In his lecture on Pathological Science, Langmuir’s first test was stated as:
The maximum effect that is observed is produced by a causative agent of barely detectable intensity, and the magnitude of the effect is substantially independent of the intensity of the cause.
What exactly does this mean?
Consider an elephant and a fly. An elephant is ambling along at 2 mph and the fly is pushing against the elephant’s backside, madly flapping his wings. From a force perspective, the inertia of the fly is added to the inertia of the elephant, resulting in the elephant’s velocity being just a fraction faster than 2 mph.
Now suppose the fly gets four of his friends to join in pushing the elephant along. We now have five times the single fly force moving the elephant along, but we can only measure an increase in the elephant’s velocity approximately twice that of a single fly pushing.
Do we really expect that if we get enough flies to push we could double the elephant’s speed? Math says that it is possible, as long as the total inertia of the flies is equal to the inertia of the elephant. However, it is unlikely to happen.
Personally, I do not think anyone has ever observed a swarm of flies ushering an elephant along. Just as I do not expect to see a swarm of mosquitos carry a person off to dine at leisure.
The idea behind this is that a driving force should be substantial enough to be easily measurable and the resulting effect should vary as the driving force varies. There is plenty of wiggle room in this statement.
“Easily measurable” implies that a proper instrument can give an accurate, precise, and repeatable result. For example, some analytical instruments can measure down to the part per billion (ppb) range.
Similarly, the effect (output) should vary (either increase or decrease) as the driving force (input) varies. The variability of the output should be statistically significant, or discernable from experimental error. Note that the driving force is the independent variable – the one that the experimentalist sets, and he should be able to set it quite accurately.
If the data from an experiment do not meet this simple test, that speaks volumes about the validity of the conclusion.