yogsoth-ai Variable Identification
Identify all variables from a mechanism description and annotate their roles within the hypothesis structure.
HARD-GATE
Not met → stop, return error: mechanism description is insufficient to identify variables.
Pipeline
- Precondition check: verify mechanism chain completeness
- Variable extraction: enumerate all variables from every mechanism chain (including implicit variables)
- Role classification: assign a role to each variable (IV/DV/mediator/moderator/control/confound)
- Deduplication and merging: when the same variable appears in different mechanisms, merge it and annotate multiple roles
- Operationalizability assessment: evaluate whether each variable can be measured (high/medium/low/unclear)
- Output a structured variable list
Output Format
[
{
"name": "Variable name",
"role": "IV | DV | mediator | moderator | control | confound",
"description": "What this variable represents",
"source_mechanism": ["mechanism name(s) where this variable appears"],
"operationalizable": "high | medium | low | unclear",
"operationalization_notes": "How it might be measured or manipulated"
}
]
Identify at least 2 variables (1 IV + 1 DV).