QUANTITATIVE RESEARCH

BASIC RESEARCH CONCEPTS

doc. Phdr. Jana Harťanská, PhD.

1
 1. THE RESEARCH PROBLEM
• is the focus of the research; a problem that a
researcher wants to investigate, stated as
research questions
E.g.
1. How can we predict which students might have
trouble learning certain kinds of subject matter?
= correlation research
2. 2. How can a headmaster improve a school
morale? = interview research
i.e. info can be collected, must have empirical or
observable referents

2
 Good research questions must be:

• Feasible (time, energy, money)

E.g. How do the students at XY School feel about the
new guidance programme recently instituted in the
di3strict?
• Not so feasible: What would be the effect on the
achievement of giving each student his
microcomputer for a semester?
• Clear (what exactly is being investigated and what
the key words mean; not too broad)

3
• Significant (worth investigating);
e.g. How might answers to this question
advance knowledge in my field / improve
educational practice / improve the human
condition?

• Ethical (causing no harm, damage)

4
 Three Types of Research Problems

1. Descriptive problems = answer the question:

What is it like?, they find and describe a
situation, state or existence of a phenomenon.
The methods used:
• observation
• scaling
• questionnaire
• interview
5
E.g.
1. What is a proportion of an English teacher
teaching activities?
2. What does a typical traditional English lesson
look like?
A descriptive problem can also be evaluative
(e.g. learners learning results, effectiveness of
a new English coursebook, material conditions
of school)

6
2. Relational research problems frequently
suggest a relationship of some sort to be
investigated, i.e. two qualities / characteristics
are tied together or connected in some way.
E.g.
What is the dependence between the university
entrance test result and the successfulness of the
study at the university?
BUT: they do not explain which of the
phenomena (A or B) causes the relation.
(Is it A which causes B or vice versa?)
7
3. Casuistic research problems
= find / identify them, e.g.
• What is the influence of using communicative
language games upon English learners´ effective
filter?
• Is a non-directive educational style more effective
for creating positive learners´ attitude to teacher
than a directive style?
They use experimental method, compare
two/more groups of research subjects which
differ in one of the phenomena and statistics
(Student t-test.)

8
In any research topic, the 3 types of the RP can
be formulated:
Descriptive: What kinds of praise do
teachers use?
Relational: What is the relation between the
kinds of teacher´ s praise and learners´
performance?
Casuistic: What is the effectiveness of a
postponed praise on the learners´
performance in the comparison with an
immediate one?
Note: descriptive methods cannot have
Hypotheses.
9
 2. VARIABLES
• VARIABLE is a concept – a noun that stands
for variation within a category of objects (e.g.
chair, gender, eye colour, achievement,
motivation, running speed)

• They differ, vary – to qualify the category as a

variable. If all research subjects are identical –
no variable! They are called constants

10
E.g. Effects of reinforcement on student
achievement: 3 groups of 3 ninth-graders,
1student gets verbal praise, 1gets money,
1gets extra points

“reinforcement“ is a variable - has 3

variations, while the grade level of the
students (all are ninth-graders) is a constant

11
 Quantitative vs categorical variables

• Quantitative variables exist in some degree

along a continuum from „less“ to „more“.
- We can assign numbers to different
individuals/objects to indicate how much of the
variable they possess (e.g. height);
- how much interest they have in a subject (5 –
indicating very much, 4 much, 3 some, 2 little, 1
very little interest)

12
• They can often be subdivided into smaller and
smaller units (e.g. „length“ – can be measured
in miles, yards, feet...)

• Categorical Variables do not vary in degree,

amount, or quantity BUT are qualitatively
different (eye colour, gender, religious
preference, occupation preference, research
methods). Cannot be quantified. Have only
two values.

13
E.g. We wish to compare computerized and
noncomputerazed classrooms, the variable
involved would be the presence or absence
of computers.
Here are several variables:
1. Type of automobile owned (categ.)
2. Learning ability (quantit.)
3. Ethnicity (categ.)
4. Cohesiveness (quantit.)
5. Heartbeat rate (quantit.)
6. Gender (categ.)
14
 INDEPENDENT VERSUS DEPENDENT
VARIABLES
• One V can cause a change of another V.
A teacher´ s teaching style can lead to certain
learner´ s learning result or performance.
• Independent variable: is the cause of the
change. It is presumed to have an effect on
another variable. Not all are possible to be
manipulated. It is possible to investigate more
than one independent variable in a study.

15
• The variable that the independent variable
is presumed to effect is called dependent
variable. It changes in the dependence on
the independent variable.

