You need to think how you will analyse your results at the beginning.
Don't assume you can get data and then mine it for meaning. The way you design your program will, effectively, define appropriate statistical analyses. Just because you can put numbers into an equation or statistical program doesn't mean that the outputs mean anything. Meaning is dependent upon appropriate analysis of properly designed studies, which is why analysis and design are inseparable.
Some tests assume the data will be normally distributed. Even the simple mean will not be appropriate for skewed data.

