Introduction

The variables
to research

The educational objectives

Knowing how to define a variable.
Knowing how to distinguish the categories of variables.
Knowing how to distinguish between a simple variable and a composite variable,
and the dimension indicator.
Know how to distinguish the independent variable from the dependent variable.
Knowing how to determine the level of measurement of a variable.
Knowing how to connect and analyze
independent and dependent variables.

The summary
1. Definition of the variable
What is an indicator?
3. The types of variables
4. How do we determine the level of measurement?
5. One characteristic, several variables
6. Mathematical gap and semantic gap
7. The simple variable and the composite variable
8. The independent variable and the dependent variable
I do my exercises.

Page 1 Introduction to Research Donald [email protected](506) 858-4886 Center of

Research and Development in Education (CRDE) University of Moncton, N.B. Canada E1A 3E9
 1. Definition of the variable

The variable is, first of all, an indicator of any phenomenon.

for example, heat or smoke emitted are indicators of fire and money
is one of the indicators of wealth. Then, this indicator can change.
value, hence its name variable. Finally, the various values it takes are
measurable and quantifiable.

A variable is an indicator
measurable and quantifiable
who can change value.

Many events can be classified as phenomena. But what is it?

What is a phenomenon, exactly? It can be, in particular, behaviors.
humans, social situations, an economic value, a concept or a
chemical reaction.

It is also evident that this phenomenon must occur with a certain

regularity and that it must not be the product of chance if we truly want
establish its connection with other phenomena. In certain places on the planet,
an earthquake will not lead to in-depth studies because it
produced rarely. In other places, where they are frequent and regular, they
are taken very seriously, not only because of their effects
devastating, but also because they occur with a certain regularity.

The variable must have

at least
two different values.
Values can, in theory, extend to infinity. For example, the
The variable sex has two values, male and female. The variable distance, by
against, has several values and can even be measured according to various
units of measurement.

The main function of the variable

is to be sensitive
to a change of state of a phenomenon,
just like the thermometer in your home
will react to the slightest variations
in the temperature.

Page2 Introduction to Research Donald [email protected](506) 858-4886 Center of

Research and Development in Education (CRDE) University of Moncton, N.B. Canada E1A 3E9
A concept or a perception is a phenomenon that is harder to measure than the
weight of an object. However, we do it, but with less precision. A
social behavior can also be measured, but, once again, with many
imprecisions. Let us also add that a social phenomenon or behavior
Human beings do not always appear in a single and definitive form.

On the other hand, the measurements are more precise and stable in the physical universe,
because most of the time they are sensitive and precise instruments that are
used to measure phenomena. And, several of these phenomena
physicists follow strict laws, such as that of gravity. However, it is necessary to
that the variable is quantifiable according to an accepted measurement scale and, if
possible, universal.

What is an indicator?

An indicator is a characteristic or an aspect of an object or a.

phenomenon. Speaking of a house, for example, we say that it is
spacious, clean, etc. All these characteristics can become
indicators, variables. When talking about intelligence, we will refer to the
capacity to memorize, reason or learn.

An indicator or a variable,
it is an aspect
which characterizes a phenomenon,
but it is not the phenomenon itself.
In research, comparisons between individuals focus on
characteristics that, precisely, distinguish them from one another. We compare the
individuals, for example, regarding the number of hours per week spent doing
physical exercise. The variable in this case is the number of hours dedicated to
exercises. The number of hours varies from one individual to another and is,
consequent, quantifiable.

One cannot compare individuals in terms of the existence of intelligence,

because everyone owns one. This indicator is, in fact, a constant. At all
Useful fins, why bother measuring a characteristic that would be
the same for everyone! However, it is possible to compare the
individuals regarding any manifestation of their intelligence. At the risk
To avoid repeating oneself, another way to define a variable would be as follows.

Page 3 Introduction to Research Donald [email protected](506) 858-4886 Center of

Research and Development in Education (CRDE) University of Moncton, N.B. Canada E1A 3E9
 A variable is
any characteristic
which varies
and of which we can
measure the variations.
3. Types of variables

There are four types of variables: nominal, ordinal, interval, and

of proportion. The following table by Marie-Fabienne Fortin (1996) provides a
Description of each type of variables. We will add examples.

