WHAT IS RESEARCH? Use this sheet to help you: helpsheet

Use this sheet to help you:
• understand concepts that are central to research as conducted in the
• improve your ability to conduct effective research
5 minute self test
Complete the puzzle by specifying key concepts in research using the hints
1. A form of research often
suited to careful analysis
of human behaviours and
2. A statement to be tested in
the research process
5. The ability of the research
methods to measure what is
1. A form of research that
usually involves application of
systematic “scientific” methods
3. The ability of the research
methods to produce
consistent results
4. An unpleasant condition
often experienced by
6. Also known as “information”
© The University of Melbourne 2010.
These materials were produced by the Teaching and Learning Unit, University of Melbourne. The University of Sydney
has reproduced these materials under licence from the University of Melbourne.
This helpsheet outlines the nature of research as it is generally understood in the Business
Of course, research can be understood in different ways in different disciplines. This
helpsheet, therefore, should not be seen as a guide to research in all of its possible
The information presented below is adapted mainly from Sekaran (1992), Yin (1994)
and Perry (1994), Clarke (2003), and Hurworth, (2003). For further information about the
research process see: helpsheets: The Research Process and Research Essentials. See
also: helpsheets, Case Studies 1 and Case Studies 2.
Quantitative and qualitative research: key
Academic research is often divided into what is known as “quantitative research” and
what is known as “qualitative research”. Key differences between these are summarised
Purpose: To determine cause and effect
Purpose: To describe on-going processes
relationships, i.e., between X and Y.
in the real world.
Hypotheses: These are usually devel-
Hypotheses: These are stated before the
oped and refined during the investiga-
study and tested during the study.
tion. Questions raised determine where
the study goes.
Theories: These deductively determine
Theories: These are developed by induc-
the study.
tive reasoning.
Variables: These are controlled and
Variables: No explicit variables. The
manipulated to shed light on the hy-
focus of qualitative research is to study
naturally occurring phenomena.
Data collection: It is important that data
is collected objectively.
Data collection: Data can involve
subjectivity (e.g., interaction with participants).
Design of study: This is stated at the out-
Design of study: This can change as the
set and does not change.
study develops.
Page 1
Presentation of data: This is usually
numerical (e.g., statistical).
Presentation of data: This is usually done
in the form of analysis using language
(narrative form).
Validity and reliability: This is ensured by
means of statistical tests.
Validity and reliability: This is ensured by
means of triangulated (multiple sources)
of data/evidence.
Data samples: These are carefully
chosen to represent larger populations
Data samples: Individual cases are
studied to shed light on other groups/
Threats to validity: Avoided by means of
statistical methods.
Threats to validity: Avoided by means
of logical analysis to rule out alterative
Subject of analysis: These are simplified
and reduced as much as possible.
Subject of analysis: These are studied
as a whole, as they occur in reality as a
complex system.
Conclusions: These are stated with
statistical measures of confidence (e.g.,
alpha levels).
Conclusions: These are suggestive and
always expressed tentatively.
From (Clarke, 2003)
The differences between quantitative and qualitative research can also be expressed
as follows.
Quantitative (deductive/linear)
Quantitative research typically follows a linear path: one starts with a testable
hypothesis, collects data, analyses the data and then accepts or rejects the hypothesis.
Hypothesis Data Collection Analysis of results
Accept/reject hypothesis
‘The more interest rates rise …’
Qualitative (inductive/spiral)
Qualitative research has a very different structure.In the case of qualitative research,
the researcher starts with a tentative idea or question, e.g., ‘what is it like working for a
major software company in the current economic climate?’ One then observes and
asks questions, and analyses what one finds. This guides more specific questioning.
Further investigation reveals themes and patterns in the research, which lead to
appropriate theoretical study. Finally, one ends with tentative conclusions based on the
theoretical insights that one has acquired during the research.
In quantitative research, the researcher generally starts with a theoretical statement or
position (the hypothesis) and tests this for accuracy. This is therefore deductive research
(moving from theory).
Page 2
In qualitative research, one starts with observation and ends with a theoretical position
or stance. This is, therefore, inductive research (moving to theory).
Of course, both kinds of research are not exclusive. Qualitative research may end in a
hypothesis that can be quantitatively tested later. Quantitative research may involve
qualitative research elements.
When is qualitative research needed?
As quantitative research is generally well-known, it may be useful to outline when
qualitative research is needed.
