Infrastructure · Data Structures
Correlation
Normalized statistical association; no causal conclusion by itself
Definition
Mechanism
A normalized measure of statistical association between two Variables under a selected correlation method. The coefficient lies in [-1, 1] and is interpreted according to that method. Correlation alone neither establishes nor rules out Causation, its direction, or an intervention effect.
Invariants
- The coefficient lies in [-1, 1].
- The correlation alone establishes neither the presence nor absence of Causation, its direction, nor an intervention effect, regardless of coefficient magnitude.
Risks
Failure modes
Causal overreach
the result alone is used to assert a causal relation, its absence or direction, or an intervention effect.
Source ambiguity
direct causation, reverse causation, confounding, and selection can produce similar coefficients, which correlation alone cannot distinguish.
Method blindness
dependence outside the selected method's sensitivity yields a weak coefficient that is read as independence.
Relationships
Builds on
Used by
References
Society
Ingest only what is strictly necessary
Society
System health acting as an inhibitory emotion
Reference
Data schema
{
"type": "object",
"required": [
"variable_a",
"variable_b",
"coefficient"
],
"properties": {
"variable_a": {
"type": "string",
"description": "First variable ID"
},
"variable_b": {
"type": "string",
"description": "Second variable ID"
},
"coefficient": {
"type": "number",
"minimum": -1,
"maximum": 1,
"description": "Normalized correlation coefficient [-1, 1] under the selected method"
},
"p_value": {
"type": "number",
"description": "Statistical significance of the correlation"
},
"sample_size": {
"type": "integer",
"description": "Number of observations"
}
}
}