Type Description
Allows you to organize subjects, events, or objects by categories.
It involves assigning numbers without numerical value, that is to say
Nominal which cannot be added together or ranked in size.
Gender, religion, social status, the spoken language.
Subjects, events, or objects are classified according to an order of
grandeur. The numbers indicate the rank, not quantities.
Ordinal absolute numbers. As such, numbers cannot be
added or subtracted.
Rank obtained among a group of individuals.
The intervals between the numbers are considered equal. They
can be added and subtracted. These are not numbers here.
At intervals absolute since the calculation is done from an arbitrary zero.
Ex. The thermometer, the Likert scale, the IQ, the aptitude, the
yield.
The scale has an absolute zero that has empirical significance. The
numbers on the scale represent the actual amount of the
Of proportion measured characteristic. Numbers can be subject to all
mathematical operations.
Weight, volume, duration, length.

3.1 Nominal variable

Religion can become a nominal variable. The individual then becomes part of a
religion: Catholic, Anglican, Muslim, or other. Religion is a
variable since it is a characteristic that distinguishes individuals from one another.
We can code each religion. However, the number assigned to each one
they have no mathematical value: one cannot add religions,
for example. The order in which they are numbered means nothing either,
since they cannot be arranged in order of magnitude on a continuum.

Page 4 Introduction to Research Donald [email protected](506) 858-4886 Center of

Research and Development in Education (CRDE) University of Moncton, N.B. Canada E1A 3E9
Another example of a nominal variable is civil status. The individual can be
find in one of the following categories: single, married or married woman,
divorced or divorced woman, or widower or widow. Any of these categories is
a variation of the civil status variable, but there is no degree between each
state. It matters little how they are numbered in a particular way, for these
categories are, so to speak, equivalent to each other and independent
one from the other. Whether the single category is coded 1 or 3, it doesn’t matter.
since no arithmetic operation will be performed on the codes.

3.2 Ordinal, interval, and ratio variables

Certain characteristics can be measured from a continuum ranging

from zero to infinity. The measure used in these cases is constructed so that
each variation should be associated with a decrease or an increase on
the measurement scale.

An ordinal variable primarily reflects an order among values. One thing or

an individual is arranged according to their relative position to others. We speak, by
example, of the first, of the second, of the third, and so on. When a
A variable ranks individuals, it is then an ordinal variable.

The nature of the ordinal variable differs from the nature

of the nominal variable. The values of the variable
ordinales are arranged in an order of magnitude, so well
that we can talk about a continuum, a progression
in the values. On the contrary, the values of a variable
nominal cannot be put in an order
any: there is no underlying continuum.

Interval and ratio variables are strangely similar. They

share a typical characteristic: the distance between each of the values
the continuum is equal. Thus, the basic arithmetic operations (addition,
subtraction, multiplication, and division) are allowed. This is not the case
nominal and ordinal type variables.

It is not surprising to note that statistical analyses are by far more

numerous for proportion variables or at intervals than for the
nominal and ordinal variables. And more powerful too!

In another module on measurement, a section will help you to make the

distinction between the four types of variables. This section is taken up here
underneath.
Page5 Introduction to Research Donald [email protected](506) 858-4886 Center of
Research and Development in Education (CRDE) University of Moncton, N.B. Canada E1A 3E9
 4. How do we determine the level of measurement?

First question
Can the values of the variable be arranged in an order?
any intrinsic, a progressive continuum, from the smallest to the largest?
If not, it is a nominal variable.
If yes, it is not a nominal variable. Move on to the second question.

Second question
Are the values of the variable equidistant from each other?
If not, it is an ordinal variable.
If yes, it is not an ordinal variable. Move on to the third question.

(3) Third question

Does the zero value of the variable represent an arbitrary value instead?
Is it a natural absence of the phenomenon?
If not, the variable is a proportion variable.
If yes, it is an interval variable.

5. One characteristic, several variables

A characteristic can be measured at various levels. Let's take income.
as an example.

At the nominal level, income will take on two different values:

no income
any income

At the ordinal level, income will correspond to various categories.

ordered in the following way:
low income
average income
high income

In terms of proportion, the income will be calculated in dollars.

The scale will range from 0 dollars to infinity.
As the absence of income is possible,
We can therefore qualify this measure as proportionate.

A measurement scale is ordinal when the levels are not equidistant.

between them. These numbers assigned to each of the levels of a measure
Page6 Introduction to Research Donald Long [email protected](506) 858-4886 Center of
Research and Development in Education (CRDE) University of Moncton, N.B. Canada E1A 3E9
Ordinals are nothing more and nothing less than indicators of position: they tell us
provide information on order, rank. Let's take the following example.

Three students in a class obtained respectively the

scores of 92, 81, and 79 in a French exam. They
thus respectively 1er, 2eand 3andto this test.
However, the difference between the 1erand the 2eis 11, while
that the distance between the 2andand the 3eis 2.
The ordinal variable (1er, 2andand 3e) is therefore less precise
that the interval variable (92, 81, and 79).

