Bounds on Treatment Effects from Studies with Imperfect Compliance

By Alexander Balke, Judea Pearl. 1997

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

Causal Inference through a Witness Protection Program (2014, methodology): “Equation (5) implies |ωx′w − ωx′w′ | ≤ w, and as such by setting w = 0 we have that ωxw + ωx′w − ωx′w′ ≥ κ1x′.w + κ1x.w − κ1x′.w′ + κ1x.w′ − χxw′(Ū + L+ 2 w) + L (20) implies ηxw ≥ η1x.w+η1x.w′−η1x′.w′−η0x.w′ , one of the most complex relationships in (Balke and Pearl, 1997).”
The Econometrics of Randomized Experiments (2016, background): “Under these assumptions these bounds are sharp (Balke and Pearl, 1997).”
Interventional Few-Shot Learning (2020, background): “In the nonlinear case, it is shown in [6] that observations on IV provide a bound for the true causal effect.”
Instrumental Variables: An Econometrician's Perspective (2014, background): “llows that pr(Yi=1,Xi=0|Zi=1) (5.1) ≤pr(Yi=1,Xi=0|Zi=0). There are three more restrictions in this setting with a binary outcome, binary treatment and binary instrument; see Imbens and Rubin (1997b), Balke and Pearl (1997) and Richardson, Evans and Robins (2011) for details. For the ﬂu data, the sim”
A Likelihood-Based Analysis for Relaxing the Exclusion Restriction in Randomized Experiments with Imperfect Compliance (2013, methodology): “Nonparametric bounds on the average treatment effects of a randomized experiment with imperfect compliance over the whole population have been developed by Balke and Pearl (1997) under the exclusion restriction, supposing a binary treatment and a binary outcome.”
Do Get-Out-the-Vote Calls Reduce Turnout? The Importance of Statistical Methods for Field Experiments (2005, background): “…of the average treatment effects of telephone and personal canvassing.9 Gerber and Green (2000) found that get-outthe-vote calls have a significant negative effect of 7 Some argue that the treatment effect for the entire population is a more meaningful quantity (e.g., Balke and Pearl 1997).”
Controlling Selection Bias in Causal Inference (2011, methodology): “We envision our method being used as a first step in a pre-processing stage, before the application of any bounding (Balke and Pearl, 1997) or estimation procedure.”
Econometric Methods for Program Evaluation (2018, background): “Balke and Pearl (1997), Imbens and Rubin (1997), Heckman and Vytlacil (2005) and Kitagawa (2015) discuss testable implications of this assumption.”
Estimation of Average Treatment Effects With Misclassification (2007, background): “Note that equation ( 5 ) comes from m(v)τ(v0) = m(v0)τ(v) = 0, which is the equation that provides”
A review of instrumental variable estimators for Mendelian randomization (2015, background): “It has been established that the core IV assumptions only identify bounds for the causal effect rather than a single causal estimate.(24,25) However, these bounds are extremely imprecise and even in large samples are rarely informative.”
Experimental designs for identifying causal mechanisms (2013, methodology): “Then, it is straightforward to obtain the sharp bounds using the linear programming approach (Balke and Pearl, 1997).”
Bounds on Direct Effects in the Presence of Confounded Intermediate Variables (2008, methodology): “Balke (1995) and Balke and Pearl (1997) describe a computer program that takes symbolic description of linear programming problems and returns symbolic expressions for the desired bounds.”
Econometric Evaluation of Social Programs, Part II: Using the Marginal Treatment Effect to Organize Alternative Econometric Estimators to Evaluate Social Programs, and to Forecast their Effects in New Environments (2007, methodology): “Following Balke and Pearl (1997), Manski (1989, 1990, 1996, 2003), and Robins (1989), we can use the identified combinations”
Augmented designs to assess principal strata direct effects (2011, methodology): “The following proposition provides a simple generalization to bounded outcomes of the bounds on the ADE for binary outcomes derived by Cai et al. (2008), using the symbolic Balke–Pearl method (Balke and Pearl, 1997).”
Two-Sample Instrumental Variable Analyses Using Heterogeneous Samples (2017, background): “In the context of binary instrument and binary exposure, Assumption 7 means that P(t = de) = 0 and is often called the no-defiance assumption (Balke and Pearl, 1997).”