Vous êtes sur la page 1sur 3

SCALES OF MEASUREMENT

Overview
There are a number of ways of measuring things in psychology. Take a moment to think
about how you might measure attitudes, anger, language development, attention, or neural
processes. These are all research topics that require creative methodology to measure what
you're trying to measure (or, in other words...have internal validity). Here, you'll get a brief
introduction to four different scales of measurement that are used in a variety of
psychological fields. Understanding the differences between these and being able to identify
the scale is important in selecting which statistic you use to analyze your data. These are
presented in order of precision, from the least to the most precise.
Nominal Scale
This is basically a way of categorizing or grouping behavior, where the actual numbers are
simply labels or identifiers. For example, let's say you were interested in whether a particular
display in a store was more effective in inducing males or females in purchasing the product.
You could categorize your observations into two categories of 'look but fail to purchase' and
'look and purchase.' You could then assign a '1' to females and '2' to males and record your
data in each category. Using '1' for females is completely arbitrary -- it's just an identifier.
You could have used 'A,' '2', 'bluk,' or whatever. The point here is that your data are in
categories with the number being simply used as a label without any meaning or indicating of
order. Usually, nominal data are presented in terms of percentages in each category.
More Examples of Nominal Scale Use
Counting how many people help someone else in a set-up scene (e.g.,
someone pretending to be blind cross the street).
Counting the number of people out of a group of 30 who are relieved from
their depression three months after treatment.
Examining whether extroverts or introverts are more likely to engage in
altruistic behavior.
Asking whether females or males have higher self-esteem.
Ordinal Scale
Ordinal scales, more precise than nominal scales, are basically sets of rankings. There is no
way to know the size of the differences between any data points--just that one is greater than
the other. A simple example of this is to arrange a class by height and assign each person a
number according to their height ranking. If number 1 is the shortest, all we know is that
number 2 is taller, 3 is taller than 2, and so on. We don't know how much taller 1 is from 2, 2
from 3, etc. Not only do we not know the differences, we don't know the actual height of
anyone.
More Examples of Ordinal Scale Use
High school class rankings.
Social economic class (low, middle, high).
Interval Scale
Interval scales keep the same rank characteristic as ordinal scales, but interval scales also
show the differences between each data point. That is, the difference between 1 and 2 on an
interval scale is the same as the difference between 4 and 5, or 8 and 9, or 100 and 101. In
other words, the interval is the same. Oftentimes in psychology things are measured by a
Likert scale in which one rates a statement (often by how much they agree with the
statement). Here's an example:



We can generally assume that the interval (psychologically speaking) between 1 and 2 is the
same as the interval between 4 and 5. It's very difficult to demonstrate equal intervals here,
but it's often assumed in order to analyze data. What an interval doesn't mean is that a rating
of 4 is twice as great as 2. In fact, the numbers themselves don't tell you anything about the
meaning of the numbers themselves. In the scale above, we could have used a scale from -2
to 2, like this:


It very important to understand that zero doesn't have any meaning in an interval scale. As
you can see from the two Likert scales above, the choice of numbers is fairly arbitrary since
the scale need not begin with any particular number. What is important (once again) is that
the interval between each value on the scale is the same.
More Examples of Interval Scale Use
IQ Scores. (The difference between an IQ of 92 and 98 is the same as the
difference between 130 and 136. Using 100 as the average is arbitrary.)
Thermometer readings on a Fahrenheit scale. (The difference between 98.6
and 99.6 is the same as the difference between 101.8 and 102.8 -- 1 degree.
The value of zero doesn't mean "the absence of heat."
Ratio Scale
The most precise and powerful of scales, ratio scales have all the components of an interval
scale but here, the zero point is meaningful and means the absence of whatever it is you're
measuring. Thus, you cannot have a negative data point using a ratio scale. Not only are the
intervals the same but here, you can compare scores in ratio to other scores. That is, a score
of 20 is 20 times greater than 1 (20:1) and 10 times greater than 2 (20:2) and twice as great as
Indicate your agreement with the following statement: I am satisfied with
psychology as my major.
1 2 3 4 5
Strongly Disagree Disagree Neutral Agree Strongly Agree
-2 -1 0 1 2
Strongly
Disagree
Disagree Neutral Agree Strongly Agree
10 (20:10). For example, measuring something with a ruler would give you a measure in a
ratio scale. Zero literally means "no length" (i.e., that it doesn't exist). Something that is twice
inches is half the length of something that is four inches. This is not true in an interval scale.
More Examples of Ratio Scale Use
A speedometer.
Walking speed (see Dr. Levine's Pace of Life research by clicking on
'Geography of Time' excerpt).
Really, any time or length measurement would be on a ratio scale.
The cost of a cup of coffee

Vous aimerez peut-être aussi