What problem does it solve?
This skill prevents cognitive scientists from interpreting fitted model parameters that may not be identifiable by guiding them through comprehensive parameter recovery and model recovery diagnostics matched to their actual experimental design, sample size, and optimization pipeline.
It highlights common failure modes (non-identifiability, insufficient trials, local minima, misspecification) and reminds researchers to verify assumptions before trusting any numerical estimates.
Core Features & Use Cases
- Structured planning protocol: Prompts researchers to state their question, justify method choices, declare expected outcomes, and note assumptions before running recovery simulations, ensuring thoughtful readiness.
- End-to-end recovery workflow: Covers grid or space-filling sampling, simulation with real design constraints, identical fitting procedures with multiple starts, and evaluation via correlations, bias, RMSE, coverage, and visualization templates from the references.
- Model recovery and reporting guidance: Details confusion-matrix construction, trial-count sweeps, objective landscapes, and a reporting checklist so reviewers can confirm parameters and models are trustworthy.
- Use case: Before publishing a drift-diffusion analysis, use this skill to simulate the full range of drift rates, boundaries, and noise levels your experiment might produce and ensure fitted parameters recover reliably for the collected trials.
Quick Start
Ask the skill whether your model's parameters can be reliably recovered given your planned experimental design.
