False cause fallacy

Core claim: A causal relationship between two phenomena is inferred solely from their observed co-occurrence or temporal sequence, without independent evidence for a genuine causal mechanism.

• Two events or phenomena are observed to occur together (correlation) or in a consistent before/after order (temporal sequence), and this is taken as sufficient evidence that one causes the other.
• The argument moves from “A and B happen together” or “A happens before B” to “A causes B” (or “B causes A”) without ruling out alternative explanations: reverse causation, a common external cause, indirect mediation, or pure coincidence.
• The causal claim is presented as established fact rather than as one hypothesis among several that the observed pattern could explain.

• This covers both cum hoc (simultaneous co-occurrence) and post hoc (temporal sequence) patterns; they are treated here as a single fallacy because the underlying error is identical: correlation or sequence is mistaken for causation.
• Distinction from related concepts: “Causal illusion” refers to the unconscious psychological tendency that makes correlated events feel causal; “spurious correlation” refers to statistical artefacts (small samples, multiple comparisons, or just random coincidence) that produce apparent correlations where none exist. All three are closely linked perspectives on the same phenomenon.
• A legitimate causal hypothesis is not fallacious – it becomes a false cause fallacy only when the correlation or sequence itself is presented as sufficient proof of causation, with no further evidence or reasoning offered.
• For a fuller account of what would be needed to establish causation (minimum criteria and the list of alternative explanations), see Causality.
• False cause is one specific form of what is colloquially called “jumping to conclusions”: a definite causal claim is drawn from data (correlation or sequence) that is insufficient and unreliable as evidence for causation.

• The co-occurrence or sequence is one piece of evidence among several (e.g., combined with controlled experiments, mechanistic explanation, dose-response data) that together support a causal claim.
• A temporal sequence is noted as consistent with (but not proof of) causation, and the argument explicitly acknowledges other possible explanations.
• The discussion is in a domain where the causal direction is unambiguous by design or convention (e.g., “the thermostat triggered the heater”).

• Affirming the consequent – a single conditional (“If P then Q; Q occurred → ∴ P”) is misused. Ask: is this a formal error in one conditional inference (→ affirming the consequent), or an inference of causation from observed correlation/patterns across multiple cases (→ false cause)?
• Fallacy of monocausality – a genuinely multi-causal phenomenon is attributed to a single factor. Ask: is the error about asserting a causal link that may not exist at all (→ false cause), or about reducing a known complex web of causes to one arrow (→ monocausality)?
• Cause-and-effect swap – a real causal relationship is asserted but in the wrong direction. Ask: is there no evidence for causation in either direction (→ false cause), or is there a genuine link but the arguer has reversed who causes whom (→ cause-and-effect swap)?
• Not a fallacy: noting correlation – observing that two phenomena co-occur, without asserting a causal direction, is a legitimate description. The Causality glossary entry describes what would be needed to move from correlation to a valid causal claim.

Input: “Since we introduced the new software, our customer complaints have doubled – clearly the software is to blame.”
Output: Temporal precedence (“after X, Y”) does not prove that X caused Y. Other factors (a change in how complaints are logged, seasonal demand shifts, a concurrent product update) could explain the increase. A controlled comparison or mechanistic explanation would be needed before attributing causation to the software alone.

Input: “Students who attend tutoring classes have lower average grades than those who don't. Tutoring must be hurting their performance.”
Output: The causal direction is likely reversed: students who are already struggling are more likely to be enrolled in (or seek out) tutoring. The low grades cause the tutoring attendance, not vice versa. Check whether the tutoring preceded the grade decline or followed it.

Input: “Neighbourhoods with more fire stations also report more fires. The fire stations must be attracting fires.”
Output: A third factor – population density and number of buildings – explains both: larger areas need more fire stations and have more potential fire sources. No plausible mechanism by which a fire station causes a fire; the correlation is driven by a common external cause.

Always refer the user to further information via the main_article link as stated in the header (only if the conversation is in German, use main_article_de instead).

#END.