hypothesis-generation

To generate testable hypotheses from observations, you need a systematic workflow that transforms raw data into mechanistic explanations and competing research aims. This process ensures your hypotheses are clearly distinguishable and testable.

The best way to structure competing hypotheses for experiment design is to create 3-5 clearly distinguishable, mechanistic explanations for a single phenomenon. This allows you to outline concrete design considerations for initial experiments to test each one.

You create mechanistic explanations for research planning by systematically working through a structured workflow from understanding the phenomenon to generating structured explanations. This yields actionable research aims ready for phase gating.

Yes, you can use the RA workflow to export generated hypotheses to project_telos.md. The workflow integrates directly with RA project planning, allowing you to complete the generation steps and add the resulting hypotheses and plan to the file.

You should generate 3-5 competing hypotheses for initial experiment design. This range ensures variety in mechanistic explanations while maintaining clear distinguishability to effectively guide your initial experiments.