Pith. sign in

REVIEW 3 major objections 4 minor 43 references

A generalized plurality rule identifies multiple treatment effects with invalid instruments, and a sampling confidence interval keeps nominal coverage at parametric rate.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

T0 review · deepseek-v4-flash

2026-08-01 07:18 UTC pith:RQ23EEP6

load-bearing objection Solid extension of single-treatment IV robustness to multiple exposures, with a genuinely new sampling CI, but the parametric-rate length claim leans on an untestable high-level condition (Assumption 3) that can fail; coverage is fine. the 3 major comments →

arxiv 2607.21481 v1 pith:RQ23EEP6 submitted 2026-07-23 stat.ME

Identification and Robust Inference for Multiple Treatment Effects with Possibly Invalid Instruments

Ziwei Mei , Qingliang Fan , Zijian Guo This is my paper
classification stat.ME
keywords causal inferenceinvalid instrumentsmultiple treatmentsgeneralized plurality rulegeneralized majority rulesampling confidence intervallocally invalid instrumentsMendelian randomization
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

Instrumental-variable (IV) methods estimate causal effects when unmeasured confounders exist, but they require instruments that do not affect the outcome except through the treatment. This paper deals with the common situation of several treatments and some invalid instruments, where the classical single-treatment majority rule no longer identifies the effects. The authors show that identification holds under a 'generalized plurality rule' — the true effect vector must be supported by strictly more instrument-defined hyperplanes than any competing vector — and that a simpler 'generalized majority rule' on the count of valid instruments guarantees it. For finite samples, they propose a sampling confidence interval that perturbs the estimated instrument-validity measures, keeps only selections satisfying the majority rule, and unions the resulting two-stage least-squares intervals. If the paper is right, applied researchers can obtain honest confidence intervals for multiple causal effects without knowing which instruments are valid, even when some invalid instruments are nearly indistinguishable from valid ones.

Core claim

On the paper's own terms, the central claim is that for p_d ≥ 1 treatments and p_z instruments, the vector of causal effects is identified by the generalized plurality rule, which says the true effect β* receives votes from more instruments than any other candidate effect, where each instrument 'votes' for every candidate on its hyperplane L_k = {β : Υ_{k·}(β−β*) = π_k}. A sufficient, easy-to-check condition is the generalized majority rule: |V*| > (|S*| + p_d − 1)/2, where V* is the set of truly valid instruments and S* the set of relevant ones. For inference, the paper constructs a sampling confidence interval (SCI): for each just-identifying subset it estimates the validity of every instr

What carries the argument

The key geometric object is the hyperplane voting interpretation of instruments: in a multiple-treatment linear IV model, each relevant instrument defines a (p_d−1)-dimensional hyperplane in the p_d-dimensional effect space, and an instrument votes for any candidate effect lying on its hyperplane. The generalized plurality rule (Condition 2) requires β* to lie on more hyperplanes than any other candidate; Theorem 1 shows that the count condition |V*| > (|S*| + h_0 − 1)/2 suffices, where h_0 encodes the rank structure of the relevance matrix and reduces to p_d in the benign case. For inference, the mechanism is 'perturb-and-aggregate': Gaussian perturbations of the validity estimates generate

Load-bearing premise

The procedure's parametric-rate length guarantee rests on Assumption 3's inequality (39), which says that for every just-identifying subset with strongly invalid instruments, at most (|S*|+p_d−1)/2 instruments can have centering bias inside the local-to-zero window C_1 log(C_2/α0)/√n.

What would settle it

Rerun the paper's own simulation design (S1) with n=5000 and τ=0.05, where one instrument is locally invalid and hard thresholding undercovers badly: the sampling interval must cover the true effect at about the nominal 95% level while its length shrinks with n. If empirical coverage is materially below 95% or the interval length does not shrink at the 1/√n rate, Theorems 2 and 3 are falsified. Alternatively, construct a design satisfying (39) but with p_d=2, p_z=9, and 5 valid instruments, and check that the SCI length remains O(1/√n).

Watch this falsifier — get emailed when new claim-graph text bears on it.

