How do you eliminate tests?
Posted: Wed Dec 07, 2011 3:01 am
RE: How do scientists decide which tests to exclude from the experiment?
Throwing out outliers is definitely a good first step to make sure that all of the data you are looking at is "clean" data. But, you need to be sure when you are throwing out outliers that you are doing so in a scientific way! Outliers should be those situations where something went wrong during the experiment or the results for one test are different enough from other tests that you think an error must have happened during the test. You might want to come up with a definition for what you consider "outlier" just to make sure that when you are reporting your results people will be able to understand why you chose to discount certain data points.
To cut down the rest of your data, choosing random data points is probably the best option. Choosing random tests out of a hat is a great way to make sure that you are objective because you want to make sure that you aren't just cutting out results because you don't like them as much or they aren't what you were expecting!
This looks like an awesome project and its always a good thing when you have too much data instead of too little! Good luck!
Anna