I

1 What is a research
question and why is it
important?
Jonathan D.J ansen

OUTCOMES FOR THIS CHAPTER

After studying this chapter, you should be able to do the following:

• State reasons why researchers need research questions

• Describe qualities of good research questions
• Identify criteria that mark good research questions
• lllustrate the differences between research questions and interview
questions
• Give examples of poor and good research questions, and of main and
subsidiary questions

1.1 INTRODUCTION

Any research - be it a small independent study for an

organisation, a mini-research project for an undergraduate
class or a formal thesis or dissertation for a higher degree -
stands or fallsby the quality of the research question.'
Research questions should be original, and research should
not have been conducted on your specific research question
before. Moreover, the research question should be clear and
focused, and many literature sources should be integrated to
help you formulate your distinctive question (What mak es a
good research question? 2014). The research question
identifies the phenomenon you wish to study (The
relationship between the research question, hypotheses,
specif ic aims,and long term goals of thep roject, 2003). It
should, typically, appear right at the beginning of your
research report since the research question isthe single most
important measure of whether your research is sound or not.
Identifying and formulating a research question isvery
difficult - aswill become clear in this chapter -and it is a
useful investment of time to make sure that the research
question you choose is a good one. But before you dive into
your research question, it might be useful to think about the
broader problem statement under which the question(s) will
fall.
 The pro blem statement in research
p
Before posing a specific research question, it is useful to state
what the research is about. Consider the following:

• My research sets out to determine the relationship

between class size and student achievement in primary
school education in developing countries.

N ate that the statement clarifies the key variables in this

study, namely, class size (the independent variable) and
student achievement (the dependent variable). N ate also that
the problem statement specifiesthe level of schooling for the
inquiry -primary school education. The specification is done
on purpose because the nature of the relationship between
the vaiiables does seem to vary from early primary education
to high school education , for example. Itis important to
notice that the statement delimits the context for the
investigation -developing countries. This is also done on
purpose since there is likely to be a diff erence in student
achievement outcomes depending on the levels of resources
available in r,elation to rich versus poor nations, for instance.

While the problem statement clarifies the broad direction of

the research, you still need a specific research question(s)
that indicateswhat it is about the relationship between class
size and student achievement you wish to investigate .Asking
a good question does not come easily.

The habit of asking g ood questions

It is difficult to ask good questionsin research if you are not
in the habit of asking good questions in life. Everyday life
brims with puzzles. Young children especially are intrigued
by life's puzzles and are in the habit of constantly asking
questions. How do birds know to fly south for the ·winter?
Where do we come from? Why do people float in space? Are
we alone in the universe?

As you grow older, there are still as many questions as in

childhood. \"1hy is there so much poverty in otherwise rich
nations? Why do young people knowingly engage in high­
risk sexual behaviou r? \.vhat happens when we die?
Similarly, educational researchers occupy themselves with
hundreds of questions that defy simple explanation . Here are
a few:

• Do smaller classes lead to higher gains in student

I
achievement?

• 'Why, in most societies, do more girls than boys drop out of

high school?
• How do successful principals organise their schools?

In this chapter I introduce students who wish to embark on

small or large research proj ects to the art of identifying and
posing a good research question.

Some of the issues concerning research questions that I

address in this chapter are the following:

• 'Why research questions are important

• The features of good research questions
• How you will know whether you have a good research
question
• How research questions differ from field questions (such
asinterview questions)

• 'What constitutes exampl es of poor and good research

questions
• The different categories of research questions
• The difference between main research questions and
subsidiary (secondary) research questions

• The iterative nature of research questions

• The global character of research questions

t.2 WHY DO YOU NEED A RESEARCH

QUESTION?

The research question specifieswhat intrigues you and

focuses on what you will [Link] becomesthe beacon that
guidesyou over months or even years of research as you
strive to find answers to this thing called "the research
question ". In practical terms, you need a research question
for at least two reasons.
 Your research question should never be used to enable you to
p "prove"something. It merely directs you to appropriate
research literature and other [Link] question tells
you what research literature to read and how to narrow down
your bibliogr aphic search. For example if you are researching
the question "How do successful principals organise their
schools?"then you are likely to focus only on the literature
concerned with school leadership and management rather
than the literature on learning styles or models of teaching.
Thishelps the research librarian . Ifyou come in and say, "I
am interested in studying how successful principals organise
their schools", she or he then has a clear idea of where to
direct you in the library stacks or on the internet or the
specialist periodical shelves. Without a clear question, you
could waste valuable time tracking down the appropriate
literature pe1tinent to your study.