A variable becomes ordinal when the scores assigned can be ordered.

individuals according to a continuum, a progression, a hierarchy of values, without
even if the distance between these values is equal. Social classes are one
example: the lower class, the middle class, and the upper class. The gap between
each one is certainly not the same. One cannot say that the upper class is
twice as good as the lower class: no arithmetic operation is
not possible either with an ordinal measure. A strong student in mathematics does not
it is not twice as much as a weak student.

The interval variable is also ordered according to a hierarchy of values.

but these values are equidistant from each other. Arithmetic calculations are
So possible. This scale also includes an arbitrary zero. The
Celsius thermometer contains a zero, but it is arbitrary. The quotient
Intellectual is a measured interval, just like the Likert scale.

The proportion variable is similar to the interval measure, but zero

is real. It corresponds to a reality, to an absence of the phenomenon. If we measure
the level of education through years of schooling, this measure can be
qualified as proportion simply because it is possible not to have
No education. The same applies to income measured in dollars.

The distinction
between the interval scale and the ratio scale
is real, but it is not cumbersome, nonetheless,
because the majority of statistical analyses apply
to the two types of variables.

The gap between the nominal measurement and the three other levels of measurement is significantly larger.
important and more significant. Few statistical analyses are applicable to the measurement

Page 7 Introduction to Research Donald [email protected](506) 858-4886 Center of

Research and Development in Education (CRDE) University of Moncton, N.B. Canada E1A 3E9
nominal. On the other hand, even for ordinal measurement, several statistical analyses are
possible.
6. Mathematical gap and semantic gap
Here is an example of a variable with various increasing categories that has been designed
to measure the frequency of occurrence of a phenomenon:

0 Never
Sometimes
Quite often
Often

Each individual receives a score from 0 to 3. The group average will therefore vary.
from 0 to 3. The figures associated with each of the descriptors progress in a way
uniform: the mathematical gap between each of the categories is equal and
The mathematical gap between two contiguous categories is the same.
for any other pair of adjacent categories.

However, the descriptors as such are indeed imperfect and

can, in themselves, not truly express categories
equidistant. The semantic gap between the descriptors is not
necessarily equal between all contiguous pairs. It is necessary to
to cope with it, as it proves impossible to describe some
growing categories with precision. These descriptors serve
At best, landmarks. The descriptors include
the imperfections that we must accept,
for lack of better options.

In any case, it is essential to remember that the four categories

The above croissants are, theoretically, only four reference points on a
continuum ranging from 0 to infinity. This variable ultimately results from a
arbitrary and imperfect découpage. Nevertheless, there is a progression of a
category to another. This progression may not be uniform from one category to
the other. There is therefore some imprecision in the measurement, an imprecision that is difficult
to circumvent or to eliminate.

7. The simple variable and the composite variable

Certain characteristics are measured using a single variable, by

example age, sex, heat, weight, type of locality of a residence or
Page 8 Introduction to Research Donald [email protected](506) 858-4886 Center of
Research and Development in Education (CRDE) University of Moncton, N.B. Canada E1A 3E9
the schooling. However, for many characteristics, we resort to several
indicators to obtain a measure that is more appropriate and more comprehensive than we
let's call dimension. It is the composite variable.

You notice that the composite variable has various names: scale,
dimension, index, constructed variable, and many others. It changes from
name according to conventions and whims, but its nature remains the same. The
more often, the term indicator accompanies the term dimension. The table
The following shows that several indicators can define several dimensions.
You also understand that if there are as many dimensions as indicators,
therefore there are no dimensions...

Indicator Dimension
o
Indicator n 1
Indicator no2 Dimension A
Indicator no 3
Indicator no4 Dimension B
Indicator no 5
Indicator n6

Each of the indicators of the composite variable constitutes in a way a

specific aspect. Some of these aspects can be grouped under a
dimension. Statistical techniques and analyses are used to
to ensure, for example, that each of the indicators actually measures the
dimension (internal consistency test in combination with analysis
factorial).

These indicators are not considered in themselves as variables. Taken in

All these indicators measure rather a variable that we have decided.
of calling dimension. Most personality and performance tests,
attitude, motivation, and others, are products of the statistical approach
which consists of grouping homogeneous indicators under the heading of the
dimension. These tests often include several dimensions.

If schooling can be measured using a single variable, the status

socioeconomic is more demanding and often encompasses several
indicators. We need to obtain information about the individual's profession,
his salary, his education, and so on. A composite score will be made up of
merging these indicators.