If this is right

  • Applied researchers can report honest confidence intervals for multiple endogenous treatments without knowing which instruments are valid, as long as the valid instruments exceed the majority bound.
  • The method works with summary-level Mendelian randomization data, where only SNP-exposure and SNP-outcome coefficients and standard errors are available.
  • The SCI avoids a full grid search over the effect space by searching over just-identifying subsets, keeping the computation feasible for moderate dimensions.
  • The parametric-rate length guarantee means the procedure does not achieve coverage by inflating the interval width.
  • For a single treatment the generalized majority rule reduces to the classical majority rule, so the results unify single- and multi-treatment IV inference.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The paper leaves open whether replacing the Gaussian perturbation with a bootstrap or wild bootstrap would improve finite-sample coverage; that trade-off is not explored.
  • The same perturb-and-aggregate scheme could be adapted to other post-selection inference problems, such as confidence intervals after moment selection in GMM, wherever a plurality condition can be formulated.
  • If the local plurality condition (39) fails, the length guarantee breaks down; an adaptive widening of the SCI could restore honesty at the cost of conservatism.
  • Proposition 1 suggests that under random coefficients, the generalized plurality rule holds almost surely once |V*| > p_d, implying that the majority rule may be unnecessarily restrictive for identification.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit.

Referee Report

3 major / 4 minor

Summary. The paper studies the linear instrumental variable model with multiple endogenous treatments and possibly invalid instruments. It introduces generalized plurality and majority rules for identifying the vector of treatment effects, emphasizing the geometric distinction from the single-treatment case. For inference, it proposes a Sampling Confidence Interval (SCI) that aggregates TSLS intervals over perturbed estimates of instrument validity, with the aim of remaining valid even when hard-thresholding cannot separate locally invalid from valid instruments. The main theoretical results are Theorem 2 (asymptotic coverage of the SCI under a generalized majority condition) and Theorem 3 (parametric-rate length under Assumption 3). The paper also reports simulations and a Mendelian randomization application.

Significance. The identification results are a useful extension of single-treatment plurality logic to multiple treatments, with a clear geometric interpretation and a transparent sufficient condition. The sampling scheme is a nontrivial generalization of Guo (2023) that avoids multidimensional grid search, and the supplementary proofs are detailed. However, the advertised parametric-rate length guarantee rests entirely on Assumption 3, Eq. (39), a high-level local-plurality condition that is neither derived from primitive conditions nor verifiable from data. In addition, the adaptive tuning of C0 in Remark 4 does not match the fixed constant used in the proof of Proposition 3. These issues prevent the paper from fully establishing its central efficiency claim, although the coverage theorem is more robust.

major comments (3)
  1. [Section 4, Assumption 3, Eq. (39); Theorem 3] The parametric-rate length guarantee (41) is obtained only under a high-level condition not implied by the identification conditions. A concrete failure: |S*|=10, p_d=2, |V*|=6, so the generalized majority rule (16) holds. Let five invalid IVs pass exactly through a false effect beta_f, and two further invalid IVs have |pi_k^(ell)| <= C1 log(C2/alpha0)/sqrt(n) for the subset ell at beta_f. Then Condition 2 holds, but the left side of (39) is at least 7 > 5.5, so Assumption 3 fails. The corresponding TSLS interval is then included in the SCI and is centered near beta_f, so Len(alpha)=O(1), not O(log/sqrt(n)). Thus the length claim is not a byproduct of identification; it is an extra, untestable local-plurality assumption. Please derive (39) from primitives or reframe the length guarantee as conditional.
  2. [Section 3.3, Remark 4 and proof of Proposition 3] The proof sets a specific constant C0 in rho_n(M) that depends on alpha0 and the dimension, but Algorithm 1 uses an adaptive rule: start at 0.05 and multiply by 1.25 until more than 5% of the generalized-majority selections are nonempty. No argument shows that this adaptive choice preserves the event in Proposition 3 or the coverage/length bounds. The simulations and application implement the adaptive rule, so the reported performance is not the procedure for which theorems are proved. Either prove guarantees for the adaptive C0 or fix C0 as in the proof and report sensitivity.
  3. [Section 4, Assumption 2 and Remark 5] Assumption 2(b) requires every just-identifying subset H in H* to have eigenvalue lambda_min(Upsilon_H^T Upsilon_H) bounded away from zero. This is much stronger than Condition 1, which only requires the full valid set to have full column rank. In Mendelian randomization, valid IVs are often weak individually, so there may exist just-identifying valid subsets with very small eigenvalues under Condition 1. Under such subsets, the TSLS estimator is not sqrt(n)-consistent, and Proposition 2, Proposition 3, and Theorem 2 are unavailable. Remark 5 acknowledges weak identification is out of scope, but the abstract and title do not state this restriction. It should be stated prominently.
minor comments (4)
  1. [Section 3.3, Eq. (28)] The symbol M is used both for the number of resamples and for the filtered set in (28). Use a distinct notation such as N to avoid confusion.
  2. [Section 3.2, Eq. (27)] The definition of locally invalid IVs uses the TSHT candidate set Lhat from (26), but the sampling procedure later uses different candidate sets bT_m. Clarify whether the concept is relative to TSHT or to the SCI procedure.
  3. [Section 5-6] The method enumerates all subsets of size p_d from S_hat. For small p_z this is fine, but the paper does not discuss computational feasibility for larger p_z. A brief complexity discussion would be helpful.
  4. [Section 4, Assumption 3] The constants C1 and C2 in (38)-(39) are introduced as absolute constants but never specified or connected to primitive parameters. The theorem would be more informative if these constants were made explicit or shown to depend only on identifiable quantities.

