Online survey experiments are a useful method of testing social scientific theories. Of commonly used online convenience samples, Amazon’s Mechanical Turk platform is especially useful as a relatively inexpensive means of conducting experiments. In this research methods case, I discuss how I used Mechanical Turk to study the effects of news coverage on public opinion about public policy. I describe the practical steps taken to conduct the experiment, before outlining some important trade-offs and considerations scholars should consider when using Mechanical Turk, and for experimental research more broadly. I then outline some important lessons I learned from the experiment, which I hope might be useful for scholars interested in conducting similar research.