Rancangan

Penelitian
(Research Design)
Sri Raharjo
Direktorat Penelitian
5 Agustus 2016
 Tujuan
• Memahami beragam pendekatan dalam
merancang penelitian
• Dapat memilih metodologi yang sesuai dalam
merencanakan penelitian
Industri manufaktur, Pertanian, Peternakan...
Arkeologi, sosial ekonomi, komputer, satelit...
Source: © Mark Saunders, Philip Lewis and Adrian Thornhill 2006
 Research Philosophy
• This layer contains the philosophical stances
associated with the philosophies.
• Each of the possible choices at this level
require careful thought as they provide
structure, guidance and possible limitations to
following decisions and ultimately the way a
researcher can collect and analyse data to
create valid findings.
 Research Philosophy:
Positivism
• Positivism generates hypotheses (or research
questions) that can be tested and allows
explanations that are measured against accepted
knowledge of the world we live in.
• This position creates a body of research that can
be replicated by other researchers to generate
the same results.
• The emphasis is on quantifiable results that lend
themselves to statistical analysis.
• These could be the laws of gravity applied to an
apple falling from a tree.
 Research Philosophy:
Realism
• Realism is similar to positivism in its processes and belief
that social reality and the researcher are independent of
each other and so will not create biased results.
• However, where they differ is that realism thinks that
scientific methods are not perfect.
• It believes that all theory can be revised and that our
ability to know for certain what reality is may not exist
without continually researching and leaving our minds
open to using new methods of research.
• Realism may therefore use several types of research
methods to triangulate results in their search for a more
reliable outcome.
 Research Philosophy:
Interpretativism
• Interpretativism refers to approaches
emphasizing the meaningful nature of people's
participation in social and cultural life.
• Researchers working within this tradition analyse
the meanings people confer upon their own and
others' actions
• Take the view that cultural existence and change
can be understood by studying what people think
about, their ideas, and the meanings that are
important to them.
 Research Philosophy:
Pragmatism
• Pragmatism argues that both constructivism and
objectivism are valid ways to approach research.
• Pragmatism allows a researcher to view the topic
from either or both points-of-view regarding the
influence or role of social actors and uses these
to create a practical approach to research.
• This may be used to find solutions to problems.
 Research Philosophy:
Objectivism
• Objectivism recognises that social phenomena and
their meanings exist separately to social actors.
• An example of social phenomena could be heavy rain
and social actors are people wanting to have a picnic
outside.
• The rain exists, it is real and would be acknowledged by
everyone to exist and is therefore independent to the
people who have had their day out ruined by it.
• In your research this may be how a law (the social
phenomenon) impacts on a group of people (social
actors).
 Research Philosophy:
Constructivism
• Constructivism argues the opposite to
objectivism.
• It is a standpoint that believes social phenomena
are actually constructed by social actors.
• So, if you had a constructive ontological
worldview you would believe that, for example, a
new law is the product of the behaviour of the
group of people it now has an impact on
 Research Approach
• It contains the terms deductive and inductive.
• A decision on this level may be strongly
indicated by the decisions made at the
previous level, or you may need to assess your
research aim, your limitations and personal
opinions to decide which method will work
best for your work.
 Research Approach: Deductive
• Deductive means that you start with a statement or
question and your research sets out to answer it.
• The aim would be to conclude with a yes or no response
to the question.
• Questions may be statements or informed speculation
about the topic that the researcher believes can be
answered.
• The thought process of deduction moves from theory to
the research question, to data collection, findings to a
rejection or confirmation of the research question.
• This should lead onto a revision of the theory and often
starts the process over again.
 Research Approach: Inductive
• Inductive means that you are researching to create
theory.
• The process moves in the opposite direction to the
deductive approach taking its focus from the working
title of the researcher not the existing theory.
• This means the research goes from research question
to observation and description to analysis and finally
theory.
• Therefore if little research exists on a topic then an
inductive approach may be the best way to proceed.
Deductive vs. Inductive Approaches

