Correlation is used to measure the relationship between two variables, and to verify if this relationship is straight and direct. But to use this test, the data must have a normal distribution, i.e. bell-shaped, and must be measured continuously or in interval levels. The result shows the direction and strength of association. Spearman's rho coefficient is somewhat similar, it is a non-parametric test used when the data to be analyzed are not normal or technically do not meet Pearson's assumptions. This test is in charge of measuring the monotonicity between the relationship of non-normal continuous variables, and is more useful when there are ranges instead of exact measurements. Finally we have the chi-square test, which is used when we want to know if there is a relationship between two categorical variables, checking if the observed frequencies differ significantly from those expected under independence. I think it is important to understand how normality, linearity and level of measurement work, since not taking them into account can lead to wrong conclusions.