Haynes Goddard writes:
Reviewing my notes and books on categorical data analysis, the term “nominal” is widely employed to refer to variables without any natural ordering. I was a language major in UG school and knew that the etymology of nominal is the Latin word nomen (from the Online Etymological Dictionary: early 15c., “pertaining to nouns,” from Latin nominalis “pertaining to a name or names,” from nomen (genitive nominis) “name,” cognate with Old English nama (see name (n.)). Meaning “of the nature of names” (in distinction to things) is from 1610s. Meaning “being so in name only” first recorded 1620s.)
So variables without a natural order such as gender (male-female), transport mode (walk, bicycle, bus, train, car) and so on are just coded 0, 1 and so on. Yet the textbook writers do not explain that nominal just means name which it seems to me would help the students better understand the application.
Do you know when this usage was first introduced into statistics?
I have no idea but maybe you, the readers, can offer some insight?
That’s precisely the way I teach it. I took Latin in high school, so that’s what just made sense when I learned about “nominal” measurements.
Same here, I always say “nominal from the Latin word for name.”
That is what I do too. I tell my students that they are merely labels, like a student ID number. It may have some underlying logical ordering, but it’s irrelevant for our research purposes.
The obvious guess is Stevens’s paper in which he introduced his system of scale types (he mentions there that this was presented by him in 1941):
Stevens, S. S. (1946). “On the Theory of Scales of Measurement”. Science 103 (2684): 677–680.
However, it may well be that the term “nominal scale” was around in statistics before that time.
There’s also the economic version of this, which refers to GDP (nominal) as GDP at some official market/currency rate. In this case the meaning takes on ‘baseline’. I think that particular definition of nominal, much like many words in the English language, was a completely bastardized Georgian or Texan derivative of ‘norminal’. And like the words ‘conversate’ and ‘strategery’, we know that this simply the linguistic norm of converting hard sounds to soft ones …. ah reckun….
Never misunderestimate the power of Texans to introduce new words or new meanings for old words.
For example, ‘Merkin’. Def: A citizen of the United States. Example: “Ahm a Merkin.”
(Not to be confused with the official meaning of ‘Merkin’. Look it up!
My understanding about a nominal variable is that the *only* relevant information is its name (the names of its categories).