Which correlation should be used to measure the relationship between gender and grade point average for a group of college students?

Neither Pearson's r nor Spearman's rho is appropriate for measuring the relationship between gender (a categorical variable) and GPA (a continuous variable). A point-biserial correlation would be the most suitable method. This correlation is a specific type of Pearson correlation adapted for use when one variable is dichotomous (like gender: male/female).