Circularity Check

0 steps flagged

No significant circularity: the main theorems are derived from stated assumptions and standard asymptotic arguments; the high-level Assumption 3 is an extra regularity condition, not a restatement of the conclusion.

full rationale

The paper's central claims are not circular. Theorems 2 and 3 are proved from Proposition 3 plus standard TSLS asymptotics and union-bound arguments. Proposition 3 itself is a genuine probabilistic result: it shows that by sampling normal perturbations of the estimated validity vector, with high probability one perturbation lands within a shrinking neighborhood of the true π*, so the corresponding selected IV set contains all valid IVs. This does not assume the coverage or length conclusion. The proof of Theorem 2 then shows that, conditional on this event, at least one TSLS interval in the union SCI(α) is valid, and hence the union covers β*1; this is standard union-inference logic, not a tautology. Theorem 3's parametric-rate length bound does rely on Assumption 3, especially inequality (39), but the paper transparently labels this as an additional assumption rather than deriving it from the conclusion. Indeed, the paper states that (40) is 'slightly stronger than the generalized plurality rule in Condition 2, accounting for the variable selection error in finite samples.' Whether (39) is primitive, verifiable, or too strong is a legitimate robustness/assumption-strength concern, not evidence of circularity. The self-citations to Guo et al. (2018) and Guo (2023) are used as prior baselines, comparisons, and tuning heuristics; the identification theorems and the inference proofs do not reduce to these citations. In particular, the sampling idea is generalized from Guo (2023), but the coverage and length guarantees are proved here directly from the paper's own conditions, not imported from Guo (2023). No fitted parameter is renamed as a prediction, and no uniqueness theorem from the authors' prior work is invoked to force the choice of procedure. The real-data application and simulations are external checks rather than part of the derivation. Therefore there is no specific reduction of a claimed result to its own inputs, and the appropriate circularity score is 0.

Axiom & Free-Parameter Ledger

4 free parameters · 7 axioms · 0 invented entities

The paper does not introduce new physical entities; 'locally invalid IV' is a definitional category, not a postulated entity. The extrapolation burden is carried by the linear constant-effects model, the strong-relevance assumptions, and the high-level Assumption 3 for the rate guarantee.

