Concepts

Concepts

1. "Gold Standard" - Randomised control trial (RCT)

   • participants randomly assigned to two or more treatments
   • blind trial to participants and assessors to reduce bias

2. We want to find experimental units as similar as possible so that when we impose treatments we can assume the measured differences are due to those treatments rather than the background variation

3. Definitions

   • Experimental unit: the unit randomly assigned to a treatment
   • Sampling unit: the object measured in an experiment
   • Replication: the repetition of an experimental situation by application to more than one experimental unit
   • Factor: a structure that specifies an allocation of the units into groups
   • Treatment: the entire description of what can be applied to an experimental unit
   • Block: a division of the experimental units into groups of similar structure/assumed variability
   • Plot/design structure: the way the experimental units are organised physically/temporally, irrespective of treatments
   • Treatment structure: the set of factors that the experimenter wants to compare
     • Simple structure: involves a single factor with many levels
     • Factorial structure: involves some combination of factors