abstract: Mediation analysis is used to investigate how an independent variable, X, is related to an outcome variable, Y, through a mediator variable, M (MacKinnon, 2008). If X represents a randomized intervention it is difficult to make a cause and effect inference regarding indirect effects without making no unmeasured confounding assumptions using the potential outcomes framework (Holland, 1988; MacKinnon, 2008; Robins & Greenland, 1992; VanderWeele, 2015), using longitudinal data to determine the temporal order of M and Y (MacKinnon, 2008), or both. The goals of this dissertation were to (1) define all indirect and direct effects in a three-wave longitudinal mediation model using the causal mediation formula (Pearl, 2012), (2) analytically compare traditional estimators (ANCOVA, difference score, and residualized change score) to the potential outcomes-defined indirect effects, and (3) use a Monte Carlo simulation to compare the performance of regression and potential outcomes-based methods for estimating longitudinal indirect effects and apply the methods to an empirical dataset. The results of the causal mediation formula revealed the potential outcomes definitions of indirect effects are equivalent to the product of coefficient estimators in a three-wave longitudinal mediation model with linear and additive relations. It was demonstrated with analytical comparisons that the ANCOVA, difference score, and residualized change score models’ estimates of two time-specific indirect effects differ as a function of the respective mediator-outcome relations at each time point. The traditional model that performed the best in terms of the evaluation criteria in the Monte Carlo study was the ANCOVA model and the potential outcomes model that performed the best in terms of the evaluation criteria was sequential G-estimation. Implications and future directions are discussed. / Dissertation/Thesis / Doctoral Dissertation Psychology 2018

abstract: In investigating mediating processes, researchers usually use randomized experiments and linear regression or structural equation modeling to determine if the treatment affects the hypothesized mediator and if the mediator affects the targeted outcome. However, randomizing the treatment will not yield accurate causal path estimates unless certain assumptions are satisfied. Since randomization of the mediator may not be plausible for most studies (i.e., the mediator status is not randomly assigned, but self-selected by participants), both the direct and indirect effects may be biased by confounding variables. The purpose of this dissertation is (1) to investigate the extent to which traditional mediation methods are affected by confounding variables and (2) to assess the statistical performance of several modern methods to address confounding variable effects in mediation analysis. This dissertation first reviewed the theoretical foundations of causal inference in statistical mediation analysis, modern statistical analysis for causal inference, and then described different methods to estimate causal direct and indirect effects in the presence of two post-treatment confounders. A large simulation study was designed to evaluate the extent to which ordinary regression and modern causal inference methods are able to obtain correct estimates of the direct and indirect effects when confounding variables that are present in the population are not included in the analysis. Five methods were compared in terms of bias, relative bias, mean square error, statistical power, Type I error rates, and confidence interval coverage to test how robust the methods are to the violation of the no unmeasured confounders assumption and confounder effect sizes. The methods explored were linear regression with adjustment, inverse propensity weighting, inverse propensity weighting with truncated weights, sequential g-estimation, and a doubly robust sequential g-estimation. Results showed that in estimating the direct and indirect effects, in general, sequential g-estimation performed the best in terms of bias, Type I error rates, power, and coverage across different confounder effect, direct effect, and sample sizes when all confounders were included in the estimation. When one of the two confounders were omitted from the estimation process, in general, none of the methods had acceptable relative bias in the simulation study. Omitting one of the confounders from estimation corresponded to the common case in mediation studies where no measure of a confounder is available but a confounder may affect the analysis. Failing to measure potential post-treatment confounder variables in a mediation model leads to biased estimates regardless of the analysis method used and emphasizes the importance of sensitivity analysis for causal mediation analysis. / Dissertation/Thesis / Ph.D. Psychology 2013

Background Studies based on a cohort design requires access to both subject-specific and period-specific information. In order to conduct an occupational cohort study, access to exposure information and the possibility and permission to link information on outcomes from other registers are generally necessary. The analysis phase is also aggravated by its added complexity because of the longitudinal dimension of the cohort’s data.This thesis aims at increasing the knowledge on hazards from work on fatalities and cancer within the domain of cohort studies on miners and metal refiners and to study the complexity of the analysis by discussing and suggesting analytical strategies. Methods The study population for this thesis consisted of a cohort of 2264 blue-collar aluminium smelter workers (paper I) and a cohort of 13000 blue-collar iron-ore miners (papers II-IV), both followed for over 50 years. The outcomes were collected from the Swedish Cause of Death Register and the Swedish Cancer Register. The primary methods of analysis were either Standardized Morbidity Ratios (SMR) or internal comparisons based on Cox or Poisson regression modeling. In paper IV, a g-estimation based on an accelerated failure-time model was performed to estimate the survival ratio. Results The results from paper I suggested that working as a blue-collar worker metal refiner was associated with increased rates of incidental lung cancer. Elevated rates among short term workers were observed for several outcomes. Paper I also showed that the choice of reference population when calculating SMR could influence the conclusions of the results. In paper II, several outcomes were elevated among the miners compared to the reference population from northern Sweden. However, no outcome except lung cancer was associated with cumulative employment time. The most recurrent pattern of the results was the negative association between cumulative employment time underground and several outcomes. The results from paper III showed that cumulative employment time working outdoors was associated with increased rates of cerebrovascular disease mortality. However, employment with heavy physical workloads did not explain the previously observed decreasing rates in the selected groups of outcomes. The adjustment for the healthy worker survivor effect by g-estimation in paper IV suggested that exposure from respirable dust was associated with elevated mortality risks that could not be observed with standard analytical methods. Conclusion Our studies found several rates from the cohorts that were elevated compared to external refererence populations but also that long term employments generally were associated with decreasing rates. Furthermore, incidental lung cancer rates was found elevated for the metal refiners. Among the miners, mortality rates of cerebrovascular diseases depended on if work was performed outdoor (higher rates) or underground (lower rates). Methodologically, this thesis has discussed different analytical strategies for handling confounding in occupational cohort studies. Paper IV showed that the healthy worker survivor effect could be adjusted for by performing g-estimation.