16
 The realtionship between IV and DV: IV
effects DV

Independent Dependent
variable(s) variable(s)

(presumed or possible (presumed results)

cause)

17
• The result of learners is DV, teaching style is IV.
• IV (categorical) is the number of teachers, the
DV is the amount of science teachers.
• To investigate a variable, it must be
operationally defined. Operation definition
puts a variable into observable, findable,
measurable connections. It determines
operations which take place while measuring
them.
• E.g. Variable: FLL ability, OD: score of a learner
in an aptitude test

18
 3. HYPOTHESES
• It is a prediction of some sort regarding the
possible outcomes of a study.

E.g. Research question:

Will the students who are taught English by
a teacher of the same gender like the subject
more than students taught by a teacher of a
different gender?

19
Hypothesis:
Students taught English by a teacher of the
same gender will like the subject more than
students taught English by a teacher of the
same gender.

„Liking for“ E is DV, „gender of teacher“ is IV.

Extraneous variables might be: personality
and ability of teacher, student, materials
used, style of teaching...

20
 Formulating Hypotheses
Main characteristics: they express relations
between two variables!
Bad !: Good teachers use humour in the class.
Good: Teachers who use humour in the class reach
better education results than teachers who do not
use it.
In hypotheses, there are between variables
expressed differences, relations, or
consequences.
21
Differences: „more“; „more often“; „higher“;
„different“.
E.g. Second-graders prefer school more than
fifth-graders.
Relations: positive relation, negative relation, correlation.
E.g. There is a positive relation between parents´
education and the intelligence of the child.
Consequences: „if – then“; „the more - the...”
E.g. If teacher increases the number of praises, then
learners will learn more.

22
• The formulation of the H signifies the way
how it will be proved or rejected.
The H with differences uses statistical research
methods so that to find significant differences.
E.g. Student t-test -in the case of the
differences expressed in averages (average
grade, average performance/achievement ) or
CHI-quadrate – if there are the data expressed
in frequencies (frequency of teacher questions
or of student absences in a period of time)

23
 GOLDEN RULES OF A HYPOTHESIS
1. H is a statement. It is not expressed in a
question. At the end of the research we
must accept it if true, or reject it if not true.
2. H expresses a relation between two
variables.
3. H must be testable (empirically investigated).
Its variables must be measurable or
expressed in categories

24
• Frequent incorrect formulations:

- too long and complicated, unclear sentences!

- too many variables, no clear relations among them
(then they cannot be investigated, tested!)

Each research must have several hypothesis –

One broad H can have several
sub-hypotheses: H1 – H1.1
H1.2
At the end of the work, researcher must
comment and explain all the research data,
results.

25
4. QUANTITATIVE RESEARCH METHODS
Most frequent are:
• Observation
• Scaling
• Questionnaire
• Interview
• Content text analysis
• Experiment

26
Each research must be:
1. VALID – its research instrument investigates
what it is supposed to (content validity:
presence of all the needed elements +
construction validity: abstract concepts, e.g.
knowledge, ability, intelligence: Does this
research instrument investigate academic
literacy or literacy needed for life?

2. RELIABLE- means accuracy and reliability at

the same time (can be reached by the
repetition of measurements)

27
 4.1 SCALING
• Evaluative scale is an instrument enabling to find
qualities, frequency, or intensity. Researcher
expresses own evaluation by marking the position
on the scale.
• TYPES OF SCALES
1. Ordinal scales determine order of importance,
favour, occurance
Pros: easy to create, respond;
Cons: do not show the difference between the
order.
28
2. Interval / Frequency scales – there are equal
values = distances between them. Usually,
they have 3,5,7 degrees. They have an odd
number. They can also measure intensity,
proximity. Except numerical scales there are a
number of graphic scales. E.g.
In this class I feel
useful 7 6 5 4 3 2 1 useless

Cons: respondent can overvalue or

undervalue the observed features

29
3. Bipolar scales – are created by contradicting
features
Respect the following rules:
1. Use the same part of speech (mostly only
nouns, only adjectives on the both ends).
(expert 1 2 3 4 5 DK layman)
2. The second expression in the scale should not
be a negative form of the first one.
(good ex.: polite 1 2 3 4 5 rude)
3. The same expression cannot be used in the
scale twice.
4. The opposite expressions must be chosen very
carefully.
30
 Problems with “quasi-interval“ scales:

- in lexical descriptions, the intervals can be unequal.

E.g.
I like to play
always –very often – often – sometimes – never

the intervals between always and very often

and between very often and often is not clear.

31
4. Likert scales are used for measuring of
respondents´ attitudes and opinions.
On scale, respondent expresses the measure
of his agreement/disagreement.
Pros: easy to evaluate
E.g. Teacher should treat learners as a friend.
Strongly agree no disagree strongly
agree categori- disagree
cal opinion

Do not use negative formulation!

