Tests and Confidence Sets with Correct Size when Instruments are Potentially Weak

Mikusheva, Anna; Poi, Brian P.

doi:10.1177/1536867x0600600303

Cited by 88 publications

(59 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, as yet, there is no software for performing such analyses. Other alternative approaches have also been suggested recently, and one based on the conditional likelihood ratio has been programmed in the statistical software package Stata [123,124]. We would suggest that currently researchers using IV analyses in Mendelian randomization studies always report F-statistics from the first-stage regressions to examine the strength of the instrument and avoid making causal inferences, or seek advice from an IV expert where this is close to or less than 10.…”

Section: Weak Instruments In Mendelian Randomization Studiesmentioning

confidence: 99%

Mendelian randomization: Using genes as instruments for making causal inferences in epidemiology

Lawlor

Harbord

Sterne

et al. 2008

Statistics in Medicine

3,534

3,153

View full text Add to dashboard Cite

Observational epidemiological studies suffer from many potential biases, from confounding and from reverse causation, and this limits their ability to robustly identify causal associations. Several high-profile situations exist in which randomized controlled trials of precisely the same intervention that has been examined in observational studies have produced markedly different findings. In other observational sciences, the use of instrumental variable (IV) approaches has been one approach to strengthening causal inferences in non-experimental situations. The use of germline genetic variants that proxy for environmentally modifiable exposures as instruments for these exposures is one form of IV analysis that can be implemented within observational epidemiological studies. The method has been referred to as 'Mendelian randomization', and can be considered as analogous to randomized controlled trials. This paper outlines Mendelian randomization, draws parallels with IV methods, provides examples of implementation of the approach and discusses limitations of the approach and some methods for dealing with these.

show abstract

Section: Weak Instruments In Mendelian Randomization Studiesmentioning

confidence: 99%

Mendelian randomization: Using genes as instruments for making causal inferences in epidemiology

Lawlor

Harbord

Sterne

et al. 2008

Statistics in Medicine

3,534

3,153

View full text Add to dashboard Cite

show abstract

“…Had the rule-of-thumb been based on the proportional bias being less that 20% rather than 10%, our F-statistic would have not been deemed problematic. 28 The implementation of this procedure in STATA 11 is by Mikusheva and Poi (2001), and the associated command is condivreg. 29 While we have opted to run all our regressions using STATA's svy:reg command to account for survey design, as is appropriate for our data, the condivreg estimation (table 8) does not permit the use of svy or weights.…”

Section: Other Determinantsmentioning

confidence: 99%

Domestic violence and women's autonomy in developing countries: theory and evidence

Eswaran

Malhotra

2011

Canadian J of Economics

209

137

View full text Add to dashboard Cite

This paper sets out a simple non-cooperative model of resource allocation within the household in developing countries that incorporates domestic violence as a vehicle for enhancing bargaining power. We demonstrate that the extent of domestic violence faced by women is not necessarily declining in their reservation utilities, or necessarily increasing in their spouses'. Using the National Family Health Survey data of India for 1998-99, we isolate the effect of domestic violence on female autonomy, taking into account the possible endogeneity of domestic violence through the choice of appropriate instruments. We provide some evidence for the evolutionary theory of domestic violence, which argues that such violence stems from the jealousy caused by paternity uncertainty in our evolutionary past. The findings have strong policy implications suggesting that it will take more than an improvement in women's employment options to address the problem of spousal violence. JEL classification: J16, D7Violence conjugale et autonomie des femmes dans les pays en voie de développement : théorie et résultats. Ce mémoire met en place un modèle simple d'allocation de ressources sans collaborationà l'intérieur d'un ménage dans des pays en voie de développement où la violence conjugale est un véhicule utilisé pour accroître le pouvoir de marchandage. On montre que l'importance de la violence conjugale faite aux femmes ne décline pas nécessairement avec leurs exigences minimales, non plus qu'elle s'accroît avec les exigences minimales de leursépoux. A l'aide des résultats de l'Enquête nationale sur la santé des familles de l'Inde pour 1998-99, on précise l'effet de la violence conjugale sur l'autonomie des femmes, en tenant compte de l'endogénéité possible de la violence conjugaleà l'aide d'instruments appropriés. On fournit un certain supportà la théorieévolutionnaire de la Without implicating them, we would like to thank Domestic violence and women's autonomy 1223 violence conjugale, laquelle suggère que cette violenceémerge de la jalousie causée par l'incertitude de la paternité dans notre passéévolutionnaire. Les résultats ont de fortes implications pour la politique publique : ils suggèrent qu'il faudra beaucoup plus qu'une amélioration dans les opportunités d'emploi des femmes pour résoudre le problème de la violence conjugale.

show abstract

“… Note : The entries in the table show the Confidence Intervals around the parameter value 0 given by the Conditional Likelihood Ratio Test proposed by Moreira () as well as the Confidence Intervals based on the Anderson and Rubin statistic. These are computed using the condivreg command in Stata described in Mikusheva and Poi (). These confidence intervals are robust to the use of weak instruments.…”

Section: Resultsmentioning

confidence: 99%

The National Minimum Wage and the Substitutability Between Young and Old Workers in Low Paid Occupations

Lanot

Sousounis

2016

The Manchester School

View full text Add to dashboard Cite

We study the effect of the National Minimum Wage (NMW) on the workforce age composition within the low paying sectors of the British economy. Our interest is in the degree of substitutability between labour inputs (young and old employees). We find evidence that both the introduction and the regular upratings of the NMW have a significant effect on the observed changes to the relative wages and to the relative wage bills but not to relative employment. We estimate the elasticity of substitution, between ‘young’ age groups and older workers (55+) to be zero, while that of ‘prime’ (22+) age and older workers, to be around 0.79. Our estimates therefore imply significant complementarity between younger and old employees.

show abstract

Tests and Confidence Sets with Correct Size when Instruments are Potentially Weak

Cited by 88 publications

References 13 publications

Mendelian randomization: Using genes as instruments for making causal inferences in epidemiology

Mendelian randomization: Using genes as instruments for making causal inferences in epidemiology

Domestic violence and women's autonomy in developing countries: theory and evidence

The National Minimum Wage and the Substitutability Between Young and Old Workers in Low Paid Occupations

Contact Info

Product

Resources

About