CORREL
Returns the Pearson correlation coefficient between two datasets. The coefficient ranges from -1 (perfect negative correlation) to +1 (perfect positive correlation), with 0 indicating no linear relationship.
CORREL(array1, array2)Arguments
array1The first range of values.array2The second range of values. Must have the same number of data points as array1.
=CORREL(A2:A20, B2:B20)A value close to 1 indicates a strong positive correlation between the two variables
=CORREL(C2:C50, D2:D50)A negative value indicates that as one variable increases, the other tends to decrease
=CORREL(A2:A10, B2:B10)A value near 0 suggests no linear relationship between the two datasets
- •CORREL measures linear correlation only — non-linear relationships may show a low coefficient even if strongly related
- •Both arrays must have the same number of data points
- •Correlation does not imply causation
- •Use PEARSON for the same result with a slightly different function name
- •Interpreting a low CORREL result as proof of no relationship at all - CORREL only measures linear correlation, so two variables can be strongly related in a non-linear (curved) way and still show a coefficient near 0
- •Mismatching the size of array1 and array2, which raises a #N/A error since every value in one array needs a corresponding value in the other
- •Assuming a strong correlation means one variable causes the other - CORREL only measures statistical association, not causation
PEARSONCalculates the identical Pearson correlation coefficient as CORREL, just under a different function name.RSQReturns the square of the correlation coefficient (R-squared), a related measure of how well a linear model explains the variance between two variables.COVARIANCE.SMeasures how two variables vary together in absolute terms, the unstandardized measure that CORREL normalizes into a -1 to 1 scale.Does a CORREL result near 0 mean there's no relationship between my variables?
Not necessarily - it means there's no linear relationship. Two variables can have a strong curved or cyclical relationship and still produce a correlation coefficient close to 0.
Why does CORREL return a #N/A error?
array1 and array2 probably don't have the same number of data points - correlation requires every value in one array to be paired with a corresponding value in the other.
Does a high CORREL value prove causation?
No, correlation only measures how strongly two variables move together, not whether one causes the other. A high correlation could result from coincidence or a third, unmeasured factor influencing both.
Need to translate a formula using CORREL?
Use our translator to convert your complete formula