Page 9 Introduction to Research Donald [email protected](506) 858-4886 Center of

Research and Development in Education (CRDE) University of Moncton, N.B. Canada E1A 3E9
 A common mistake is to believe that a characteristic
may be measured reliably by a single indicator.
To understand a phenomenon well
in all its components,
It is essential to define these components and measure them.

A phenomenon
complex is measured

by
a together complex
of
variables

Before you proceed to the next section, it is important to remind yourself that
Polit and Hungler (1995) argue that the research aims to understand
why and how the values of a variable change and how they are
associated with the different values of other variables.

8. The independent variable and the dependent variable

If there is a dependent variable, there is ipso facto a variable

independent. These two go together like heads and tails. The variable
independent, it is the variable manipulated by the experimenter. The variable
dependent, it is the variable that is affected by the independent variable.

If we compare men and women in terms of their

job satisfaction in a factory, the independent variable
it is gender, while the dependent variable is the
job satisfaction. The research question is the
next: What is the effect of belonging to a gender
about job satisfaction?

Sex and job satisfaction are variables that exist by themselves.

the same. They become independent or dependent according to the hypothesis of
research that establishes their status and their mutual relationship.

Page 10 Introduction to research Donald [email protected](506) 858-4886 Center of

Research and Development in Education (CRDE) University of Moncton, N.B. Canada E1A 3E9
 A dependent variable in a situation
can become an independent variable
in another situation.

For example, job satisfaction (dep.) can be analyzed by gender

It would also be possible to analyze absenteeism among the staff.
(dep.) based on their job satisfaction (ind.). It is not just the effect
of an independent variable on a dependent variable that is interesting. We
can analyze the effect of one or more independent variables on one or
multiple dependent variables.

In order to avoid unnecessary confusion, we will study this relationship between variables.
independent and dependent variables through four issues of
search.
Problem no 1
Is there a relationship between the level of education and
Question the attendance of sociocultural activities?
There is no relationship between the level of education and
Hypothesis the attendance of sociocultural activities.
We are trying to find out to what extent the socioeconomic level
Characteristic influences the number of times that individuals attend
to sociocultural activities.
V. Independent V. Dependent
Level of education Attendance

Problem no2
Is there a relationship between the level of education and the
attendance at sociocultural activities, between salary and
Question
attendance at sociocultural activities, and between age
and the attendance at sociocultural activities?
There is no relationship between the level of education and the
Hypothesis No. 1 attendance at sociocultural activities.
There is no relationship between salary and participation in activities.
Hypothesis no 2 socio-cultural.
There is no relationship between the age of individuals and attendance.
Hypothesis no 3 socio-cultural activities.
We seek to determine the influence of each of the three variables.
independent on attendance at cultural activities. Three
Characteristic
separate analyses will be necessary. We do not do
no link between the independent variables.
V. Independent V. Dependent
Level of education Attendance
Salary Attendance
Age Attendance
Page 11 Introduction to Research Donald [email protected](506) 858-4886 Center of
Research and Development in Education (CRDE) University of Moncton, N.B. Canada E1A 3E9
 Problem noo3
Is there a relationship between education level, the
Question salary, age, and gender of individuals
and their participation in socio-cultural activities?
There is no relationship between the level of education and salary,
Hypothesis the age, and the gender of individuals and their participation in activities
socio-cultural.
The four independent variables will be considered in one.
Characteristic and even statistical analysis. We seek to establish a link between
their combined effect on attendance.
V. Independent V. Dependent
Level of education
Salary Attendance
Age
Sex

Problem no4
Is there a relationship between education level and salary?
the age, the sex of individuals and their participation in activities
Question
sociocultural on one hand, and the number of responsibilities
what community ones do they have, on the other hand?
There is no relationship between education level, salary, age,
the sex of individuals and their participation in activities
Hypothesis
socio-cultural, on the one hand, and the number of responsibilities
community ones they have, on the other hand.
We are looking to establish a relationship between several variables.
independent (as a set) and several variables
Characteristic dependent (like another set). Multivariate analyses
will be necessary because several (2) dependent variables
will be considered simultaneously.
V. Independent V. Dependent
Level of education
Salary Attendance
Age Responsibility
Sex

What distinguishes a univariate analysis

? of a multivariate analysis?
A statistical analysis that only involves one dependent variable
R it is univariate. A statistical analysis that focuses on several
dependent variables are multivariate.

Page 12 Introduction to Research Donald [email protected](506) 858-4886 Center of

Research and Development in Education (CRDE) University of Moncton, N.B. Canada E1A 3E9
 Exercise no1
Independent and dependent variables

Directives
For each of the thumbnails
determine
where the independent and dependent variables are.

