CogViz Research Studio

A planner for psychology experiments. Compare research designs, plan the analysis, see what your prediction implies and export the whole plan.

Hosted free on Streamlit. If nobody has used it for a while, it can take about a minute to wake up.

Simulation-based power: the share of simulated studies of each size that detect an effect of d = 0.5. An illustration of one CogViz tool.

Five steps from question to plan

A check mark appears on each step once it holds what the next step needs.

  1. QuestionState what you want to find out. A worked example fills every field if you want to see one first.
  2. DesignCompare designs side by side, with what each gains and costs.
  3. AnalysisPlan the test, and check it fits your variables.
  4. PredictionsSimulate data from your predicted means and estimate power.
  5. PlanExport a Markdown plan and a preregistration draft.

What's inside

  • Power by simulationFrom predicted means, a spread and a sample size, CogViz simulates 1,000 studies and reports how often the test reaches p < .05 (Lakens & Caldwell, 2021).
  • Preregistration draftThe plan follows the section order of the OSF Preregistration template, with every missing field marked "To complete".
  • Sources you can checkEach built-in design, analysis and model cites the work it comes from, in APA 7 style with DOI links.
  • A model library21 models of perception, choice, response time, effort, learning and memory, including signal detection and the drift-diffusion model.
  • A plan checkEach step flags gaps and mismatches, such as a binary outcome paired with a t-test.
  • Your files, not a server'sSave the study as a small file and open it later; the server stores nothing.

Share it