Download presentation
Presentation is loading. Please wait.
Published byThomas King Modified over 9 years ago
1
Non-experimental methods Markus Goldstein The World Bank DECRG & AFTPM
2
Objective Find a plausible counterfactual Every method is associated with an assumption The stronger the assumption the more we need to worry about the causal effect »Question your assumptions Reality check
3
Program to evaluate Hopetown HIV/AIDS Program (2008-2012) Objectives Reduce HIV transmission Intervention: Peer education Target group: Youth 15-24 Indicator: Pregnancy rate (proxy for unprotected sex)
4
I. Before-after identification strategy (aka reflexive comparison) Counterfactual: Rate of pregnancy observed before program started EFFECT = After minus Before
5
YearNumber of areas Teen pregnancy rate (per 1000) 20087062.90 20127066.37 Difference+3.47
6
Effect = +3.47 Intervention Counterfactual assumption: no change over time Question: what else might have happened in 2008-2012 to affect teen pregnancy?
7
Examine assumption with prior data Number of areas Teen pregnancy (per 1000) 200420082012 7054.9662.9066.37 Assumption of no change over time looks a bit shaky
8
II. Non-participant identification strategy Counterfactual: Rate of pregnancy among non-participants Teen pregnancy rate (per 1000) in 2012 Participants66.37 Non-participants57.50 Difference+8.87
9
Counterfactual assumption: Without intervention participants have same pregnancy rate as non-participants Effect = +8.87 Participants Non-participants Question: how might participants differ from non- participants?
10
Test assumption with pre-program data ? REJECT counterfactual hypothesis of same pregnancy rates
11
III. Difference-in-Difference identification strategy Counterfactual: 1.Nonparticipant rate of pregnancy, purging pre-program differences in participants/nonparticipants 2.“Before” rate of pregnancy, purging before-after change for nonparticipants 1 and 2 are equivalent
12
Average rate of teen pregnancy in 20082012Difference (2008-2012) Participants (P)62.9066.373.47 Non-participants (NP)46.3757.5011.13 Difference (P-NP)16.538.87 -7.66
13
57.50 - 46.37 = 11.13 66.37 – 62.90 = 3.47 Non-participants Participants Effect = 3.47 – 11.13 = - 7.66
14
After Before Effect = 8.87 – 16.53 = - 7.66 66.37 – 57.50 = 8.87 62.90 – 46.37 = 16.53
15
Counterfactual assumption: Without intervention participants and nonparticipants’ pregnancy rates follow same trends
16
74.0 16.5
17
74.0 -7.6
18
Questioning the assumption Why might participants’ trends differ from that of nonparticipants?
19
Examine assumption with pre-program data counterfactual hypothesis of same trends doesn’t look so believable Average rate of teen pregnancy in 20042008Difference (2004-2008) Participants (P)54.9662.907.94 Non-participants (NP)39.9646.376.41 Difference (P=NP)15.0016.53 +1.53 ?
20
IV. Matching with Difference-in- Difference identification strategy Counterfactual: Comparison group is constructed by pairing each program participant with a “similar” nonparticipant using larger dataset – creating a control group from similar (in observable ways) non- participants
21
Counterfactual assumption: Question: how might participants differ from matched nonparticipants? Unobserved characteristics do not affect outcomes of interest Unobserved = things we cannot measure (e.g. ability) or things we left out of the dataset
22
73.36 66.37 Matched nonparticipant Participant Effect = - 7.01
23
Can only test assumption with experimental data Apply with care – think very hard about unobservables Studies that compare both methods (because they have experimental data) find that: unobservables often matter! direction of bias is unpredictable!
24
V. Regression discontinuity identification strategy Applicability: When strict quantitative criteria determine eligibility Counterfactual: Nonparticipants just below the eligibility cutoff are the comparison for participants just above the eligibility cutoff
25
Counterfactual assumption: Question: Is the distribution around the cutoff smooth? Then, assumption might be reasonable Question: Are unobservables likely to be important (e.g. correlated with cutoff criteria)? Then, assumption might not be reasonable However, can only estimate impact around the cutoff, not for the whole program Nonparticipants just below the eligibility cutoff are the same (in observable and unobservable ways) as participants just above the eligibility cutoff
26
Target transfer to poorest schools Construct poverty index from 1 to 100 Schools with a score <=50 are in Schools with a score >50 are out Inputs transfer to poor schools Measure outcomes (i.e. test scores) before and after transfer Example: Effect of school inputs on test scores
28
Non-Poor Poor
30
Treatment Effect
31
Applying RDD in practice: Lessons from an HIV-nutrition program Lesson 1: criteria not applied well –Multiple criteria: hh size, income level, months on ART –Nutritionist helps her friends fill out the form with the “right” answers –Now – unobservables separate treatment from control… Lesson 2: Watch out for criteria that can be altered (e.g. land holding size)
32
Gold standard is randomization – minimal assumptions needed, intuitive estimates Nonexperimental requires assumptions – can you defend them? Summary
33
Different assumptions will give you different results The program: ART treatment for adult patients Impact of interest: effect of ART on children of patients (are there spillover & intergenerational effects of treatment?) –Child education (attendance) –Child nutrition Data: 250 patient HHs 500 random sample HHs –Before & after treatment Can’t randomize ART so what is the counterfactual
34
Possible counterfactual candidates Random sample difference in difference –Are they on the same trajectory? Orphans (parents died – what would have happened in absence of treatment) –But when did they die, which orphans do you observe, which do you not observe? Parents self report moderate to high risk of HIV –Self report! Propensity score matching –Unobservables (so why do people get HIV?)
35
Estimates of treatment effects using alternative comparison groups Standard errors clustered at the household level in each round. Includes child fixed effects, round 2 indicator and month-of-interview indicators. Compare to around 6.4 if we use the simple difference in difference using the random sample
36
Estimating ATT using propensity score matching Allows us to define comparison group using more than one characteristic of children and their households Propensity scores defined at household level, with most significant variables being single-headed household and HIV risk
37
Probit regression results Dependent variable: household has adult ARV recipient
38
ATT using propensity score matching
39
Nutritional impacts of ARV treatment Includes child fixed effects, age controls, round 2 indicator, interviewer fixed effects, and month-of- interview indicators.
40
Nutrition with alternative comparison groups Includes child fixed effects, age controls, round 2 indicator, interviewer fixed effects, and month-of- interview indicators.
41
Summary: choosing among non- experimental methods At the end of the day, they can give us quite different estimates (or not, in some rare cases) Which assumption can we live with?
42
Thank you
Similar presentations
© 2024 SlidePlayer.com. Inc.
All rights reserved.