MetricVsMeasurement

What is the difference between a metric and a measurement? DonGray 2003.10.20

The term "measurement" suggests looking at the world and deriving a number from that observation, with the implication that someone else would have gotten the same number. I think of a metric as a potential measurement. The dictionary tells me, "a standard of measurement". -- LaurentBossavit 2003.10.20

My dictionary doesn't give me a satisfactory distinction between the two either. And that's OK. Dicitonaries tells HOW we use words and what the word means in THAT context. I'm reading a book on "metrics", but the author seems to interchange measurement and metric without distinction. DonGray 2003.10.20

I like this distinction:

Measure is a number, created from observation somehow which allows ranking and comparison. 5 granks is both different from and bigger than 3 granks.
Metric is a ratio of a measure to some other number, placing the original number in context. So "gross sales" is one number, and a measure. "Gross sales per store" is another number and a metric.
Indicator is the result of comparing a metric to another number. "Gross sales per store this week" vs. "gross sales per store last week."

I think that distinction comes from some combination of Austin's Managing And Measuring Performance In Organizations, Good To Great, maybe with some of Weinberg's general systems stuff thrown in. At any rate, the discinction has been very useful to me in two ways:

Realizing what we are talking about: an observation, a context, or a conclusion.
Noticing and working with leaps to conclusion. Often we leap to conclusion, declaring that the sky is falling without really paying attention to how we know what's bad, the context of the number or even the basic measurement.

I'm not stuck on people using these actual words in a metrics and measurement conversation, as long as the meanings are distinct. The software measurement and metrics books that I have seen do not make this distinction. You can see the distinction sometimes in pure science.

-- JimBullock, 2003.10.22

Good distinctions. Of course, what we call measures are not usually obtained directly from observation. Gross sales this week is generally not obtained by watching all sales take place and adding them in your head. Even the adding is an extra step that leaves room for interpretation and error. The further you get from direct observation, the more chance there is for such mistakes. - JerryWeinberg 2003.10.22

Yep. I had some fun with this in last summer's CSE-403 class doing a workshop on "general systems" which went pretty well. The distinction between an actual measure (smooth oval in the notation) and something we could measure (cloud) was kind of lost on them until I hit on the right rule:

If you can count it on your fingers and toes, or if there's a meter we have that we can hook up and read a number, it gets an oval. Otherwise we're guessing and it gets a cloud.

This distinction between things we actually measure and things we could measure is an important part of Jerry's stuff on general systems. I don't recall seeing this distinction in other general systems references.

And of course, the rule as stated is easier for a bunch of people to remember in part because it's easy to make it multi-mode. Every time I said "count it on your fingers and toes" I held up a hand and counted my fingers. When one student proposed a measure that was really a feeling inside his head, I got to pretent to attach wires of a meter to his ears. Then of course "has a meter" goes with that gesture.

I've tried to get that distinction through without the visual for years, and it's way, way harder.

-- JimBullock, 2003.10.23 (Purposeful Antics)

Bryce Ragland, the author, uses a series of charts of body temperature to illustrate the difference between measure, metric and indicator. The charts worked better for me than the words.

Jerry, The further you get from direct observation, the more chance there is for such mistakes.

Bullseye. I will put that insight to use.

SteveSmith 2003.10.27

I don't recall seeing that paper before, but yeah, that. I also usually use pictures when explaining the distinction (vs. using it.) Most of the time, they are ad-hoc pictures about something of immediate interest. "We're ready to ship." "But there are 87 bugs." Ah. OK, now we have an opportunity for an indicator discussion.

Indicators can also be decidedly non-linear and, the comparison to a reference value doesn't have to be, well, a numeric comparison, nor do the metric and the reference have to fit in Cartesian notation. But drawing those alternatives is hard.

-- JimBullock 2003.10.27

The discussion reminds me of S.J. Gould's "The Mismeasure of Man", about the IQ metric. Do we have a page about MisMeasurements ? -- LaurentBossavit 03.10.27

How big might such a page get? In terms of what and compared to what? How much is too much (or too little?)

-- JimBullock 2003.10.28 (Or as like him as need be for most puposes, but how would you know?)

Updated: Wednesday, November 12, 2003