A good research question provides you with af ocus for your

data collection. It prevents you from drifting from your
original purpose and keeps you focused on your starting
[Link] does not mean you should slavishly follow your
research question irrespective of what you find in the course
of your inquiry; often your work in the field collecting data
might throw up other or related questions of interest that
you could not have imagined when you started the research.
You might then either wish to modify or discard your original
research question or questions. However, be conscious of the
fact that every time you drift into other questions there is a
loss of time and resources and you may well get so distracted
that you never get to finish your original research [Link]
the more reason to ensure that you start with a strong
research question.

For these and many other reasons, it is especially important

that the researcher gets the research question right. You
could waste valuable time and resources pursuing something
only to find out much later that you started off, right at the
beginning, with a weak research question. What, then, is a
good research question?
1.3 WHAT ARE THE QUALITIES OF A GOOD
RESEARCH QUESTION?

A good research question has many different features and

ten of these are described briefly below. Consider the
following research question:

Do smaller classes lead to g reater g ains in student

achievement?

Using the above example, notice that a good research

question is the following:

1. Concise. It comes to the point, focusessharply on the

issue of interest, and is stated in a cogent and specific
way. In the example, the research question is posed in
one sentence and in one line.
2. Clear (unambiguous). Itleaves no doubt as to what is
being proposed for study; it is understandable by even
non-specialists (in this case, those outside the field of
educational research); it does not require further
explanation.
3. Operationalisabl e. It can be executed or implemented in
practice. In the example, the researcher simply has to
find measures of class size and measures of studen t
achievemen t - and then determin e the relationship
between them. In other words, such a study is reasonably
easy to implement.
4. Open-ended. Itis posed in a way that suggests no obvious
answer; it is therefore inteHectually honest, and does not
expect or lean towards an obvious answer. It could be
that small classes (with few students) produce higher test
results (the most common way to measure achievemen t),
but we do not know for sure. A good question therefore
lacks bias and allows for any kind of result to emerge
from the study.
5. Elegant. It is posed simply and yet in a way that is
parsimonious (limited words) and conveys rich meaning.
In the above example, the research question asks about a
relationship between only two variables - class size (the
number of students in a class) and student achievement
(academic performance). It is as simple and elegant as
that.
6. Timelv. It is a mark of a l!:ood auestion that it addresses
 6. Timely. It is a mark of a good question that it addresses
some pressing issues of importance at the time. The
implementation of a n ew curriculum needs to be
understood;the attitudes of teachers towards a new
policy making corporal punishment illegal needs to be
probed; the growing levels of violence in schools need to
be theorised about; and so on. Of course, some research
questions are timeless and some might not address an
immediate practical need but speak to a long-standing
intellectual puzzle.
7. Theoretically rich. It leads to other questions. It can be
compared to peeling an onion: you discover that there
are many more layers to the onion as you go deeper. Does
class size matter at all, or isthere a lower and upper limit
of numbers in a classbefore size starts to impact on
student achievement? Are some teachers better at
teaching large classes than others? Does it not matter
who is in the class -such as children from poor families
or children with learning disabilities? In other words,
even though the research question isposed simply and
elegantly, it immediately raises and generates other
interesting research questions - it is, in short, a
theoretically rich question.
8. Puzzle features. It addresses some deep puzzle in the
literature or complex problem in society, something to
which the answer is not clear cut. A good question
therefore evokes attention; it speaks to a question which
scholars are eager to find answers to. For example
medical researchers are interested in explaining why
some people live with the Human Immunodeficiency
Virus (HIV) but never develop Acquir ed
Immunodeficiency Syndrome (AIDS) even when they are
not using medication. A research question of this order
addresses a complex puzzle with a high level of public
interest.
9. Self-explanatory. It speaks for itself .A good question
does not need further elaboration. It iswritten in such a
way that the topic of interest is clearly specified and the
variables of interest easily identified. While the use of
specialist language is sometimes necessary, a mark of a
good research question is that a larger group (i.e. beyond
the subject specialists) can understan d the question.
10. Grammatically correct. Itis attentive to accuracy of
 grammar. A sound researcher does simple thmgs well,
has an eye for detail and is meticulous about the use of
words. External reviewers of research repo1ts tend to
take a dim view of poorly posed research questions in
whic.h simpl P. thingslikP. gmm m:itic.::i l ::[Link].y ::irP.
overlooked.

