Mendelian randomization (MR) has been a popular method in genetic epidemiology to estimate the effect of an exposure on an outcome using genetic variants as instrumental variables (IV), with two‐sample summary‐data MR being the most popular. Unfortunately, instruments in MR studies are often weakly associated with the exposure, which can bias effect estimates and inflate Type I errors. In this work, we propose test statistics that are robust under weak‐instrument asymptotics by extending the Anderson–Rubin, Kleibergen, and the conditional likelihood ratio test in econometrics to two‐sample summary‐data MR. We also use the proposed Anderson–Rubin test to develop a point estimator and to detect invalid instruments. We conclude with a simulation and an empirical study and show that the proposed tests control size and have better power than existing methods with weak instruments.
- Award ID(s):
- 1654234
- PAR ID:
- 10142670
- Date Published:
- Journal Name:
- Annual Review of Economics
- Volume:
- 11
- Issue:
- 1
- ISSN:
- 1941-1383
- Page Range / eLocation ID:
- 727 to 753
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
more » « less
Abstract -
We discuss the ivcrc command, which implements an instrumental-variables (IV) estimator for the linear correlated random-coefficients model. The correlated random-coefficients model is a natural generalization of the standard linear IV model that allows for endogenous, multivalued treatments and unobserved heterogeneity in treatment effects. The estimator implemented by ivcrc uses recent semiparametric identification results that allow for flexible functional forms and permit instruments that may be binary, discrete, or continuous. The ivcrc command also allows for the estimation of varying-coefficient regressions, which are closely related in structure to the proposed IV estimator. We illustrate the use of ivcrc by estimating the returns to education in the National Longitudinal Survey of Young Men.more » « less
-
The Survey of Attitudes Toward Statistics (SATS) is a widely used family of instruments for measuring attitude constructs in statistics education. Since the development of the SATS instruments, there has been an evolution in the understanding of validity in the field of educational measurement emphasizing validation as an on-going process. While a 2012 review of statistics education attitude instruments noted that the SATS family had the most validity evidence, two types of challenges to the use of these instruments have emerged: challenges to the interpretations of scale scores and challenges using the SATS instruments in populations other than undergraduate students enrolled in introductory statistics courses. A synthesis of the literature and empirical results are used to document these challenges.more » « less
-
more » « less
Summary A major challenge in instrumental variable (IV) analysis is to find instruments that are valid, or have no direct effect on the outcome and are ignorable. Typically one is unsure whether all of the putative IVs are in fact valid. We propose a general inference procedure in the presence of invalid IVs, called two-stage hard thresholding with voting. The procedure uses two hard thresholding steps to select strong instruments and to generate candidate sets of valid IVs. Voting takes the candidate sets and uses majority and plurality rules to determine the true set of valid IVs. In low dimensions with invalid instruments, our proposal correctly selects valid IVs, consistently estimates the causal effect, produces valid confidence intervals for the causal effect and has oracle optimal width, even if the so-called 50% rule or the majority rule is violated. In high dimensions, we establish nearly identical results without oracle optimality. In simulations, our proposal outperforms traditional and recent methods in the invalid IV literature. We also apply our method to reanalyse the causal effect of education on earnings.
-
more » « less
ABSTRACT The elemental abundances in the broad-line regions of high-redshift quasars trace the chemical evolution in the nuclear regions of massive galaxies in the early Universe. In this work, we study metallicity-sensitive broad emission-line flux ratios in rest-frame UV spectra of 25 high-redshift (5.8 < z < 7.5) quasars observed with the VLT/X-shooter and Gemini/GNIRS instruments, ranging over $\log \left({{M}_{\rm {BH}}/\rm {M}_{\odot }}\right) = 8.4-9.8$ in black hole mass and $\log \left(\rm {L}_{\rm {bol}}/\rm {erg \, s}^{-1}\right) = 46.7-47.7$ in bolometric luminosity. We fit individual spectra and composites generated by binning across quasar properties: bolometric luminosity, black hole mass, and blueshift of the C iv line, finding no redshift evolution in the emission-line ratios by comparing our high-redshift quasars to lower redshift (2.0 < z < 5.0) results presented in the literature. Using cloudy-based locally optimally emitting cloud photoionization model relations between metallicity and emission-line flux ratios, we find the observable properties of the broad emission lines to be consistent with emission from gas clouds with metallicity that are at least 2–4 times solar. Our high-redshift measurements also confirm that the blueshift of the C iv emission line is correlated with its equivalent width, which influences line ratios normalized against C iv. When accounting for the C iv blueshift, we find that the rest-frame UV emission-line flux ratios do not correlate appreciably with the black hole mass or bolometric luminosity.
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Journal Article:
- https://doi.org/10.1146/annurev-economics-080218-025643
