Explanation in causal inference: developments in mediation and interaction

VanderWeele, Tyler J.

doi:10.1093/ije/dyw277

Cited by 765 publications

(1,265 citation statements)

References 12 publications

Supporting

Mentioning

1,261

Contrasting

Order By: Relevance

“…These pathways could be separate and distinct, but might also be interacting with one another in complex ways over time. Recent developments in mediation analysis rooted in a causal framework have made advances that enables the analysis of multiple mediating pathways and interactions (49). Future studies would benefit from applying these methods to datasets that include repeated measurements of life-course factors over time in order to disentangle the complex relation between social position, work exposures, and sick leave.…”

Section: Discussionmentioning

confidence: 99%

Adult social position and sick leave: the mediating effect of physical workload

Corbett

Gran²,

Kristensen³

et al. 2015

Scand J Work Environ Health

View full text Add to dashboard Cite

show abstract

Section: Discussionmentioning

confidence: 99%

Adult social position and sick leave: the mediating effect of physical workload

Corbett

Gran²,

Kristensen³

et al. 2015

Scand J Work Environ Health

View full text Add to dashboard Cite

show abstract

“…The surrogate then affects (or at least strongly correlates with) the outcome of interest. These estimators require that the surrogate satisfy a strong conditional independence assumption (see, e.g., Rosenbaum (1984); Begg and Leung (2000); Frangakis and Rubin (2002);VanderWeele (2015)). Athey et al (2016) overcome this limitation by observing multiple surrogates, each of which may not satisfy the required conditional independence.…”

Section: Literature Reviewmentioning

confidence: 99%

LATE with Mismeasured or Misspecified Treatment: An Application to Women's Empowerment in India

2017

View full text Add to dashboard Cite

We show that a local average treatment effect (LATE) can sometimes be point identified and consistently estimated when treatment is mismeasured, or when treatment is estimated using a possibly misspecified structural model. Our associated estimator, which we call Mismeasurement Robust LATE (MR-LATE), is based on differencing two mismeasures of treatment. In our empirical application, treatment is women's empowerment: whether a wife has significant control of household resources. Due to measurement difficulties and sharing of goods within a household, this treatment cannot be directly observed without error, and so must be estimated. Our outcomes are health indicators of family members. We first estimate a structural model to obtain the otherwise unobserved treatment indicator. Then, using changes in inheritance laws in India as an instrument, we apply our new MR-LATE estimator. We find that women's empowerment substantially decreases their probability of being anemic or underweight, and children's likelihood to suffer from cough, fever or diarrhea. We find no significant positive or negative effects on men's health. JEL Codes: D13, D11, D12, C31, I32.

show abstract

“…We think that a reasonable interpretation of this causal effect is the one expressed by Vanderweele [10] in his recent book which explicates potential outcomes approaches to the estimation of mediation and interaction. This is a counterfactual definition [11], but one that does not rely on the intervention specification that is required for interpreting causal effects within a potential outcomes framework.…”

Section: Motivating Examplesmentioning

confidence: 93%

“…The meaning of race depends on early SES and the meaning of early SES depends on race-they are inextricably linked. This concept would seem closest to the sufficient cause model of interaction and suggest the use of methods for assessing this "mechanistic" interaction [10]. As Vanderweele notes, this type of interaction focuses on "why and how" particular exposures affect outcomes in addition to the "for whom" and "to what extent" of other additive interaction approaches.…”

Section: Starting From the Theorymentioning

confidence: 99%

Commentary: on the application of potential outcomes-based methods to questions in social psychiatry and psychiatric epidemiology

Schwartz

2017

Soc Psychiatry Psychiatr Epidemiol

View full text Add to dashboard Cite

and VanderWeele [1] provide a clear rationale for choosing an additive model. Similarly, VanderWeele, Jackson, and Li [2] describe how marginal structural models can overcome obstacles in the use of longitudinal data to resolve the longstanding controversy about reverse causation in the relationship between religious service attendance and depression. In both instances, methods from the potential outcomes framework solve important technical problems in the estimation of causal effects in social and psychiatric epidemiology.These applications of potential outcomes methods entailed the movement of both the potential outcomes approach and intersectionality concepts out of the realms in which they were developed. Within epidemiology, methods grounded in the potential outcomes framework were explicated largely in the context of estimating the effectiveness of treatments for HIV. The causal effects in this context can be easily conceptualized as intervention effects. How does this interpretation of "causal effects" change when applied to social and psychiatric constructs? Similarly, how do we interpret concepts like "intersectionality", when they move from the rich qualitative observations in which they were developed, into a quantitative sphere that is often seen as atheoretical [3]?Intersectionality and religious service attendance as contested terrain for potential outcomesIn a potential outcomes approach, a causal effect is generally defined as the difference in outcome for the same individual under two different treatments; it represents what would happen if we performed a hypothetical intervention

show abstract

Explanation in causal inference: developments in mediation and interaction

Cited by 765 publications

References 12 publications

Adult social position and sick leave: the mediating effect of physical workload

Adult social position and sick leave: the mediating effect of physical workload

LATE with Mismeasured or Misspecified Treatment: An Application to Women's Empowerment in India

Commentary: on the application of potential outcomes-based methods to questions in social psychiatry and psychiatric epidemiology

Contact Info

Product

Resources

About