RESEARCH DESIGN | CHAPTER 1 | NO-REPEAT EDITION

How to Choose a Research Approach

A simple, complete guide to qualitative, quantitative, and mixed methods

Based on Creswell & Creswell, Research Design (5th edition)

Promise: Every major concept is explained once, then applied. No separate section repeats the same
definition in different words.

UNDERSTAND RECOGNISE APPLY

Which worldview or design a question How to build and justify your own
Why the three approaches differ
needs study

Language: simple English + short Hinglish explanations

RESEARCH DESIGN STUDY NOTES CHAPTER 1

1. The entire chapter in one story

Running question: Why do many educated young people remain unemployed even after earning a
degree?

Before collecting data, a researcher must make three linked decisions. Together they form the research
approach.

Decision What it means Unemployment example

Is there one measurable pattern, many lived meanings,

Your assumptions about reality and what
1. Worldview an injustice to change, or simply a practical problem to
counts as trustworthy knowledge.
solve?

2. Research The named blueprint that organises the Survey, experiment, phenomenology, case study,
design study. explanatory sequential design, etc.

3. Research The actual procedures used to obtain and Sampling, questionnaire or interview, statistical or
methods analyse evidence. thematic analysis, interpretation, and quality checks.

Hinglish: Worldview = tum kis nazariye se knowledge ko dekh rahe ho; design = poora blueprint; methods
= ground par exactly kya karoge.

One important correction

Qualitative and quantitative research are ends of a continuum, not airtight boxes. A study may lean
strongly toward one side and still borrow a small feature from the other. Mixed methods is different: it
deliberately uses both forms of evidence and integrates them to answer one connected problem.

Core rule: Start from the research problem. Do not begin with a favourite method and force the problem
to fit it.

Page 2
RESEARCH DESIGN STUDY NOTES CHAPTER 1

2. First decision: what kind of evidence is needed?

Approach Best used when How the reasoning works What the study finally gives

Meanings, experiences, context, Open questions -> detailed accounts A deep explanation of how
Qualitative or a process are not yet -> categories/themes -> people experience or
understood. interpretation. understand the issue.

An estimate, comparison,
Variables are known and a theory, Theory -> variables/hypothesis ->
relationship, prediction, or
Quantitative difference, relationship, or effect measurement -> statistics ->
causal effect when the design
can be tested. conclusion.
supports it.

A combined explanation in
One data type alone leaves an Collect both forms -> connect or
which breadth and depth
Mixed methods important gap that the other can merge them at a planned
answer different parts of the
fill. integration point.
same problem.

Qualitative. The researcher usually works in the participants' natural setting, asks broad questions,
gathers words, observations, documents, or images, and builds meaning from the details. The design
can evolve because early findings may reveal better questions. The researcher is part of interpretation,
so reflexivity and transparent reasoning matter.
Example question: How do unemployed graduates explain the gap between their degree and the skills
employers demand?

Quantitative. The researcher defines variables in advance, measures them consistently, and uses
statistics to test expectations. Larger or probability-based samples may support population
generalisation. Causal claims need control over alternative explanations; numbers alone do not
automatically prove cause.
Example question: Does completing a certified internship increase a graduate's probability of employment
within six months?

Mixed methods. The researcher has a specific reason for needing both numeric and qualitative
evidence. A study is not mixed merely because it contains a survey and interviews; the two parts must
meet in sampling, sequence, analysis, or interpretation.
Example: First estimate unemployment patterns in a survey, then interview selected graduates to explain the
surprising patterns.

Page 3
RESEARCH DESIGN STUDY NOTES CHAPTER 1

3. Worldviews: four different lenses on the same problem

A worldview is the researcher's broad set of assumptions about reality, knowledge, values, and the
proper way to investigate. It silently shapes the questions considered important and the evidence
considered convincing.

Postpositivist lens - test and revise. A reality exists beyond us, but every measurement is imperfect.
Researchers reduce bias, test theories, examine causes, and accept conclusions as provisional rather
than final truth. This lens commonly supports quantitative studies.
Question: After controlling for degree, region, and family background, how strongly does internship experience
predict employment?

Constructivist lens - understand meaning. People form multiple meanings through interaction, history,
and culture. The researcher listens to participants, uses broad questions, and interprets how their
different realities are constructed. This lens commonly supports qualitative studies.
Question: What does 'being employable' mean to graduates from different social and educational backgrounds?

