FutureIQ KnowledgeBase

Dose-Response Relationship

principle

A causal claim becomes more credible when effects rise and fall with exposure. A graded response helps separate a plausible cause from a correlation, while also revealing thresholds below which an exposure may have no meaningful effect.

Near toll booths, birth effects were stronger. With every step farther away they weakened—then vanished beyond roughly two kilometres. Geography behaved like a dosage dial.

E1

When exposure turns the effect dial

A dose-response pattern adds structure to correlation: more of the suspected cause accompanies more of the outcome, while less accompanies less. That gradient is what a real causal-mechanism often predicts. Distance from a pollution source can reduce exposure; reduced exposure should then reduce harm. The sequence makes competing stories less comfortable because they must explain not merely why two things coincide, but why the effect changes systematically with dose.

The same logic also prevents the opposite mistake: treating the mere presence of a hazard as proof of meaningful danger. Radioactivity can be detectable in steel yet too slight to matter biologically. Dose, sensitivity and the threshold of consequence all sit between “present” and “harmful.”

E1 E2

Where it shows up

The toll-booth gradient

Stronger birth effects near toll booths, weaker ones farther away and none beyond about two kilometres form a spatial dose-response curve. The pattern resembles a natural-experiment because location supplies comparison groups with differing exposure.

E1

Radioactive, but not consequential

Steel’s radiation registers on unusually sensitive instruments, yet the reported amount is too low to make a practical difference. Detection answers whether something exists; dose-response asks whether enough exists to produce an effect.

E2

A gradient strengthens a case; it does not close it

Dose-response is supporting evidence, not automatic proof. Distance could travel with other differences, and selecting observations through a shared gate can introduce collider-bias. Some effects also have thresholds: below them, smaller doses may produce no measurable consequence rather than a proportionally smaller one.

E1 E2

Ask for the exposure ladder

Tomorrow, when you encounter a claim that X causes Y, divide the available cases into at least three exposure levels—high, medium and low—and write down the effect predicted at each. Then look for the gradient and a credible delivery path from exposure to outcome. If the pattern is flat, reversed or visible only after selective filtering, investigate alternatives such as reverse-causality before acting on the claim.

Episodes that teach this

Related concepts

Causal MechanismCorrelation vs CausationBidirectional CausalityCollider BiasConfirmation BiasNatural ExperimentConfounding VariablesHormesisReverse CausalityTeacher Effect