Above all, a good research question has panache. The

Concise Oxf ord English Dictionary (2011) definesthe word
"panach e" as follows:"Flamboyant confidence of style or
manner; a tuft or plume of feathers, especially as a headdress
or on a helmet".

A research question with flair or panache is therefore

provocative, interesting, current and stimulating. It makes
people sit up and take notice, without the researcher going
overboard towards a sensationalist kind oflanguage.
Questions like "Do teachers practise outcomes-based
education in their classroom?"are routine, everyday.
Moreover, this question cannot be answered by a simple
"yes" or "no" response ( What makes a research question
answerable?, 2006). Simple "yes" or "no" questions are
regarded as low-level and not acceptable in scholarly
contexts. But questions like "What are the interactional
effects of diff erent assessment policies on teachers'
practices?" -well, this kind of question begins to show
panache.

While you might have a well-posed question, following the

criteria given above, you also need to know that your
research question makes sense to other researchers in the
field - in this case, in the field of educational research. How
will you know this?

i.4 HOW WILL YOU KNOW IF YOU HAVE A

GOOD RESEARCH QUESTION?

It is quite possible that you came up with a well-posed

research question but that somebody else had already done a
similar [Link] would do well to remember the maxim of
that wise man, King Solomon: "There is nothing new under
the sun."So, how do you know that your research question is
original and that it will make some new contribution to what
we already know? Where do you go next? I suggest three
The scholarly literature
For researchers, the answer is consulting the research
literature. You read and read what othershave said about the
topic - in our example, about the relationship between class
size and student achievem ent. You go to the library, you surf
the internet and you go to conferences where experts on this
subject tend to assemble. You read what they have wlitten on
the subject of dass size and student achievement.

This kind of research is typically published in academic

journals - some might be online and accessible from your
computer, others might be housed as hard copies in your
library. Some might be general education jou rnals in which
thiskind of research (on classsize and student achievement)
appears; others might be specialist jou rnals dealing with
your topic. In either case, your task is to find out what has
already been researched and published on your topic.

A useful hint is to look for journals that summarise or

synthesise res·earch on specific topics - this saves you time,
and you can then find in one place much of the research
published on class size and student achievement, for
example. It is especially important, when you find journal
articles on your topic, to use the references or bibliography to
guide your further reading on the topic. Two international
j ournals that specialise in synthesising the literature on
special topics in the field of education are the Review of
Educational Research and the Rev iew of Research in
Education -both publish ed by the American Educational
Research [Link] is a useful place to start.

Whether you find journals or books,j ournals that synthesise

research or general education journals, your task is to find
out what has already been researched on a topic (such as the
relationship between class size and student achievement)
and then to determine whether and how you can add any
value to what we already know.

Expert opinio n
In addition to thejournals, you would do well to also consult
a research supervisor or mentor or simply an expert in the
field (ideally, one person could fulfil all three roles) and ask
this person directly: "Is this draft research question of mine
any good? Will it add value to the field of study? Have other
people already studied this topic and answered this
question? Isthis question worth pursuing?" Ifyou have a
strong and competent supe1visor, she or he could be of great
value not only in answering these questions but in guiding
you through the appropriate readings and research to answer
them.

Specialist conf erences and seminars

By attending education research conferences, you also find

out what people are studying and who works in your field of
interest, such as the relationship between class size and
student achievement in our example. One of the most useful
sites for examining conference agendas is the website of the
American Educational Research Association (AERA;
[Link] and the specific areas of that site called
SIGs or Special Interest Groups. Normally there is at least
one SIG that speaks to, or is organised to bring together,
researcherson a similar [Link] Annual Conference
Programme of the AERA then lists topics, speakers and their
addresses and you can usually expect to find your area of
research being addressed.