Qualitative research is often appropriate in the following situations:
When the research is looking at an area that is not well-studied or understood
When a subject needs to be studied in a great deal of depth
When a holistic perspective is needed
When attitudes or behaviours of people need to be studied
When measurement techniques like questionnaires are not considered suitable
When you are more interested in the process of something (how it works) and not
the product (the outcome)
When you want to put “content” on statistical results to make the results meaningful
When observation of people is considered to be important
From (Hurworth, 2003)
Case study method uses both qualitative and quantitative research methods. See:
helpsheets, Case Studies 1, Case Studies 2 and Case Studies: Research Method
Page 3
Qualitative and quantitative data
There are several types of data that can be researched. Some of these are listed below:
Qualitative Data
Utterances: Not just words but complete
utterances and the meaning they have
in context (e.g., phrases used between
workers on the shop floor).
Physical actions: For example, what
workers do with parts of their bodies
when in conversation with fellow workers or their employers; how prospective
employees behave in interviews, etc.
Quantitative Data
Nominal categories: Nominal categories
include things like “worker”, “boss”, “employed”, and “unemployed”.
Ordinal categories: Ordinal categories
are ranked categories such as “strongly
agree”, “agree”, “undecided”, “disagree”, etc.
Text analysis: This refers, for example, to
Interval categories: This refers to quantifi-
the reoccurrence of words, their mean-
able differences between categories,
ing and significance in texts.
for example, IQ.
Pictures and diagrams: Diagrams and
Ratio: These categories can be related
pictures can be studied as qualitative
as proportions (e.g., income level).
(Clarke, 2003)
Constructs and hypotheses
Whether quantitative or qualitative, research must always carried out with a research
question or statement (or hypothesis) in mind. It is important to also be sure that the
constructs being studied are operational and can be researched successfully.
What is a construct?
A construct is an abstract entity, not something physical. For example, “employee
satisfaction” or “customer satisfaction” are constructs, as is “short term pressure on
interest rates”. You must be clear how the constructs in your proposed study can be
made operational (objectively testable). You need to ask yourself questions such as
• ‘How do I decide what “customer satisfaction” means?’
• ‘How can it be measured?’
• ‘What counts as an example of “customer satisfaction”? … and so on.’
Similarly, constructs like: “intelligence”, “company performance”, and “interest”, need
to be made operational before you can commence your research project.
Page 4
What is a hypothesis?
A hypothesis is a statement that relates two or more constructs. For example: The
greater the stress experienced in the job, the lower the job satisfaction of employees.
Here you need a clear operational definition of “stress” and “job satisfaction”.
A good hypothesis is tested by the research that you propose to do. Note that assumptions are not the same as hypotheses. Assumptions are not tested in the research, but
hypotheses are. You can also commence and carry out research by taking assumptions
to be true, but you cannot do this with the hypothesis(es). In the example above, you
might assume the truth of the proposition that stress results in employee dissatisfaction
and begin your research. However, you cannot assume the truth of the hypothesis that
the greater the stress the lower the job satisfaction. This is, in fact, what you will test.
Assumptions in the use of constructs
The following assumptions are implicit in the selection of constructs. One must be wary
of them and consider them carefully when designing projects (from Clarke, 2003).
All researchers select what they want to study and leave out other things that they
consider irrelevant. However, this selection process is subjective and another researcher
may regard something to be relevant that you regard as irrelevant. How would you
justify your selection decisions?
Anything that is studied can also be distorted or changed in the process of carrying out
a study. A survey, for example, can prompt attitudes that participants initially did not
have. Interviews can result in answers that may not have been given outside the interview context. Observers in a company can change the way workers in the company
interact. How would you minimise such distortions in your study?
All research has a subjective component. How a survey or interview is carried out inevitably reflects the researcher’s views and biases. How would this be minimised?
Reducing or simplifying a construct can also misrepresent it. For example, you may want
to operationalise “customer satisfaction” to be only the number of times a customer
returns to the company for further products. But this might result in an inaccurate assessment of satisfaction. Customers might return for other reasons (for example, the company is the nearest one, etc). How would you avoid problems of reductionism in your
When something being studied is removed from its usual surrounding context, it can be
different or behave differently. Studying something in a laboratory or outside the context of the normal company operations may result in an artificial or inaccurate view of
what is really going on. How can you minimise problems of decontextualisation?
Page 5
Sometimes the researcher can view something as important when it does not really
exist. Constructs can be made “real”, when they are merely the result of the individual
researcher being too closely involved in the proposed study and seeing something as
“observed”, when it is really only what the researcher wants to see. How would this be
avoided in your study?