Transformative lens - expose power and support change. Knowledge is connected to politics and
inequality. Research should include marginalised groups, question power relations, and aim at reform or
action rather than merely describing them. Participation and justice are central.
Question: How do caste, gender, disability, and institutional discrimination restrict access to graduate jobs, and
what reforms do affected groups demand?

Pragmatic lens - use what solves the problem. The research question and useful consequences guide
method choice. Researchers may combine different assumptions and data forms if doing so produces a
workable answer. Pragmatism is therefore common in mixed methods.
Question: Which combination of labour-market data and graduate experiences will best guide an effective
placement programme?

Do not memorise rigid pairs: Worldviews and approaches are common associations, not prison cells. The
real test is whether the worldview, question, design, and methods form a coherent argument.

Page 4
RESEARCH DESIGN STUDY NOTES CHAPTER 1

4. Quantitative designs: measure, compare, test

Quantitative design is chosen after the researcher knows which variables must be measured. The design
determines what kind of claim the evidence can support.

Survey design. Collect standardised responses from a sample to describe a population's attitudes,
characteristics, or behaviour. A cross-sectional survey gives a one-time snapshot; a longitudinal survey
follows change over time. Good sampling matters because a precise estimate from a biased sample is
still biased.
Use it for: What proportion of final-year students have completed internships, and how does this vary by
institution type?

Correlational or observational design. Measure variables as they naturally occur and estimate whether
they move together. This can describe association or prediction, but it cannot by itself rule out reverse
causality or omitted variables.
Use it for: Is internship completion associated with higher starting wages after accounting for degree and
location?

True experiment. Manipulate a treatment, randomly assign eligible participants to treatment and
control groups, and compare outcomes. Random assignment helps balance other causes, making a
causal claim more credible.
Use it for: Randomly offer an intensive placement programme and compare employment rates with a control
group.

Quasi-experiment. Estimate an intervention's effect without full random assignment, using a

comparison group, policy cutoff, timing change, matching, or another identification strategy. It can be
powerful, but the researcher must defend why the groups are comparable.
Claim discipline: Survey or correlation -> association. Strong experiment or defensible quasi-experiment
-> possible causation. Never upgrade a weak design by using stronger language.

Page 5
RESEARCH DESIGN STUDY NOTES CHAPTER 1

5. Qualitative designs: choose the kind of understanding

Qualitative designs are not five names for the same interview study. Each seeks a different final product.

Narrative research - a person's story. Study the life or experience of one or a few individuals and
organise events into a meaningful chronology. The final product is a story showing turning points,
context, and interpretation.
Phenomenology - the shared essence of an experience. Study several people who have lived the same
phenomenon and identify what they commonly experienced and how they experienced it. The final
product is the essential structure of that lived experience.
Grounded theory - a theory built from participant data. Study a process, action, or interaction; collect
and compare data repeatedly; and develop an explanatory model grounded in the evidence. The goal is
theory, not merely a list of themes.
Ethnography - the culture of a group. Spend sustained time with a culture-sharing group and study its
shared behaviour, language, routines, and beliefs in context. The unit is the group's cultural pattern, not
just individual opinion.
Case study - a bounded case examined deeply. Investigate a clearly bounded programme, institution,
event, community, or person using multiple sources such as interviews, documents, observation, and
records. The boundary tells the reader what is inside and outside the case.
Memory chain: Story = narrative | shared experience = phenomenology | process theory = grounded
theory | culture-sharing group = ethnography | bounded unit = case study.

Page 6
RESEARCH DESIGN STUDY NOTES CHAPTER 1

6. Mixed methods designs: the two datasets must meet

Mixed methods is justified when qualitative and quantitative evidence answer different but connected
parts of the problem. The researcher must state the sequence, relative priority, and integration point.

Convergent design - collect both together. Gather qualitative and quantitative data during roughly the
same phase, analyse them separately, and then merge the results. Agreement strengthens confidence;
disagreement becomes a finding that needs explanation.
A graduate survey and employer interviews are conducted together, then compared to see whether both
identify the same skill gaps.

Explanatory sequential - numbers first, explanation second. Begin with quantitative results, select the
findings or cases that need clarification, and follow with qualitative work. The second phase explains
why the initial pattern occurred.
A survey finds rural graduates with similar marks have lower placement rates; interviews then investigate the
reasons.