So let us assume that you have read thejournals, consulted

the experts and attended or reviewed specialist conferences.
You are now in a position to ask again: Will my research
question add anything new or valuable to what we already
know on, say, the subject of class size and student
achievement? Here are some guidelines for answering this
question:

1. Doesthe existing research speak to my context, for

example rural South Africa n schools?
2. Doesthe existing research shed light on this relationship
in contexts where teachers are mainly unqualified?
3. Do the research findings from different studies contradict
each other? Ifso, how can my research bring greater
clarity to these conflicting findings?
4. Are the existing research findings outdated and perhaps
not capturing more recent developmentsor trends in
education?
5. Does the existing research shed light on this relationship
in contextswhere textbooks and other learning materials
are in short supply?
6. Does the existing research capture the outer range of
classsize (such as more than so children in a class), as is
common in developing countries?
7. Does the existing research confuse class size with
classroom size? That is, having so children in a small
room has different implicationsfrom having so children
in a large classroom space.

These are simply examples of the kinds of questions that you

can ask of the literature reviewed and the expertise consulted
as you set out to determine whether or not you have a good
research question .

i.5 HOW ARE RESEARCH QUESTIONS

DIFFERENT FROM INTERVIEW
QUESTIONS?

A research question is the broad , general question to which

you seek [Link] our example again, the research
question is:
Interview
What is the relationship between class size and stud ent questions break
achievement? the research
question into
This is different from the questions that you might put into smaller sections
to obtain
an interview schedule or that might appear in a specific answers
questionnaire for the purposes of data collection. For from participants:
example let's say the researcher decides that the way in
which she is going to collect data (information) on this
question is to interview 100 Grade 9 teachers. With this
research question in mind, the researcher might generate a
few smaller interview questions, such as:

1. How many children are there in your class?

2. Do you prefer to teach smaller classes or larger classes?

3. Do you find it easier or more difficult to teach large

classes?

These are examples of interview questions (sometimes called

field questions) posed directly by the researcher to the
teachers (or interviewees), and must be distinguished from
the research question, which serves simply to focus the
broader research project.
t.6 WHAT ARE EXAMPLES OF POOR RESEARCH
QUESTIONS?

Scholarly research implies certain standards about

acceptable or unacceptable research [Link] follmving
are examples of research questions that are not, in general,
regarded as high-level questions appropriate for scholarly
inquiry.

Developmental questions are not normally regarded as

appropriate research questions, for example: "How can first­
year students become better teachers?" This question is valid
for interventions that seek to improve teacher preparation,
but it does not fit well with open-ended inquiry in which the
researcher seeksto understand rather than simply change a
situation.

Planning questions are seldom good research questions,

for example: "What can be done to design a curriculum for
ICT (information communication technologies) in schools?"
This is a planning task, in which curriculum developers
should sit down and design a curriculum . It is not normally a
research-d1iven task, even though it could benefit from the
results of research.

Obvious-answer questions are definitely not good

research questions, for example: "Will poor children learn
better if they receive nutritiou s meals at school?"The answer
is obvious and does not require the investment of research,
especially since on thisparticular question common sense
and p1ior research have shown the links between good
nutrition and academic performance.

Needs analysis questions are not advan ced research

questions in academic contexts, for example: "What are the
IT needs of mentor teachers?" No doubt an organisation
might need to find answers to this question, but it involves
fairly simple questionnaires that will probe needs and do not
yield the kind of data that makes for advanced or scholarly
research.

Preaching questions are definitely unacceptable, for

example: "What can be done to ensure that children do not
have sex before marriage?"This might be an important
development task to faith communities or to parents
concerned about HIV and AIDS, but questions posed this
way are best resolved through churches or non­
governmental organisations, not through advanced research.

Low-level questions are not suitable for advanced

research, for example: "What medical aid scheme do most
teachers belong to?" This question clearly has value for
marketing agencies dealing with medical aid schemes but
has very little value for theorising about teachers and their
health choices; it is a low-level question because the answers
are simple statements such as:"40 per cent belong to
medical aid X and 30 per cent to medical aid Y". There is
little one could do with this kind of information in
educational research.

Committed questions are, like preaching questions, of

little value in research, for example: "Why are teachers so
stubborn when it comes to change?"This question has
already decided that teachers refuse to change because of an
attitude or personality problem, like stubbornness .Why do
research if you already know the answer?

