This component contains the materials for our power simulations, as well as the power simulation results. All simulations were conducted via the University of Wisconsin-Madison [Center for High Throughput Computing](https://chtc.cs.wisc.edu/); our power analyses will therefore be difficult to computationally reproduce without access to a high-throughput cluster. The general workflow for our power analysis is as follows: 1. **Create a set of parameters for the power simulation**. These are created by `create_conditions.R` and stored in `conditions.csv`. (Note that the data header is omitted because this line cannot be there during the power simulations) 2. **Put all files on a CHTC submit node**. The submit node is a computer from which a user can deliver jobs to the computing cluster. 3. **Submit a set of jobs using a `.sub`file**. The job sets were either based on Rottman and Young Study 2, Study 3, or a modified study based on the statistical results of Study 3, but that asks what would happen if the number of stimuli were doubled (ie if we used all the stimuli from Study 2 and Study 3). 4. **The .sub file sends each combination of conditions from `conditions.csv` to a separate computer**. The computers take these sets of conditions and send them, using `runR.sh`, to an instance of R, which runs the set of conditions using an R script (either `s2_power.R`, `s3_power.R`, or `s3dbl_power.R`). Depending on the target of replication, these .R files also read in `s2_processed.csv` and `s3_processed.csv`, which contain processed versions of Rottman and Young's Study 2 and Study 3 datasets. 5. **Output the results of the simulation**. The R instance then outputs a `.csv` file, which corresponds to the results of a simulation study with the target set of parameters. These results are sent back to the submit node of the cluster. 6. **Download the simulation results from the submit node and aggregate them**. The file that does the aggregation is `check_sims.R`, which outputs a `.csv` file and graph.