Quote:
Originally Posted by J_Player
Hi Theseus,
One good thing about trying several machines at the same test location is you can see how they perform at the same place under the same conditions of target "strength of field". If conditions are only average, then we can take a look to see how they all perform under the same average conditions. Then you don't need to guess if there was a changed condition between one machine tested and another.
...Same as they test cars on the same track with the same test conditions so there is no difference in the curves of the road, or weather that could skew the results between one car and another.
Best wishes,
J_P
|
When dealing with a large number of variables or parameters, and the desire is to still arrive at meaningful conclusions from fair testing procedures; about the only way to accomplish such a task is to incorporate BIBD (Balanced Incomplete Block Designs) and couple that with a good double-blind protocol.
Even if you do all that... as Jim has pointed out, there will still be criticism from one side of the fence or other - depending on the results published.