Statistics 2: Design + Hyp testing

Statistics 2: Design + Hyp testing

1. Developing a hypothesis

   • A tentative ( not certain) proposition (statemnet) subject to verification by investigation
   • Hypotheses are never PROVEN
   • null: no sig diff between 2 sets of data
   • Scientists work to show that the null is not vaild and the alternate is

2. Things that need to be considered

   • Control groups
     • No intervention
     • Needs to be concurrent (not historical)
     • Beware placebo effect
     • beware confounding variables
   • Sample size (randomisation)
   • Randomisation