Solution-seeking questions hold low value in serious

scholarship, for example: "What can be done to stamp out
drugs in schools?" Again, the moral task is important but this
does not suggest a research question, at least not in the way
the question is posed. Good research questions try to address
complex problems in waysthat leave the answer open to
many possibilities .

1.7 ONE EXCEPTION2

I
There is something called applied research, where
[Link].::[Link] ask [Link] [Link] [Link] to proviclP. cl:it::i that
could be used to improve a situation or solve a problem.
Researchers working in applied contextssuch as computer
studies or community development often initiate research
with such practical [Link] is acceptable research and
requires the same level of rigour and care in the formulation
of the research questions as with traditional academic
research. Consider the following two examples of good,
applied research questions:

• What are the online behaviours of students doing their

first degree through elearning?
• What are the online behaviours of students doing their
first degree through elearning?
• How can knowledge of these behaviours be used to
improve the design of online software?

There is a particular kind of applied research in teacher

education called action research. Here the researcher
becomes part of the community (e.g. teaching with existing
teachers), doing research in the course of working with such
teachers - in this case, in order to improve the practice of
[Link] intellectual demands of posing good questions
apply here as well even though the research endeavour has
practical aims. Consider the following well-posed question
emerging from an action research context:

• How do new teachers, working alongside an experienced

teacher (the researcher), engage the new curriculum over
the course of a school year?

Exercise
It might be helpful at this point to give a range of research questions and
ask you, the student, to make a judgement,based on what has been
discussed so far, as to whether the follo\\ng are good or weak research
questions. Please take the time to give reasons for your answers.

1. Does preschool education make a difference? That is, what are the
effects of preschool education on the subsequent academic
performance of children·/
2. Since smoking causes cancer, what can be done to eliminate
smoking in high schools·/
3. How can more schoolsbe built in rural areas?

I
4. Is there a correlation between student achievemen t in high-school
physics and their performance in first-year university physics?

5. What are the effects of trauma on the learning capabilities of

children from war zones?
6. How do teachers use science textbooks in rural, underresourced
schools as compared to urban,well-resourced classrooms?

7. What training do teachers need to design appropriate examination

questions·/
8. How do newly graduated teachers navigate their way through the
first year of teaching?
9. What do women academics working at large research universities
identify as baniers to promotion·/
10. What are the resources that speech-disabled adolescentsdraw on in
order to enhance their learning performance?
t.8 MOVING BEYOND THE BASICS: THREE
CATEGORIES OF RESEARCH QUESTIONS

Once you have learned the basics of posing good research

questions, it is time to ask what kinds of questions are
regarded as really outstanding research questions. In
general, "how" questions and "what " questions dwell on the
descriptive; on the other hand, really excellent questions
go deeper and focus on the explanatory. In research
generally, explanatory questions hold higher value and
significance than descriptive questions. Consider the
following question, which dealswith the announcement by
the South African government that corporal punishment
(caning) in schools is now illegal.

A descriptive question:Do teachers still practise

corporal puni shment?

This is a good question, sin ce it is worth finding out whether

teachers have changed their behaviours since the new policy
was announced. But it is a very basic question, nonetheless.

An explanatory question: Why do schools persist in

ad ministering corporal punishm ent despite thef act
that such pr actices are illegal?

This question goes beyond description (what happens) and

seeks to explain behaviour, not at the level of teachers
(individuals), but at the level of schools (institutions). A
student who researched this question came up with this
answer: corporal punishment continues because of the ways
in which such punishment has become institutionalised in
South Africa (in the same way that most South Africans

I
support capital punishment for murder), not because of the
errant or devious behaviours of individual teachers. In other
words, "why" questions invite theory and theoretical
explanation.

Of course, combining the two questionsin one research task

would be ideal in a major study on corporal punishment .The
point for now is to recognise the value of both kinds of
questions, but to indicate the added significance of "why"
questions in educational research.

There is another type of question , the exploratory

question, which is used under specific conditions. The
researcher might not yet be sure about the nature or extent
of a complex problem, and might first wish to do a limited,
initial research study before launching a more in-depth,
long-term study. For example a researcher might have a
hunch that students on former white South African
university campuses live socially segregated lives even
though the campuses are open to all. Before making this a
major study, he might want to explore the dimensions of this
problem to see whether thisis serious enough to merit
further inquiry. This situation calls for an exploratory study,
and could be framed as follows (note the careful, tentative
phrasing):

An exploratory question:To what extent, if any, are

stud ents onjormer white campuses living socially
seg reg ated lives?

