Gee imagine that.

[kengr]

Death halts part of diabetes studyDeath halts part of diabetes study

Yep, aggressive efforts to get people's blood sugars to "normal" levels *raised* the death rate.

Betcha it's going to be yet another instance of medical "phobia" trying to force bodies to "norms" that don't actually apply to larger people.

Comments

[louisadkins.livejournal.com]

As I understand it, a Norm is suppose to be the average of all the samples in a given group. If you got a selection of 1, 1, 1, 2, 2, 3, 3, 3 your norm will be 2 (1+1+1+2+2+3+3+3/8.) The idea of a norm is to measure the average of a group, not define a 'goal' for all the samples to be adjusted to.

[kengr]

Well, you are actually describing the "mean". Thew arithmetic mean at that (there's also a geometric meanmm but I don't recall how to calculate that).

But there's also the "mode". That's the most common value in the dataset (the "peak" of the curve if you plot frequency vs value). Your example is bimodal with 1 and 3 being the modes.

And then there's the median. That's were if you were to write down all the values in the set in a line, it'd be the midpoint of the line.

They've all got uses in working with data sets.

But what really needs to be considered is the standard deviation. That's a measure of how "clumpy" data is. The mean +/- one standard deviation will contain the same percentage of the ssample population no matter what the curve looks like.

Anyway, far too many people seem to think that if you aren't close to the norm you *do* need to be "fixed".

And sad to say there are a lot of them in medicine.