Variable 1. With the regular reading method, children of

First year students can read an average of 20 words per minute.
On the other hand, the new speed reading method allows for
other students at the same level to reach 28 words at the
minute.
Independent
Dependent

Variable 2. Thanks to seat belts in cars, there are

less mortality on the roads, especially during the season
wintery.
Independent
Dependent

Variable 3. Global warming is reflected by

atmospheric phenomena such as floods, the
torrential rains, hurricanes, and tornadoes.
Independent
Dependent

Variable 4. Working relations in the factory are not improving much. We

noted this deterioration since the suspension of a
hundred workers three months ago.
Independent
Dependent

5. Television advertising by breweries during the matches of

Hockey really boosts their sales figures.
Independent
Dependent
Variable 6. The number of graduates from my school
Page13 Introduction to therdeciherchehDoéaldd Lo5ng%lod
ngd@u i mol ncton.cia(d506)8i5è8-4886 C eébetweendIel
Research and Development in Education (CRDE) University of Moncton, N.B. Canada E1A 3E9
 secondary has fallen by 5% over the last three years. It
seems that the main cause of this decrease is the
decline in the birth rate.
Independent
Dependent

Variable At the newsstand, there are plenty of magazines. I am

ask who buys the most, men or women
magazines.
Independent
Dependent

Variable 8. Single people go out less often to the

cinema halls than married people.
Independent
Dependent

Variable 9. This automobile manufacturer notes that its model

sport is preferred by men, especially those who are between 20
and 30 years old who are not married.
Independent
Dependent

Variable Do female students read more romance novels?

than girls the same age, but who work full time?
Independent
Dependent

Variable 11. Absenteeism among high school students seems to be

more common among 8-year-old boys eyear that have already repeated
a year.
Independent
Dependent

Page 14 Introduction to Research Donald [email protected](506) 858-4886 Center of

Research and Development in Education (CRDE) University of Moncton, N.B. Canada E1A 3E9
 Exercise no2
Independent and dependent variables

Directives
Find an independent variable
to each of the following dependent variables.

Spelling errors among secondary school students

The attendance of campers at Fundy Provincial Park

The job satisfaction of the staff at the corner grocery store

Absenteeism in a fish packaging factory

The number of educational broadcasts listened to per week

The number of girls enrolled in the faculties of science

The number of hours spent browsing the Internet

School dropout among boys

the sale of sporty-looking automobiles

Racial prejudices

Page 15 Introduction to Research Donald [email protected](506) 858-4886 Center of

Research and Development in Education (CRDE) University of Moncton, N.B. Canada E1A 3E9
 Exercise no3
Independent and dependent variables

Directives
Indicate three independent variables
who could influence
the dependent variable.

Problem no1
V. dependent The number of surviving drivers
in a car accident
V. independent
V. independent
V. independent

Problem noo2
V. dependent The voters' preference for Mr. Z,
candidate for the position of deputy
V. independent
V. independent
Independent V.

Problem no3
V. dependent The indebtedness of students
in the university environment
Independent V.
V. independent
Independent V.

Problem no4
V. dependent The average amount spent by an individual
at the Montreal Casino on each visit
V. independent
V. independent
V. independent

Page 16 Introduction to Research Donald [email protected](506) 858-4886 Center of

Research and Development in Education (CRDE) University of Moncton, N.B. Canada E1A 3E9
 Exercise no 4
Independent and dependent variables

Directives
Find an indicator
to measure each of the concepts.

The power
economic
of an individual
The yield
school
of a student
The production
of a factory worker
production of nails

Wealth
of an individual
The good understanding
with friends
Profitability
of a herd
dairy cows

Page 17 Introduction to Research Donald [email protected](506) 858-4886 Center of

Research and Development in Education (CRDE) University of Moncton, N.B. Canada E1A 3E9
 Self-esteem
The benefits
of the walk
daily
Dissatisfaction
at work

Stress
The influence
pairs
on the choice
of his clothes
The protection
against the MTS
The regime
food
The influence
of an advertisement
in the newspaper
The popularity
of a university

Page 18 Introduction to Research Donald [email protected](506) 858-4886 Center of

Research and Development in Education (CRDE) Université de Moncton, N.B. Canada E1A 3E9
 Generosity
towards
community works
The yield
of a pilot
the car race
The yield
students
first year
in reading

The yield
of a student
in spelling

Page 19 Introduction to Research Donald [email protected](506) 858-4886 Center of

Research and Development in Education (CRDE) University of Moncton, N.B. Canada E1A 3E9