We present an approach to investigating causal mechanisms in survey experiments that leverages the provision or withholding of information on potentially important mediating variables. The designs we propose can identify the controlled direct effect of a treatment and also what we call an intervention effect. These quantities can be used to address substantive questions about causal mechanisms, can be identified under weaker assumptions than current approaches to causal mechanisms, and can be estimated with simple estimators using standard statistical software. Furthermore, these methods are compatible with a broad range of experimental designs, including survey vignettes and conjoint designs. We illustrate the approach via two examples, one on evaluations of potential U.S. Supreme Court nominees and the other on public perceptions of the democratic peace.