Once again, it is possible to combine exploratory, descriptive

and explanatory questions within a single research project,
in the order given.

i.9 MAIN AND SUBSIDIARY QUESTIONS

I started with the notion of a single research question . Of

course, a research proj ect could have more than one research
question - although too many questions tend to disperse the
focus of the project. Another way to think about this is that
you have a main research question and subsidiary questions.
Consider the following example.

Main research question

1. How do South African teachers use learning materials in the new

curriculum assessment and policy statement document?

I
SubsidianJ questions

1.1 What are the kinds oflearning materialsavailable to teachers in

South Af 1ican classrooms?
1.2 How are learning materials distributed between urban (mostly well­
resourced) schools and rural (often underresourced) schools?

1.3 How do teachers in these two contexts (rural and urban) use the
available learning materialsgiven the demands of the new
curriculum?
Note that the subsidia1y questions are more specific and give
added focus to the main research question. Note also that
each subsidiary question relates directly to the broad er or
main research [Link] note finally that in order to
answer the main research question, it is important to start
with basic questionslike i.1, then to recognise that not all
schools have equal or similar access (1.2) and this makes it
possible to ask how teachers use the materials (i.3). The
order of questioning therefore moves from basic information
to the more advanced information suggested in the main
resear ch question .

t.10 THE ITERATIVE NATURE OF RESEARCH IN

FORMULATIN G RESEARCH QUESTIONS

In a long-term research project, the research question(s)

often change repeatedly. You seldom start and end with the
same research question; this is especially the case in
qualitative research (see Chapter 15), which is more open
ended than quantitative (such as statistical; see Chapters
8-14) [Link] researchers trawl through the literature,
they might find that the original research question is too
broad or too narrow; they then go back and amend the
research question. As they enter the field for the purposes of
data collection, they might find that the research question is
too ambitious or that the context does not allow access to the
specific data sought. Again, researchers might decide to
modify the original research question an d, in extreme cases,
to drop it [Link] latter decision must be made with
great caution as the material and personal costs of
completely changing course could be quite high, especially
when the research hasto be completed within specified
[Link] going back and forth between question and
field experience, or between question and the literature, is

I
what is ref erred t o as the iterative nat ure of research.

t.11THE GLOBAL CHARACTER OF A RESEARCH

QUESTION
It is extremely important to realise that while your research
question might focus on a specific locality or local problem,
your audience is a global one and therefore the question has
to be posed more broadly bef ore you delve into the details of
the context. The following question istherefore not
acceptablefor a scholar-researcher:

What are the attitudes of teachers in the Tshwane South

school d istrict, west of the Apies Riv er, towards the new
po licy on teacher licensing?

There is nothing ·wrong with the substantive focus of the

research question. The problem is the overspecification in
the title to include a local school district, west of the Apies
[Link] research is about teacher attitudes towards
teacher licensing -and stated as such in the title, the study
resonates ·with research evel)'\'vhere in the '\.Vorld on teacher
licensing.

Of course, the researcher will later detail the specific context

or locality '\.vithin which this study of teacher attitudes "vill
take place. Even then, the researcher has the task of
justifying "vhy that particular locality \.Vas chosen, and how
knowledge of teacher attitudesin that locality adds
knowledge to the broader literature on ho'\.V teachers feel
about teacher licensing policies.

One way of judgin g whether your research question

communicat es to a broader audience is to ask whether the
question as formulated '\.vill hold appeal for or interest
researchers in other parts of the '\.Vorld .
1.12 CONCLUSION

A research question can make or break a study.\.Yithout a

research question, it is hard to maintain focus in the research
process and to know vhat kinds of literatures to consult or
data to collect. You do not always literally need a research
question in the form of a question ; you could formulate a
hypothesis to test ("that teachers practise corporal
punishment because they believe that it is the best available
means for changing student behaviour") or simply a
statement of purpose ("the purpose of my research is to
examine teacher attitudes towards corporal punishment").
The point is that you n eed afo cus, stated simply and
pointedly.

Ho vever, for beginning researchers in particula r, a research

question is the best way of formulating your purpose and
focusing your efforts as you begin the exciting process of
gathering data to ansver a compelling question.