Inductive

Deductive
 Research Strategy Choices
• Choices refer to the research style that you will use to
collect and analyse data.
• Each one has its benefits and limitations. This needs to
be thought through and explained and balanced
throughout your work.
• Choices may be more commonly associated with
different philosophies and philosophical standpoints.
• You can choose more than one of these choices to
design and collect data as long as this is a decision that
can be justified in your written work.
 Research Strategy Choices:
Experiment
• Experimental designs are more rigid and scientific
in their structure to enable the research to be
replicated.
• These designs test the causal effects of
phenomena on a group compared to a control
group who are not subjected to any phenomena.
• The causal effect is the independent variable on
the dependent variable.
• Experimental strategies generate data that can be
statistically analysed.
 Types of Experiments
Laboratory experiments
• are those in which the independent variable is
manipulated and measures of the dependent
variable are taken in a contrived, artificial setting for
the purpose of controlling the many possible
extraneous variables that may affect the dependent
variable
 Types of Experiments

Field experiments
• are those in which the independent variables are
manipulated and the measurements of the
dependent variable are made on test units in their
natural setting.
 Experimental vs. Quasi-Experimental
1. Experimental
 only type of study design that can actually prove
causation.
 individuals are randomly allocated to at least two
groups.
 Manipulation is there.
 Randomization.

2. Quasi-Experimental
• all the three main characteristics are not followed.
• Manipulation must be present.
Quasi-Experimental Control Group

Study group Intervention Study group

before after

Compare

control group Control group

before before
 Research Strategy Choices:
Survey
• A survey strategy is often associated with a
deductive approach.
• It offers the researcher a highly economical
way of collecting large amounts of data to
address the who, what, where, when and how
of any given topic or issue.
• This strategy can generate both rich and
statistical data.
 Research Strategy Choices:
Case Study
• Case study design involves extensive study of one
or more individuals or cases in a real life context.
• To draw clear conclusions about data it is advised
that the number of cases are restricted.
• The data that is collected may include: watching
aspects of their behaviour or of the setting,
interviews with participants and record
searching.
 Research Strategy Choices:
Aciton Research
• This strategy is concerned with addressing issues to
find and implement solutions.
• As part of this search for a solution the strategy allows
the researcher to be part of the organisation or case
study that requires the solution.
• It allows for collaboration between the topic
organisation and the researcher.
• The process of Action Research moves from a clear
objective to diagnosis of the problem and generation
of a list of actions to solve the problem.
 Research Strategy Choices:
Grounded Theory
• Grounded Theory uses inductive methods to predict
and explain behaviour to build theory.
• This starts with data being collected from
observation, theory and predictions being generated
from that data and then those predictions being
tested.
• This strategy, although generating new theory, is still
grounded by existing theory and literature on the
topic.
 Research Strategy Choices:
Ethnography
• Ethnography is rooted in anthropology, which is
the study of others from a detached point-of-
view.
• However, ethnography requires the researcher to
be a part of the community or situation they are
researching.
• This may be time consuming for the researcher to
achieve full integration into a social scene to
experience and document long-term changes in
actions and opinions.
 Research Strategy Choices:
Archival Research
• This strategy centres its data collection on
existing data sets or archive documents.
• This allows for exploratory, explanatory or
descriptive analysis of changes tracked over a
long period of time.
• However, the accuracy and breadth of
information available may be an issue for a
researcher relying solely on this type of
secondary data.
 Research Method Choices:
Mono-method
• Mono-method research is when either quantitative
or qualitative data is collected rather than a
combination of both.
• This may be due to the demands of the philosophy,
philosophical choices and strategies employed.
• It could also be used to research an opposing view to
existing mono-method research
 Research Method Choices:
Mixed-method
• Mixed-methods research is when the researcher uses
quantitative and qualitative research methods in the
process of their study, data collection and analysis.
• It can be argued that by combining both types of
research, the limitations of each individual method
can be offset and gaps of data can be filled or
predicted.
 Research Method Choices:
Multi-method
• This is where the researcher uses both quantitative
and qualitative data but the researcher’s outlook is
rooted in only one of them.
• So both types of data are analysed from only one
point-of-view.
 Research Time Horizon
Cross-sectional
• Cross-sectional designs can use qualitative and
quantitative research and they measure an
aspect or behaviour of many groups or
individuals and at a single point in time.
Longitudinal
• Longitudinal designs can also use qualitative and
quantitative research but they study events and
behaviours using concentrated samples over a
longer period.
 Research Techniques and Procedures
• This is where you as the researcher need to decide from all
the previous decisions you have made what data collection
methods will work best and what type of analysis you
employ to create the results to answer your research
question.

