01 / Understand the biological system
Begin with the ecology, taxonomy and social organisation of eusocial snapping shrimp. Review how colonies are sampled, how reproductive roles are identified and how alternative explanations are tested. Comparative work must preserve differences among species, habitats and studies.
Near-term output: an annotated literature map with explicit inclusion criteria and an evidence register. Proposed field research would require qualified partners, permits and welfare review before it began.
02 / Build models that can be questioned
Investigate network science, distributed algorithms, game theory and optimisation as modelling languages. Each model should define agents, interactions, uncertainty and the quantities being measured. Mathematical elegance does not establish biological validity.
Near-term output: small, documented simulations with reproducible random seeds, sensitivity analysis and failure cases. Computational exercises must carry a simulation label and must never be reported as observations of shrimp colonies.
03 / Make participation consequential
Develop bounded tasks with learning material, consent, quality controls and a path to feedback. Candidate activities include literature annotation, reproducibility checks and classifications of appropriately licensed research material.
Participation records should distinguish submitted, reviewed, accepted and withdrawn contributions. Public totals should be generated from an auditable counting rule and eligible stored records, rather than page views or invented momentum.
04 / Establish accountable procedures
Create a clear publication taxonomy, document registry, correction route and complaints process. Appoint responsible people before making service commitments. Disclose conflicts and distinguish editorial judgement from empirical claims.
Institutional development is itself a research-adjacent responsibility. The procedure for creating a committee should be shorter than the eventual committee’s remit, where practicable.