Simulationstudy
- non-hierarchical
- full drift diffusion model
- with 2 stimuli (2 conditions for drift rate)
- fix prior distributions
- sample 2000 parameter sets from prior distribution
- for each parameter set simulate 2 datasets: for 100 trials and for 500 trials (in sum twice 2000 datasets)
IMPORTANT: This study was run when the function still had another name. To run it with the current version, the change `wiener_full_lpdf` to `wiener_full_prec_lpdf` in `S1_fullddm.full.stan`.
To run the study, the following steps are needed:
0. Copy `setup.template.R` to `setup.local.R` and set your cmdstan directory (with the `wiener_full_lpdf` function and a properly configured `stanc3`)
1. Simulate the data: `S1_simulate.R`
2. Fit the data sets: `S1_sample.R`
3. Extract the results to see if convergence criteria were met: `S1_extract_results.R`. (set results directory each time)
4. Move invalid fits to new location (if they should be stored): `S1_move_invalid_fits.R` (adjust results directory and extension each time)
5. Return to 2., adjusting warmup and sampling iterations as you see fit. Continue until all of the fits have sufficient diagnostics.
6. Analyse data with `S1_analysis.R`