Levels of Measurement

Nominal Scale
The nominal scale (also called dummy coding) simply places people, events, perceptions, etc.
into categories based on some common trait. Some data are naturally suited to the nominal scale
such as males vs. females, redheads vs. blondes vs. brunettes, and African American vs. Asian.
The nominal scale forms the basis for such analyses as Analysis of Variance (ANOVA) because
those analyses reuire that some category is compared to at least one other category.
The nominal scale is the lo!est form of measurement because it doesn"t capture information
about the focal ob#ect other than !hether the ob#ect belongs or doesn"t belong to a category$
either you are a smo%er or not a smo%er, you attended college or you didn"t, a sub#ect has some
e&perience !ith computers, an average amount of e&perience !ith computers, or e&tensive
e&perience !ith computers. No data is captured that can place the measured ob#ect on any %ind
of scale say, for e&ample, on a continuum from one to ten. 'oding of nominal scale data can be
accomplished using numbers, letters, labels, or any symbol that represents a category into !hich
an ob#ect can either belong or not belong.
Ordinal Scale
The ordinal scale has at least one ma#or advantage over the nominal scale. The ordinal scale
contains all of the information captured in the nominal scale but it also ran%s data from lo!est to
highest. (ather than simply categori)e data by placing an ob#ect either into or not into a category,
ordinal data give you some idea of !here data lie in relation to each other.
*or e&ample, suppose you are conducting a study on cigarette smo%ing and you capture ho!
many pac%s of cigarettes three smo%ers consume in a day. +t turns out that the first sub#ect
smo%es one pac% a day, the second smo%es t!o pac%s a day, and the third smo%es ten pac%s a
day. ,sing an ordinal scale, your data !ould loo% li%e this.
-. Ten pac%s a day smo%er
.. T!o pac%s a day smo%er
/. One pac% a day smo%er
The ordinal scale ran% orders the sub#ects by ho! many pac%s of cigarettes they smo%e in one
day. Notice, ho!ever, that although you can use the ordinal scale to ran% the sub#ects, there is
some important data missing$ the first smo%er occupies a ran% the same distance from the second
smo%er as the second smo%er occupies a ran% the same distance from the third smo%er.
'onseuently, no information e&ists in the ordinal scale to indicate the distance one smo%er is
from the others e&cept for the ran%ing. (icher than nominal scaling, ordinal scaling still suffers
from some information loss in the data.
Interval Scale
,nli%e the nominal scale that simply places ob#ects into or out of a category or the ordinal scale
that ran% orders ob#ects, the interval scale indicates the distance one ob#ect is from another. +n the
social sciences, there is a famous e&ample often taught to students on this distinction.
Suppose you are near the shore of a la%e and you see three tree stumps stic%ing out of the !ater.
,sing the !ater as a reference point, it !ould be easy to measure !hich stump rises highest out
of the !ater. +n this !ay, you can create a relative measure of the height of the stumps from the
surface of the !ater. *or e&ample, the first stump may breach the !ater by t!enty0four
centimeters, the second by t!enty0si& centimeters, and the third by t!enty0eight centimeters.
,nli%e the nominal and ordinal scales, you can ma%e relative distance measurements among
ob#ects using the interval scale.
1o!ever, the distance the stumps e&tend out of the !ater gives you no indication of ho! long
the stumps actually are. +t"s possible that the bottom of the la%e is irregular ma%ing the tallest
stump loo% tallest only in relation to the !ater. ,sing interval scaling, you have no indication of
the absolute length of the stumps. Still, the interval scale contains richer information that the t!o
lo!er levels of scaling.
Ratio Scale
The scale that contains the richest information about an ob#ect is ratio scaling. The ratio scale
contains all of the information of the previous three levels plus it contains an absolute )ero point.
To use the e&ample above, the ratio scale allo!s you to measure the stumps from the bottom of
the la%e$ the bottom of the la%e represents the absolute )ero point.
The distinction bet!een interval and ratio scales is an important one in the social sciences.
Although both can capture continuous data, you have to be careful not to assume that the lo!est
possible score in your data collection automatically represents an absolute )ero point.
Ta%e e&traversion captured using a psychometrically sound survey instrument. The items that
capture this construct may range from )ero to ten on the survey but there is no guarantee that a
score of )ero on the survey places a sub#ect at the absolute )ero point on the e&traversion
construct. 2es, you %no! that a sub#ect !ith a score of eight on the scale is more e&traverted than
someone !ith a score of seven, but those numbers only e&ist for comparison bet!een each other,
not in comparison to some absolute score of )ero e&traversion.