Research design & hypotheses
Use existing evidence to define research questions, measurable variables, experimental conditions, and an analysis plan that can be reviewed and implemented.
Inputs
Background, prior findings, available data, constraints, and hypotheses to test.
Deliverables
Research questions, variable definitions, experiment outline, analysis plan, and testable hypotheses.
Step-by-step workflow
Break the work into verifiable stages, then choose models and supporting tools.
- 01
Break down the research question
Separate the phenomenon, existing evidence, and uncertain hypotheses into a question tree.
- 02
Design variables and controls
List measurement methods, controls, and potential confounders, explaining each choice.
- 03
Evaluate the analysis plan
Generate power analysis or simulation code, run it in statistical software, and retain parameters.
- 04
Prepare a reviewable plan
Document procedures, data fields, exclusion rules, and failure criteria for researcher review before execution.
Required model capabilities
Combine models for the actual stages. The directory includes candidates matching one or more of these capabilities.
View matching modelsSupporting tools
Selection & delivery checks
- Prioritize traceable reasoning and explicit assumptions.
- Determine sample sizes through effect sizes, test methods, and power calculations rather than model guesses.
- Record exploratory work separately from pre-specified confirmatory analysis.
An example request
Derive testable hypotheses from existing research, design a controlled experiment, simulate statistical power in Python, and produce a protocol and analysis plan.Find models for this request