Information Technology Reference
In-Depth Information
The key to working with this sort of scale is to appreciate that the aver-
age derived from a summative response scale is meaningful. Let's say that
we administered a short self-esteem inventory to a class of public policy
graduate students and that one item on the inventory read, “I feel that I am
aworthwhileperson.”Assumethatitemswereratedonafive-pointscale
with higher values indicating more endorsement of the statement. Let's
further say that mean for this item based on all of the students in the class
was 4.75. Is that value interpretable? Yes, it indicates that the individuals
in the sample believed pretty strongly on average that the content of
the item was quite true for them, namely, that they were worthwhile
people.
2.1.5 INTERVAL SCALES
Interval scales of measurement have all of the properties of nominal,
ordinal, and summative response scales but include one more important
feature. Fixed distances between the numbers represent equal intervals.
The most common illustrations of an equal interval scale are the
Fahrenheit and Celsius temperature scales. According to Stevens (1951,
p. 27), “Equal intervals of temperature are scaled off by noting equal vol-
umes of expansion. . . . ” Essentially, the difference in temperature between
30 and 40
◦
F is equal to the difference between 70 and 80
◦
F. A less-obvious
but important characteristic of interval scales is that they have arbitrary
zero points. For example, the term
zero degrees
does not mean the absence
of temperature - on the Celsius scale, zero degrees is the temperature at
which water freezes.
As was true for summative response scales, it is meaningful to average
data collected on an interval scale of measurement. We may therefore say
that the average high temperature in our home town last week was 51.4
◦
F.
Note, however, that summative response scales are not quite at the interval
level. That is because in most cases it is not true that the difference between
the scale values of, say, 1 and 2 represent the same psychological distance
value as that between, say, 2 and 3.
2.1.6 RATIO SCALES
A ratio scale of measurement has all of the properties of nominal, ordinal,
summative response, and interval scales but includes one more important
feature. It has an absolute zero point, where zero means absence of the
property. Examples of ratio scales are time and measures of distance.
Because of this, it is possible to interpret in a meaningful way ratios of the
numbers on these scales. We can thus say that four hours is twice as long
as two hours or that three miles is half the distance of six miles.
2.1.7 QUALITATIVE VERSUS QUANTITATIVE MEASUREMENT
It is useful for our purposes to identify two general categories into
which we can classify subsets of these measurement scales:
qualitative
and
