Correlation can be real without being causal
Two variables can move together because one influences the other, because both respond to a third variable, or because the relationship is only a statistical artifact.
Statistics
Correlation vs Causation
A correlation can be statistically significant and still fail to prove a cause. This page shows why that matters and what to check next.
A small p-value is not a causality badge
The p-value only measures how unlikely the observed relationship is under the null hypothesis. It does not tell you why the relationship exists.
Direction still matters
If the timing or mechanism is unclear, you should treat the finding as a signal to investigate, not as a finished explanation.
What to check
Time order
Did the supposed cause happen before the effect?
Confounders
Could another variable explain both measurements?
Design quality
Was the result based on an experiment, a survey, or an observational dataset?
Back to the calculator
If you already have r and n, use the p-value guide or the Pearson calculator to test significance first. Then decide whether the study is worth a deeper causal claim.
FAQ
Correlation and causation FAQ
Does a significant correlation prove causation?+
No. A significant correlation only says the observed pattern is unlikely to be random under the null hypothesis. It does not identify the cause.
Why is correlation not enough to claim causation?+
Because confounders, reverse causality, selection bias, and study design limits can all create or distort a correlation.
What should I check before claiming causation?+
Look for experimental design, controls, time order, and alternative explanations before treating a correlation as causal.