scholarly journals Commentary on Instrumental Variable Methods for Low Compliance

2021 ◽  
Author(s):  
Ian Shrier ◽  
Tyrel Stokes ◽  
Russell John Steele
Keyword(s):  
British Journal ◽  
Instrumental Variable ◽  
Sports Medicine ◽  
Causal Effects ◽  
Recent Editorial

A recent editorial in the British Journal of Sports Medicine (BJSM) suggested instrumental var-iable (IV) analysis has advantages in estimating causal effects when there is low compliance. We originally submitted a version of this commentary to BJSM as an editorial (they do not have a letter to editor section) but it was rejected without review. The original BJSM editorial included several important errors, presented results that are inconsistent with the results of an IV analysis, and omitted definitions and important limitations. All of these factors contrib-uted to inappropriate interpretations. This commentary highlights the most important er-rors. We also believe the BJSM editorial serves as another reminder that appropriate statisti-cians should be included from the beginning of the study wherever possible. At the very least, they should be the co-authors responsible for calculating results and ensuring the write-up is consistent with the results.

scholarly journals Expression of concern: Which specific modes of exercise training are most effective for treating low back pain? Network meta-analysis

2020 ◽  
pp. bjsports-2019-100886eoc1
Keyword(s):  
British Journal ◽  
Low Back Pain ◽  
Back Pain ◽  
Clinical Practice ◽  
Exercise Training ◽  
Sports Medicine ◽  
Meta Analysis ◽  
Editorial Note ◽  
Low Back ◽  
Recent Editorial

Editorial NoteAs discussed in a recent editorial, the British Journal of Sports Medicine (BJSM) rescinds the Expression of Concern [1] for a recent network meta-analysis (NMA) [2] that was issued solely on the basis of comments by Professor Maher and colleagues [3]. The original authors (Dr Belavy and colleagues) have responded [4]. The original NMA paper did not require any changes. We editors of the BJSM have full confidence in the findings of the NMA [2]. The findings of the NMA inform clinical practice and can serve to inform clinical practice guidelines.Karim Khan, MD, PhDEditor-in-Chief, BJSMJuly 27th, 2020REFERENCES1 Expression of concern: Which specific modes of exercise training are most effective for treating low back pain? Network meta-analysis. Br J Sports Med 2020;:bjsports-2019-100886eoc1. doi:10.1136/bjsports-2019-100886eoc12 Owen PJ, Miller CT, Mundell NL, et al. Which specific modes of exercise training are most effective for treating low back pain? Network meta-analysis. Br J Sports Med 2019;:in press. doi:10.1136/bjsports-2019-1008863 Maher CG, Hayden JA, Saragiotto BT, et al. Letter in response to: “Which specific modes of exercise training are most effective for treating low back pain? Network meta-analysis” by Owen et al. Br J Sports Med Published Online First: 5 February 2020. doi:10.1136/bjsports-2019-1018124 Belavy DL, Owen PJ, Miller CT, et al. Response to Discussion: “Which specific modes of exercise training are most effective for treating low back pain? Network meta-analysis.” Br J Sports Med Published Online First: 10 June 2020. doi:10.1136/bjsports-2020-102673

scholarly journals Graphical Models for Quasi-experimental Designs

2015 ◽  
Vol 46 (2) ◽  
pp. 155-188 ◽  
Author(s):  
Peter M. Steiner ◽  
Yongnam Kim ◽  
Courtney E. Hall ◽  
Dan Su
Keyword(s):  
Graphical Models ◽  
Instrumental Variable ◽  
Regression Discontinuity ◽  
Experimental Designs ◽  
Causal Effects ◽  
Confounding Bias ◽  
Randomized Controlled ◽  
Quasi Experimental ◽  
Causal Graphs ◽  
Latent Subgroups

Randomized controlled trials (RCTs) and quasi-experimental designs like regression discontinuity (RD) designs, instrumental variable (IV) designs, and matching and propensity score (PS) designs are frequently used for inferring causal effects. It is well known that the features of these designs facilitate the identification of a causal estimand and, thus, warrant a causal interpretation of the estimated effect. In this article, we discuss and compare the identifying assumptions of quasi-experiments using causal graphs. The increasing complexity of the causal graphs as one switches from an RCT to RD, IV, or PS designs reveals that the assumptions become stronger as the researcher’s control over treatment selection diminishes. We introduce limiting graphs for the RD design and conditional graphs for the latent subgroups of compliers, always takers, and never takers of the IV design, and argue that the PS is a collider that offsets confounding bias via collider bias.