free parameters (4)
  • C0 (ρ_n(M) scaling) = Theory: 0.5·(2/c(α0))^{1/|S*|}; implementation: start 0.05, multiply by 1.25 until >5% nonempty (Remark 4)
    Controls the threshold width in (30). The coverage/length theorems are proven for a fixed C0 of that form; the adaptive rule used in practice is not covered by the proofs.
  • α0 (screening level) = α/20 (Remark 4)
    Tuning parameter in (28) used to truncate perturbations and to adjust the significance level to α−α0; affects the coverage/length trade-off.
  • M (number of resamples) = 1000 (Remark 4)
    Number of perturbation draws; theory sends M→∞, the finite choice affects finite-sample performance.
  • C* (TSHT threshold constant) = 0.5 in the main simulations
    Used in the benchmark TSHT selection (25); not part of the proposed SCI but needed for the comparisons in Figures 2–5.
axioms (7)
  • domain assumption Linear, additive, constant-effect potential outcome model (1): Y(d,z,x)−Y(d',z',x') = (d−d')'β* + (x−x')'ξ* + (z−z')'κ*, with E(Y(0,0,0)|Z,X) = Z'η*+X'ζ*.
    Section 2, model (1). The entire identification and inference framework is linear and homogeneous; violations (effect heterogeneity, nonlinearity, interactions) are outside the paper's scope.
  • domain assumption Assumption 1: i.i.d. sub-Gaussian (W_i, v_i) with bounded eigenvalue spectra.
    Section 4, Assumption 1. Needed for concentration and CLT in Propositions 2–3.
  • domain assumption Assumption 2: strong relevance: max_j |Υ*_k,j| > c_Υ for every k∈S*, and min_{H∈H*} λ_min(Υ_H*'Υ_H*) > c_Υ.
    Section 4, Assumption 2. Rules out weak IVs and weak just-identifying subsets; without it the first-stage selection (18)–(20) is not consistent. This is a strong assumption in MR practice.
  • domain assumption Condition 1: rank(Υ*_{V*·}) = p_d.
    Section 2, Condition 1. Existence of a full-rank all-valid just-identifying subset; necessary for identification.
  • domain assumption Identification condition: Condition 2 (generalized plurality) or sufficient majority condition (15) |V*| > (|S*|+h0−1)/2.
    Section 2, Condition 2 and Theorem 1. Theorems 2–3 are stated under the majority version of this condition.
  • ad hoc to paper Assumption 3, Eq. (39): for each strongly invalid just-identifying subset ℓ, the number of IVs with |π^(ℓ)_k| ≤ C_1 log(C_2/α0)/√n is at most (|S*|+p_d−1)/2.
    Section 4, Eq. (39). High-level local-plurality condition introduced to guarantee the parametric-rate length (Theorem 3); not derived from primitives and not verifiable from data.
  • domain assumption Summary-data independence: SNPs independent, and bΓ and bΥ (and their entries) asymptotically independent (Section D).
    Used in the Mendelian randomization application only; acknowledged as a pragmatic approximation in Section D.

pith-pipeline@v1.3.0-alltime-deepseek · 35614 in / 17584 out tokens · 162906 ms · 2026-08-01T07:18:34.134645+00:00 · methodology

0 comments
read the original abstract

The instrumental variable (IV) method is widely used to infer causal effects in observational studies with unmeasured confounding, but invalid instruments can compromise both population identification and finite-sample inference. This paper studies linear IV models with multiple endogenous treatments and possibly invalid instruments. Identification is more delicate than in the single-treatment setting because a single instrument no longer identifies a scalar candidate effect; instead, each relevant instrument defines a hyperplane in the multidimensional effect space. For identification of multiple treatment effects, we introduce generalized plurality and majority rules which require a sufficiently large number of IVs to be valid. For inference, data-dependent instrument selection may fail to separate certain invalid IVs from valid ones, leading to undercoverage of confidence intervals when these invalid instruments are mistakenly selected as valid. We propose a sampling confidence interval for each treatment effect, which is robust to IV selection errors. We establish asymptotic coverage and parametric-rate length of our sampling confidence interval under regularity conditions and illustrate this method in a Mendelian randomization application.

Figures

Figures reproduced from arXiv: 2607.21481 by Qingliang Fan, Zijian Guo, Ziwei Mei.

Figure 1
Figure 1. Figure 1: Illustrative graphs for Identification Note: Each panel sketches an example with 5 IVs and 2 treatments. Each axis represents the value of one treatment effect. Each straight line represents a hyperplane defined in (10) of an IV. The hyperplanes of valid IVs are in blue color, while those of invalid ones are red. A vector β = (β1, β2) ⊤ receives a vote from an IV when it falls on the corresponding hyperpla… view at source ↗
Figure 2
Figure 2. Figure 2: Coverage probability of the 95% TSHT confidence interval with different viola [PITH_FULL_IMAGE:figures/full_fig_p019_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Coverage probability of the 95% confidence interval by sampling with [PITH_FULL_IMAGE:figures/full_fig_p023_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Coverage Probabilities and CI Lengths for [PITH_FULL_IMAGE:figures/full_fig_p031_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Coverage Probabilities and CI Lengths for [PITH_FULL_IMAGE:figures/full_fig_p032_5.png] view at source ↗

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

43 extracted references · 4 linked inside Pith

  1. [1]

    Journal of the Royal Statistical Society Series B: Statistical Methodology , volume=

    Confidence intervals for causal effects with invalid instruments by using two-stage hard thresholding with voting , author=. Journal of the Royal Statistical Society Series B: Statistical Methodology , volume=. 2018 , publisher=

    2018

  2. [2]