There is sometimes a danger of not seeing the observations being made in a wider
context of the research activity itself. The researcher is also part of the wider sphere of
the experiment, and the research situation is also part of the situation being studied. The
aims of the research are also part of the observed aims, and so on. Not being conscious
of the situatedness of the research activity can blind the researcher into thinking their
work is totally objective and disconnected to what one is studying. How would you be
sure that your research does not fall into this trap?
Your participation in the research process
It is important to recognise the extent to which the researcher is a participant or a nonparticipant in research. Note the typical three types of observer listed below and associated advantages and disadvantages:
1. The uninformed non-participant
This is an observer who is not taking part in the particular observed practices being studied and who has not participated in the general practices or setting before.
Advantages: Less chance of observer effect and observer bias and a fresh, impartial
view on data being collected. However, advantages can be outweighed by the disadvantages: they may not be aware of important information being collected (or they
may mistake what is really significant).
2. The informed non-participant
This is an observer who has participated before but who is not taking part in the particular practices being studied.
Similar advantages to the uninformed non-participant. The disadvantage is that informed non-participants may be biased in the evidence they collect.
3. The informed participant
This is an observer who has participated before and is taking part in the practices being
Advantages: they can empathise with participants being studied (also a disadvantage). The biggest disadvantages of the informed participant are the potential for
observer effect (presence may change data being collected) and observer bias
(characteristics of observer may influence what is observed). Finally, those being studied
may not behave normally while being observed, especially if they know the observer is
among them.
Page 6
Validity and reliability
In the case of quantitative research in particular, it is important for the indicator used to
obtain data in your research to be both valid and reliable.
Validity refers to the “extent to which the design of the study, the instrument used to
collect the data (e.g., questionnaire) and the results investigate the research questions
they were intended to investigate”. (Clarke, 2003).
Reliability refers to “the extent to which an investigation produces consistent results”
(Clarke, 2003). The following table summarises the various kinds of validity and reliability
issues that need to be considered. (Clarke, 2003; Neuman, 1994).
Face validity: The indicator (survey,
questionnaire) really does measure the
construct under examination.
Stability reliability: This is reliability over
time (test and retest).
Content validity: The indicator measures
all aspects of the construct and not just
a part.
Representative reliability: This is reliability over sub-populations (men and
Criterion validity: The indicator corresponds with and is predictive of measurements using related indicators.
Equivalence reliability: This is reliability
across different indicators used (e.g.,
questionnaires and inter-rater reliability
Construct validity: The indicator measures the construct in a manner which is
convergent with other measures in terms
of direction (e.g., the level of education and income level converge). The
indicator also allows discrimination of
opposing constructs.
Internal validity: The indicator and research design chosen should not allow
different interpretations of the data.
Internal validity should be “high”.
External validity: Results gained from the
indicator and research design should
be generalisable to other cases that
are not being studied. External validity
should be “high”
Statistical validity: The indicator being
used should allow data to be measured
using the most suitable statistical instrument.
Consequential validity: Results from the
study should be appropriate and relevant to the research context.
Page 7
The Hypothetico-deductive method
This research method should not be seen as strange and mysterious. It is a natural combination of the emphasis of both qualitative and quantitative research methods.
It may turn out that the research you are required to produce (in terms of written work)
in your department is strictly quantitative or qualitative. However, in practice, your research will generally progress through the stages given above.
Page 8
5 minute self test
1. Qualitative
2. Hypothesis
5. Validity
1. Quantitative
3. Reliability
4. Stress
6. Data
Chalmers, A. (1982). What is This Thing Called Science? St. Lucia: University of
Queensland Press.
Clarke, D. (2003). Research Methods in Education. Unpublished manuscript, Melbourne.
Hurworth, R. (2003). Overview of Qualitative Methods. Unpublished manuscript,
Neuman, W. L. (1994). Social Research Methods (3rd ed.). Boston: Allyn and Bacon.
Perry, C. (1994). A Structured Approach to Presenting PhD Theses: Notes for Candidates
and their Supervisors. Paper presented at the ANZ Doctoral Consortium, Sydney
Sekaran, U. (1992). Research Methods for Business: A Skills-Building Approach. New York:
Wiley and Sons.
Yin, R. (1994). Case Study Research: Design and Methods (Vol. 5). London: Sage.
Page 9