|
|
Experimental design: the `two systems' model
The `two systems' model
``The Renault Repulsive rated higher than the Vauxhall Violent for fuel
economy''
The archetypal comparison experiment is to compare two equivalent things
using the same criteria. For this to be valid, the two things must genuinely
be comparable
For example, there is little to be gained by comparing the speed of information
retrieval of new software system with, for example, a printed encyclopaedia.
These systems are not genuinely comparable. Sometimes it is not obvious whether
two systems really are comparable or not. In general, if you can predict in
advance the outcome of a comparison, you are probably not comparing two
equivalent things.
|