    2013 , publisher=

    Discovery and refinement of loci associated with lipid levels , journal=. 2013 , publisher=

    2013

  3. [3]

    Cell Genomics , volume=

    Deciphering proteins in Alzheimer’s disease: A new Mendelian randomization method integrated with AlphaFold3 for 3D structure prediction , author=. Cell Genomics , volume=. 2024 , publisher=

    2024

  4. [4]

    American journal of epidemiology , volume=

    Efficient design for Mendelian randomization studies: subsample and 2-sample instrumental variable estimators , author=. American journal of epidemiology , volume=. 2013 , publisher=

    2013

  5. [5]

    International journal of epidemiology , volume=

    Mendelian randomization with invalid instruments: effect estimation and bias detection through Egger regression , author=. International journal of epidemiology , volume=. 2015 , publisher=

    2015

  6. [6]

    Biometrika , volume=

    Using negative controls to identify causal effects with invalid instrumental variables , author=. Biometrika , volume=. 2025 , publisher=

    2025

  7. [7]

    Biometrika , volume=

    Semiparametric efficient G-estimation with invalid instrumental variables , author=. Biometrika , volume=. 2023 , publisher=

    2023

  8. [8]

    Journal of Machine Learning Research , volume=

    Selective machine learning of the average treatment effect with an invalid instrumental variable , author=. Journal of Machine Learning Research , volume=

  9. [9]

    Journal of the Royal Statistical Society Series B: Statistical Methodology , volume=

    GENIUS-MAWII: for robust Mendelian randomization with many weak invalid instruments , author=. Journal of the Royal Statistical Society Series B: Statistical Methodology , volume=. 2024 , publisher=

    2024

  10. [10]

    Journal of the American Statistical Association , pages=

    Robust inference for federated meta-learning , author=. Journal of the American Statistical Association , pages=. 2025 , publisher=

    2025

  11. [11]

    Annual Review of Statistics and Its Application , volume=

    Identification and inference with invalid instruments , author=. Annual Review of Statistics and Its Application , volume=. 2024 , publisher=

    2024

  12. [12]

    Robust Mendelian Randomization Analysis by Automatically Selecting Valid Genetic Instruments with Applications to Identify Plasma Protein Biomarkers for Alzheimer’s Disease , author=

  13. [13]

    Journal of Applied Econometrics , year=

    Agglomerative hierarchical clustering for selecting valid instrumental variables , author=. Journal of Applied Econometrics , year=

  14. [14]

    Journal of the Royal Statistical Society Series B: Statistical Methodology , pages=

    On the instrumental variable estimation with many weak and invalid instruments , author=. Journal of the Royal Statistical Society Series B: Statistical Methodology , pages=. 2024 , publisher=

    2024

  15. [15]

    Journal of Business & Economic Statistics , volume=

    Using heteroscedasticity to identify and estimate mismeasured and endogenous regressor models , author=. Journal of Business & Economic Statistics , volume=. 2012 , publisher=

    2012

  16. [16]

    Proceedings of the National Academy of Sciences , volume=

    Mendelian randomization for causal inference accounting for pleiotropy and sample structure using genome-wide summary statistics , author=. Proceedings of the National Academy of Sciences , volume=. 2022 , publisher=

    2022

  17. [17]

    Journal of Econometrics , volume=

    Select the valid and relevant moments: An information-based LASSO for GMM with many moments , author=. Journal of Econometrics , volume=. 2015 , publisher=

    2015

  18. [18]

    arXiv preprint arXiv:2310.08063 , year=

    Uniform Inference for Nonlinear Endogenous Treatment Effects with High-Dimensional Covariates , author=. arXiv preprint arXiv:2310.08063 , year=

    Pith/arXiv arXiv

  19. [19]

    Journal of Business & Economic Statistics , volume=

    A heteroscedasticity-robust overidentifying restriction test with high-dimensional covariates , author=. Journal of Business & Economic Statistics , volume=. 2025 , publisher=

    2025

  20. [20]

    Journal of the Royal Statistical Society Series B: Statistical Methodology , volume=

    The confidence interval method for selecting valid instrumental variables , author=. Journal of the Royal Statistical Society Series B: Statistical Methodology , volume=. 2021 , publisher=

    2021

  21. [21]

    Review of Economics and Statistics , volume=

    Endogenous treatment effect estimation with a large and mixed set of instruments and control variables , author=. Review of Economics and Statistics , volume=. 2024 , publisher=

    2024

  22. [22]

