This paper presents identification, inference, and sensitivity analysis for causal mediation effects. The authors prove that under a particular version of the sequential ignorability assumption, the average causal mediation effect (ACME) is nonparametrically identified. They compare their identification assumption with those proposed in the literature and discuss practical implications. The popular estimator based on the linear structural equation model (LSEM) can be interpreted as an ACME estimator under the proposed assumption if additional parametric assumptions are satisfied. These assumptions can be easily relaxed within and outside of the LSEM framework, and the authors propose simple nonparametric estimation strategies. They also propose a new sensitivity analysis that can be easily implemented by applied researchers within the LSEM framework. This method directly evaluates the robustness of empirical findings to the possible existence of unmeasured pre-treatment variables that confound the relationship between the mediator and the outcome. The authors apply the proposed methods to a randomized experiment from political psychology and make easy-to-use software available to implement the proposed methods. Key words and phrases: Causal inference, causal mediation analysis, direct and indirect effects, linear structural equation models, sequential ignorability, unmeasured confounders.This paper presents identification, inference, and sensitivity analysis for causal mediation effects. The authors prove that under a particular version of the sequential ignorability assumption, the average causal mediation effect (ACME) is nonparametrically identified. They compare their identification assumption with those proposed in the literature and discuss practical implications. The popular estimator based on the linear structural equation model (LSEM) can be interpreted as an ACME estimator under the proposed assumption if additional parametric assumptions are satisfied. These assumptions can be easily relaxed within and outside of the LSEM framework, and the authors propose simple nonparametric estimation strategies. They also propose a new sensitivity analysis that can be easily implemented by applied researchers within the LSEM framework. This method directly evaluates the robustness of empirical findings to the possible existence of unmeasured pre-treatment variables that confound the relationship between the mediator and the outcome. The authors apply the proposed methods to a randomized experiment from political psychology and make easy-to-use software available to implement the proposed methods. Key words and phrases: Causal inference, causal mediation analysis, direct and indirect effects, linear structural equation models, sequential ignorability, unmeasured confounders.