Soooo.......
I'm thinking about "normalized behavior".
From a math perspective, I believe this connects back to the normal or bell curve. It essentially means that one scalar variable or one type of set inclusion (to allow for proportional fits) will have a large majority of data cluster around one value on the scale... and the rest of the values are used by smaller and smaller percentages of the data.
(I say that, but mathematicians also talk about normalizing vectors, and finding perpendiculars/normals... )
So, what do we do if the data refuses to clump under one value? What if there are multiple scalars or measurements being made at the same time, on the same subjects? (multiple factors)
We can't say that if you're in the majority group on one set of scales, that you'll also be in the majority on a different set of scales. In fact (with multiple clumps), we can't even claim that there's always one majority. Also, not all set partitions are polar - like there's six different values for race on previous censuses. (or there's more than 2 eye colors)
Is it a bad idea to strive to be like the majorities? You know, "married, with 2.2 kids, living in the suburbs, with the median income, and a High School Education"..? Do people who think they're in the majority strive for those types of goals?
Does that capture "keeping up with the Joneses"?
No comments:
Post a Comment