6

I am looking for a document or research articles classifying physical or chemical measurements (or perhaps better means of measurement) according to the reference statistical distribution and properties they have.

Example of questions studied would be

  • What kind of weight balance equipment, technology and procedure have the most gaussian behavior? In what order scale range?

  • What is the typical distribution of error in a spectroscopic ray analyzer?

  • What is the typical distribution of an amperemeter, a voltmeter, etc.?

  • What are the noise distributions of the various diode technologies, etc.?

I am particularly interested by hints of "exotical" distributions in some common apparatus.

chl
  • 50,972
  • 18
  • 205
  • 364
ogerard
  • 167
  • 5

1 Answers1

2

Some people have started to look at this issue in the chemometrics literature. For instance, about 20 years ago Robert Gibbons started to do statistical analyses suggesting instrument responses (for low-level measurement of chemicals) were nonlinear, heteroscedastic, and had non-normal (perhaps lognormal) error distributions. I found an abstract of one of those papers on Springer's site.

whuber
  • 281,159
  • 54
  • 637
  • 1,101