I totally agree with you! Melany Ramos  I think it is very important what you mention about taking into account the normality and the type of variables. I, for example, didn't even think about that before and used Pearson by default, but then I realized that if the data are not normal, the results can be misleading. That's when I started to use Spearman more, especially in surveys where people respond with Likert-type scales or rankings. It also happens to me that when I see qualitative data, chi-square is my safe option, but I still check if the frequencies are enough for the test to be valid.