should not
  be done
• New hope and that darn computer
• If eval. fails to find effects, research was
  tainted by bias.