**Choosing which test (click on the links below for more information):**

- First, decide if your data is continuous (e.g., 1.1, 2.5, 3.4, 4.5, 7.9), or categorical (e.g., yes/no, green/yellow, present/absent). If you have discrete numerical categories (e.g., 1, 2, 3, or 100, 200, 300), they can be treated as either continuous or categorical.
- Then choose a stat:

**Some common stats (in Excel)**

__Formatting data, basic correlation, descriptive stats____Regression and ANOVA____F-Test__to test the null hypothesis that the variances of two populations/groups are equal. Run this before a T-Test to know which T-Test to select. If, for the F-Test, P<0.05, the variances are not equal (meaning one group/pop is more variable than the other). Use this information to help you decide which T-Test to perform (see below).__T-Test__to test the null hypothesis that the means between two populations are equal. If the variences between the groups are unequal (as determined by F-Test with P<0.05), do a "t-Test: Two-Sample Assuming Unequal Variences." If the variences are equal (as determined by F-Test with P>0.05), then do a "t-Test: Two-Sample Assuming Equal Variences."

If the result of the T-Test gives a P<0.05, the means of the two populations/groups are significantly different.__Chi Square Test__(comparing actual vs. expected)__Chi Square Test Test__(another example)__Discriminant Analysis (LDA)__- This is a bit more complex, and for the less common situation of a continuous independent variable to compare to a categorical dependent variable, or classifying things into groups. Here is__another LDA tutorial__.