The interocular trauma test is engineers' and statisticians' humor for the plot-first rule: graph the data, and if the effect leaps out, the formal test only confirms what everyone already sees. It also works as a warning in reverse: if a result needs a p-value before anyone can see it, it is probably small. Experienced analysts plot before they test, because plots expose outliers, curvature and mixtures that summary statistics hide.
Didn't bother with a test. Box plot passes the interocular trauma test, new tool reads 3 nm higher with zero overlap.
Also heard as
- IOTT
- eyeball test
- hits you between the eyes
Related terms
t-test
Test that compares a mean against a target, or two means against each other, using the t distribution.
Significant
Describes a result unlikely to have come from chance alone at a stated risk level, which is not the same as important.
Outlier
Observation that sits far from the pattern of the rest of the data, either a genuine extreme or an error.