32
A typical scale has the following five frequency
stages:

always - very often - often - rarely – never

• Scales Data Processing

When more respondents involved, their

evaluations of each scale are scored and then
expressed in percentages.
33
 Example

A B C D E

Fully agree I do not disagree strongly

agree have a disagree
categori-
cal opinion

34
 A B C D E
_________________________________________

Ts should respect 71% 28% 1% 0% 0%

s´s needs.

Ts should act 16% 44% 25% 12% 3%

first of all like humans
and only then like
teachers.

Sum total in one line presents 100%.

More instruments with more scales have a higher
reliability.
35
 4.2 QUESTIONNAIRE
• Basic concepts:

Respondent: a person who fills the Q.

Questions: individual elements of the Q.
Item: the question in the form of a statement.
Administration: distribution of the Q.

The objective of the Q must be clear, concrete,

well expressed.
Thought over in its content, form, structure and
graphic form.
36
 Structure of questionnaire
• Q usually has three parts:
1. Introductory part: heading
- states the name + address of the institution
distributing the Q or the name of its author;
- states the Q´ s objectives;
- underlines the importance of the respondent´ s
answers in solving the research problem, asks
for returning;
- gives instructions how to fill it;
37
• The second part: questions
- does not have a special logic. Mostly, the
easiest and attractive qs are first;
- more difficult qs are in the central part;
- more confidential and factual qs are in the
end.

• The final part: thanks to the respondents for

the cooperation
38
 How to create questions
1. Clear qs: all respondents understand them
equally.
2. Do not use too broad qs.
3. Try to avoid expressions: some, usually,
sometimes - respondents can interpret them
differently.
4. Avoid double qs. Q has to ask only one thing.
5. Ask only qs the respondents can answer.

39
6. Qs must be meaningful for Rs
7. Create simple Qs- long ones are confusing, make
RS unmotivated and slow the pace.
8. Avoid negative expressions – can be interpreted
improperly and thus can be answered positively
– if needed, highlight the negative form. No
double negation!
9. Avoid possible prejudice in qs (wrong: Do you
agree or disagree with the headmaster´s
proposal about...).
10. Some Qs can be answered as it is socially
needed – not what the R thinks about it – have
options (sometimes yes, sometimes no, it
depends) 40
It is advisable to check the Qs´s correctness
in a probe - in a form of an interview.

• TYPES OF QUESTIONS
1. Closed Qs - offer alternatives, R has to mark
(underline, circle) the adequate answer.
E.g. Do you think it is necessary for gifted
learners to create special classes?
a) yes b) no c) I have no opinion

41
Then it is a dichotomic Q – c) option is very
important here; or use alternative: „Other
(describe, please)“= it is a semi-closed Q

2. OPEN QS give Rs freedom, do not limit

them. E.g. Explain, why you disagree with
creating special classes for gifted learners.

They are more difficult to answer and

process/evaluate. Each 0Q can have
about 15 various answers. They are used
mostly in explorative researches and probes.
42
3. SEMI_CLOSED QS offer first – an alternative
answer, then ask for explanation/clarification
in a form of an 0Q.

E.g. Do you agree with creating special classes

for gifted learners? Yes – no. If yes, why?

4. SCALE QS (see part 5.1)

43
• Variability in types of questions
is recommended! But avoid too frequent
changes of types.
• Validity of types of questions:
- factual information has high validity (age,
gendre, address, employment, education etc);
- factual info, which needs guessing, have lower
validity;
- much lower validity have Qs asking for
attitudes, opinions, motives.

44
- Anonymous Qs bring more truthful responses-
but can be considered not important and are
superficially responded.
- L-questions (lie-qs): a method revealing the R´s
tendency to „lie“. Sum score of the R´s responses
to the L-Qs enables to count his/her L-score.
• Reliability of the Q depends on more features:
e.g. on inner consistency of the Q = more Qs
asking for a similar kind of info.

45
• Length of the Q - is hard to specify (up to 45
minutes!), usually around 20 min for adults.
• Return of the Q is expressed in per cent. A
very good one is 75% but is frequently much
lower.
• Covering Letter explains the reason and asks
for filling and sending the Q back to the
researcher.

46
 5. Creating own questionnaire
Research Problem: You want to know what teaching
aids teachers use.
Task:
1. Make questions (how many qs would
you use?)
2. Use more types of questions.
3. Use various scales. What symbols would
you use? Think of the form of space for
answers.
4. Write a Covering Letter / Heading

47
 Literature
1. Gavora, P.: Úvod do pedagogického výskumu. 4.
vyd.. Bratislava: UKF Bratislava, 2008. ISBN 978-
80-223-2391-8.
2. Wallen, N.E. –Fraenkel, J.R.: Educational
Research: A Guide to the Progress. 1. vyd. New
York: McGrow-Hill, Inc., 1991. ISBN 0-07-
067945-2.

48