Video Details

  1. Hypothesis
    • Start with a specific question
  2. Variable
    • Choose something that you’re going to tweak
    • You want to determine what happens when you tweak that one variable
  3. Sample Group
    • determine where your data is coming from and how much is needed
    • data should be as randomised as possible
  4. Control
    • create another sample group to test against that you don’t make any modifications to
    • consider running a double blind, where not even you know which control group is which.
  5. Duration
    • determine the amount of time you need for the test
  6. Clean Lab
    • do what you can to ensure that there are no other variables that could dirty your results
  7. Manipulate one variable
    • manipulate only one variable at a time
  8. Statistical Validity
    • ensure your results have a statistical significance
    • work with a trained data scientist to help check your correlations
  9. Transparency
    • share the results/tests/methodology to the world
  10. Replication/Falsification
    • do people running the same test as you get the same results?
    • feedback from others on where you went wrong could falsify your results which is fine because you’re after the truth