Bayesian Versus Maximum Likelihood Estimation of Treatment Effects in Bivariate Probit Instrumental Variable Models

2018 ◽  
Vol 7 (3) ◽  
pp. 651-659 ◽  
Author(s):  
Florian M. Hollenbach ◽  
Jacob M. Montgomery ◽  
Adriana Crespo-Tenorio
Keyword(s):  
Maximum Likelihood ◽  
Maximum Likelihood Estimation ◽  
Instrumental Variable ◽  
Likelihood Estimation ◽  
Direct Calculation ◽  
Causal Effects ◽  
Parameter Estimates ◽  
Bivariate Probit ◽  
Probit Models ◽  
Quantities Of Interest

Bivariate probit models are a common choice for scholars wishing to estimate causal effects in instrumental variable models where both the treatment and outcome are binary. However, standard maximum likelihood approaches for estimating bivariate probit models are problematic. Numerical routines in popular software suites frequently generate inaccurate parameter estimates and even estimated correctly, maximum likelihood routines provide no straightforward way to produce estimates of uncertainty for causal quantities of interest. In this note, we show that adopting a Bayesian approach provides more accurate estimates of key parameters and facilitates the direct calculation of causal quantities along with their attendant measures of uncertainty.

scholarly journals Background Indicators

Econometrics ◽  
2019 ◽  
Vol 7 (2) ◽  
pp. 20
Author(s):  
Burkhard Raunig
Keyword(s):  
Instrumental Variable ◽  
Latent Variable ◽  
Structural Model ◽  
Simulation Experiment ◽  
Random Noise ◽  
Empirical Work ◽  
Causal Effects ◽  
Economic Uncertainty ◽  
Aggregate Consumption ◽  
The Impact

It is customary to assume that an indicator of a latent variable is driven by the latent variable and some random noise. In contrast, a background indicator is also systematically influenced by variables outside the structural model of interest. Background indicators deserve attention because in empirical work they are difficult to distinguish from ordinary effect indicators. This paper assesses instrumental variable (IV) estimation of the effect of a latent variable in a linear model when a background indicator replaces the latent variable. It turns out that IV estimates are inconsistent in many important cases. In some cases, the estimates capture causal effects of the indicator rather than causal effects of the latent variable. A simulation experiment that considers the impact of economic uncertainty on aggregate consumption illustrates some of the results.

scholarly journals Measurement errors in the binary instrumental variable model

Biometrika ◽  
2019 ◽  
Vol 107 (1) ◽  
pp. 238-245
Author(s):  
Zhichao Jiang ◽  
Peng Ding
Keyword(s):  
Measurement Error ◽  
Instrumental Variable ◽  
Measurement Errors ◽  
Causal Effects ◽  
Sensitivity Analyses ◽  
Variable Model ◽  
Differential Measurement ◽  
True Value ◽  
Instrumental Variable Model ◽  
Outcome Biases

Summary Instrumental variable methods can identify causal effects even when the treatment and outcome are confounded. We study the problem of imperfect measurements of the binary instrumental variable, treatment and outcome. We first consider nondifferential measurement errors, that is, the mismeasured variable does not depend on other variables given its true value. We show that the measurement error of the instrumental variable does not bias the estimate, that the measurement error of the treatment biases the estimate away from zero, and that the measurement error of the outcome biases the estimate toward zero. Moreover, we derive sharp bounds on the causal effects without additional assumptions. These bounds are informative because they exclude zero. We then consider differential measurement errors, and focus on sensitivity analyses in those settings.

Sign in / Sign up

Export Citation Format

Share Document