Previous research argues the Seventeenth Amendment made Senate elections more responsive. To make this claim, existing work compares the vote-seat relationships of direct and indirect elections before and after the Seventeenth Amendment. I argue this approach is problematic because it does not account for regional variation and compares elections from different time periods using presidential instead of Senate vote. I overcome these problems by simulating indirect elections using state legislatures’ partisan compositions to evaluate the responsiveness of direct and indirect elections after the Seventeenth Amendment. With this counterfactual approach, my findings suggest direct elections are not necessary for electoral responsiveness.