Four belated stats/methodology thoughts on the follow-up discussion about factors that influence coffee and tea ...
1. For a classic example of a blinded taste test by R. A. Fisher, a pioneering giant in stats, see David Salsburg's (2001) "The Lady Tasting Tea." Research question: Add milk to tea or vice versa?
2. I'd love to see a rigorous taste test by the SGU crew. Doing this well could be prohibitively time- and resource-consuming, but even a simple version (e.g., 1 taster, 1 manipulated factor) could nicely illustrate data collection and analysis, including inevitable complications.
3. Food scientists and others have devised numerous "taste test" methods. For examples, see the Society of Sensory Professionals' Sensory Wiki. Choosing among these depends on your aims (e.g., discrimination vs. preference), resources, experience, etc.
4. Anyone wanting to optimize his or her coffee- or tea-prep process might consider an approach like response surface methodology. Its details can get pretty technical, but basically it's used to find which combo of multiple factors yields the best outcome(s).