Originally posted by Doc Tsiolkovski
The IQ or any comparable issue (like normal ranges for medical labratory parameters) are generated from that results, to exactly fit the Gaussian model.
I wanted to point out, that IQ tests
are a DISCRETE sample and the Gauss normal
distribution is a continuous function with only
two parameters, so it can only be an approximative
model for a discrete set of data.
Everyone, including our research group, uses
this function for all sort of statistical
calculations, even for very small data sets.
You can calculate averages and standard deviations
as if it was a continuous function.
I have seen publications with standard deviations
calculated from 3 data points and that is complete
nonsense. A discrete set of data is not symmetrical
and it cannot be fitted exactly, only approximately.