    Journal of the American statistical Association , volume=

    Instrumental variables estimation with some invalid instruments and its application to Mendelian randomization , author=. Journal of the American statistical Association , volume=. 2016 , publisher=

    2016

  23. [23]

    2005 , publisher=

    Matrix Algebra , author=. 2005 , publisher=

    2005

  24. [24]

    2012 , publisher=

    Matrix Analysis , author=. 2012 , publisher=

    2012

  25. [25]

    Econometrica , volume=

    Consistent moment selection procedures for generalized method of moments estimation , author=. Econometrica , volume=. 1999 , publisher=

    1999

  26. [26]

    Journal of the American Statistical Association , volume=

    Sensitivity analysis for instrumental variables regression with overidentifying restrictions , author=. Journal of the American Statistical Association , volume=. 2007 , publisher=

    2007

  27. [27]

    arXiv preprint arXiv:2301.00718 , year=

    Robust inference for federated meta-learning , author=. arXiv preprint arXiv:2301.00718 , year=

    Pith/arXiv arXiv

  28. [28]

    Journal of the American Statistical Association , pages=

    Statistical inference for maximin effects: Identifying stable associations across multiple studies , author=. Journal of the American Statistical Association , pages=. 2023 , publisher=

    2023

  29. [29]

    Journal of the Royal Statistical Society Series B: Statistical Methodology , volume=

    Post-selection problems and a solution using searching and sampling , author=. Journal of the Royal Statistical Society Series B: Statistical Methodology , volume=. 2023 , publisher=

    2023

  30. [30]

    Journal of Econometrics , volume=

    The validity of instruments revisited , author=. Journal of Econometrics , volume=. 2012 , publisher=

    2012

  31. [31]

    Quantitative Economics , volume=

    Simple and honest confidence intervals in nonparametric regression , author=. Quantitative Economics , volume=. 2020 , publisher=

    2020

  32. [32]

    Statistical Science , volume=

    The GENIUS approach to robust Mendelian randomization inference , author=. Statistical Science , volume=. 2021 , publisher=

    2021

  33. [33]

    arXiv preprint arXiv:2012.14823 , year=

    Bias-aware inference in regularized regression models , author=. arXiv preprint arXiv:2012.14823 , year=

    Pith/arXiv arXiv 2012

  34. [34]

    Quantitative Economics , volume=

    Sensitivity analysis using approximate moment condition models , author=. Quantitative Economics , volume=. 2021 , publisher=

    2021

  35. [35]

    Journal of Econometrics , pages=

    Testing underidentification in linear models, with applications to dynamic panel and asset pricing models , author=. Journal of Econometrics , pages=. 2021 , publisher=

    2021

  36. [36]

    Econometric theory , volume=

    Testing identifiability and specification in instrumental variable models , author=. Econometric theory , volume=. 1993 , publisher=

    1993

  37. [37]

    Journal of the American Statistical Association , volume=

    On the use of the lasso for instrumental variables estimation with some invalid instruments , author=. Journal of the American Statistical Association , volume=. 2019 , publisher=

    2019

  38. [38]

    arXiv preprint arXiv:2208.05278 , year=

    Selecting Valid Instrumental Variables in Linear Models with Multiple Exposure Variables: Adaptive Lasso and the Median-of-Medians Estimator , author=. arXiv preprint arXiv:2208.05278 , year=

    Pith/arXiv arXiv

  39. [39]

    Detecting Invalid Instruments using L_1 -

    Han, Chirok , journal=. Detecting Invalid Instruments using L_1 -

  40. [40]

    American Journal of Epidemiology , volume=

    Multivariable Mendelian Randomization: The Use of Pleiotropic Genetic Variants to Estimate Causal Effects , author=. American Journal of Epidemiology , volume=. 2015 , publisher=

    2015

  41. [41]

    Structural variation in amyloid-

    Qiang, Wei and Yau, Wai-Ming and Lu, Jun-Xia and Collinge, John and Tycko, Robert , journal=. Structural variation in amyloid-. 2017 , publisher=

    2017

  42. [42]

    Nature Genetics , volume=

    A unified framework for joint-tissue transcriptome-wide association and Mendelian randomization analysis , author=. Nature Genetics , volume=. 2020 , publisher=

    2020

  43. [43]

    American Economic Review , pages=

    Contamination Bias in Linear Regressions , author=. American Economic Review , pages=. 2024 , publisher=

    2024