Correlation (Statistical overview)
Definition
Correlation
“Correlation is a statistical technique that can show whether, and how strongly, pairs of variables are related. There are several different correlation techniques, including the Pearson or product-moment correlation”1
Monotonic association between 2 variables(Schober et al., 2018)
- one variable increases → other increases
- One variable increases → Other Decreases
Variance
Measure of the variability of a single variable2
Covariance
Defined as the Degree to which the change in 1 continuous variable is associated with the change in another continuous variable2
Covariance is a measure of how 2 variables vary together2
Intra-class Coefficient (ICC)
- 0-1 (higher is better)
- Correlations between repeated measurements within a class of data3
Correlation coefficient
Describes Strength & Direction of association between variables
Coefficient of Determination (R^2)
Proportion of variance in 1 variable that is accounted for by the other2 .
Linear correlation
Interpretation
See We considered a correlation coefficient of .00–0.30 to be very weak, 0.31–0.50 weak, 0.51–0.70 moderate, 0.71–90 strong and 0.91–1.00 very strong as proposed by Mukaka
Mukaka MM. Statistics corner: A guide to appropriate use of correlation coefficient in medical research. Malawi Med J. 2012;24(3):69–71.