Data Collection & Data Analysis

• This includes decisions on sample groups, questionnaire
content, and questions to be asked in interviews and so on.
• All of the decisions and tools employed at this final stage
must fit in with the philosophies, philosophical stances,
strategies, choices and time-horizons already fixed upon if
valid results are to be created and withstand criticism.
 Research Method Choices
• This is where you are asked to define how you wish to
use quantitative and qualitative methods in your
research.
• This is the stage where you decide whether you use
one or both methods and if you will use them equally
or with one more dominant than the other.
• Quantitative research involves numbers. It is concerned
with quantity and measurements.
• Qualitative research is concerned with rich data such as
personal accounts, opinions and description.
 The Philosophical Position

Positivism Phenomenology
Reality is objective and Reality is subjective
Ontology: what is the singular, apart from the and multiple as seen
nature of reality? researcher by the participants
Epistemology:
Researcher is independent Researcher interacts
What is valid
from that being researched with that being researched
knowledge?
Axiology:
Value free and un-biased Value-laden and biased
Role of values
• Cross-sectional studies • Action Research
RESEARCH • Experimental studies • Case Studies
STRATEGY • Longitudinal studies • Ethnography
• Surveys • Grounded Theory
• Etc... • Hermeneutics, etc...
 Positivistic paradigm Phenomenological paradigm
Tends to produce quantitative Tends to produce qualitative
data data
Uses large samples Uses small samples
Concerned with hypothesis Concerned with generating
testing theories
Data is highly specific and Data is rich and subjective
precise
The location is artificial The location is natural
Reliability is high Reliability is low
Validity is low Validity is high
Generalises from sample to Generalises from one setting to
population another
 Comparing approaches

Characteristic Positivism Phenomenology

Questions that can be What? Why?
answered How much? How?
Direct observation,
Survey,
Associated methods Interviews,
Experiment
Participant observation
Predominantly
Data type Predominantly words
numbers
Finding Measure Meaning
 Why does the approach matter?
Whether you take a scientific (positivistic) or
phenomenological approach will influence:
• What research questions you ask
• What methods you use to collect your data
• What type of data you collect
• What techniques you use to analyse your data
 Rancangan Penelitian
(Research Design)

A series of choices specific to the methodology selected:

 Will the research be inductive or deductive?
 Interpretive or functional?
 Controlled or naturalistic?
 Conducted in the field or in an artificial setting?
 Will you be a participant in the social setting or an
observer?
 Will it involve qualitative or quantitative data?
 Will it be cross-sectional or longitudinal?
 Rancangan Percobaan
(Design of Experiment)
1. Formulate question/goal in advance
2. Comparison/control
3. Replication
4. Randomization
5. Stratification (blocking)
6. Factorial experiments
 Rancangan Analisis Data
(Design of Data Analysis)
 Know in advance how you will evaluate each
piece of data you gather to answer the research
question or test the research hypothesis.
 Will analysis be qualitative or quantitative?
 What measures will be compared?
 Will you look for differences between groups or
relationships between variables?
 Contoh Analisis Data
1. Comparison
2. Screening / characterization
3. Modeling
4. Optimization
 Classification of Study Design

Cross-sectional
1. Descriptive
Longitudinal

Case-control
2. Analytical
Cohort

Experimental
3. Interventional
Quasi-experimental
 Diet habit Cases

Contains + bowl
of the food _
cancer
cases

Population
Case control

+
control
_
 High fat diet +

Cohort

Population

Not high fat

+
diet

_
Cohort Bowl cancer
THANK YOU