Correlation: Pearson or Spearman?


Q: when should someone use pearson’s correlation or spearman’s correlation?

from my understanding, pearson r coefficient is a measure of the strength of a linear relationship, whereas spearman rho coefficient is a measure of monotonic relationship. Every linear relationship is monotonic, but not the other way around.

does using one correlation test over the other dependent on the distribution of the data? I’ve read that if your data isn’t normally distributed, then you can use spearman or even kendall instead of pearson.

Q: what if both pearson and spearman coefficients are too small (<0.3)? what kind of conclusion can I draw from these values?