When should an F-test be used?
The most commonly used F-test is the equality of two variances. The null hypothesis is that two populations have the same variance. This test makes certain assumptions and should only be used when those assumptions are met.
The F-test assumes independence within a sample and between a sample as well as normality of samples. This is a test that is very sensitive to non-normality and should be used with caution.
If you have very small sample sizes, it may be best to use a different test that does not assume normality. This is because it can be hard to detect if your sample really is normal with such limited data available.
To read more, see this webpage.