Exploratory sequential - exploration first, measurement second. Begin qualitatively when concepts are
poorly understood, then use the findings to create an instrument, variables, intervention, or later
quantitative test.
Interviews identify locally relevant employability barriers; those themes are converted into a questionnaire for a
large sample.

Embedded or complex design - one dataset supports a larger design. Place a smaller qualitative or
quantitative component inside an experiment, case study, evaluation, or multi-stage programme. Its
role and priority must be explicit.
Integration test: If removing one dataset would not change how the other is sampled, explained,
compared, or interpreted, the study may be parallel data collection rather than genuine mixed methods.

Page 7
RESEARCH DESIGN STUDY NOTES CHAPTER 1

7. Methods: turn the design into an actual study

Methods are the specific procedures used at each stage. A method is good only when it fits the question
and the claim the researcher wants to make.

1. Questions. Quantitative questions specify variables or hypotheses; qualitative questions remain

open enough to discover meaning; mixed questions connect both strands.
2. Sampling. Probability sampling supports population estimates; purposeful sampling selects
information-rich participants; mixed studies coordinate samples across phases.
3. Data collection. Closed instruments create standardised numeric data; open interviews,
observation, documents, and images preserve context; some studies combine them deliberately.
4. Analysis. Quantitative analysis estimates patterns and uncertainty; qualitative analysis codes
details into categories and themes; mixed analysis compares, connects, or merges the two results.
5. Interpretation. The researcher answers the original question, considers rival explanations, states
limitations, and keeps claims within the design's actual power.
6. Quality checks. Quantitative studies examine reliability, validity, bias, and generalisability.
Qualitative studies examine credibility, reflexivity, multiple evidence sources, and
participant/context detail. Mixed studies also test the quality of integration.
Subtle point: A tool does not define an approach by itself. An interview can contain closed numeric
ratings, and a survey can contain open responses. Look at the full logic: question, sampling, data, analysis,
and claim.

Researcher role changes across approaches

• Quantitative: standardise measurement and minimise avoidable influence.
• Qualitative: interact with participants and explain how your background may shape
interpretation.
• Mixed methods: manage both roles and make the connection between strands transparent.

Page 8
RESEARCH DESIGN STUDY NOTES CHAPTER 1

8. How to select the right approach

Start here: Write the exact research problem and question before naming an approach.

A practical decision route

1. Is the main need to test known variables, estimate a population pattern, compare groups, or
identify an effect? Choose a quantitative design that supports the intended claim.
2. Is the central issue poorly understood, context-heavy, or about meaning, experience, culture, or
process? Choose a qualitative design whose final product matches that need.
3. Will one form of evidence leave a clearly stated gap that another form can fill? Use mixed
methods and name the integration point. If not, do not add a second method for decoration.

Then check feasibility

• Experience and comfort: statistics, fieldwork, interviewing, interpretation, or team support.
• Resources: time, funding, software, travel, transcription, sample access, and research assistants.
• Ethics and access: vulnerability, privacy, gatekeepers, intervention risk, and data permissions.
• Audience: discipline, supervisor, journal, policy client, or community expectations.
Hard truth: Mixed methods is not automatically superior. It demands competence in two traditions plus
integration. Use the least complicated rigorous approach that fully answers the question.

Final alignment check

Read the proposal as one chain: problem -> question -> worldview -> design -> sample/data -> analysis
-> claim. If one link promises something the next link cannot deliver, the study is misaligned.

Page 9
RESEARCH DESIGN STUDY NOTES CHAPTER 1

9. One economics topic designed three genuinely different ways

Topic alone never determines the approach. The question determines what evidence is needed.

Version A: qualitative
• Question: How do first-generation graduates experience the transition from university to
unemployment?
• Logic: constructivist; phenomenology; purposeful sample of graduates who share the experience.
• Evidence: repeated open interviews and reflective notes; thematic analysis seeks the shared
essence and important differences.
• Output: a deep account of uncertainty, family pressure, identity, skill mismatch, and job-search
experience.

Version B: quantitative
• Question: Does access to an intensive placement programme increase employment within six
months?
• Logic: postpositivist; randomised experiment if feasible, otherwise a defensible quasi-experiment.
• Evidence: baseline characteristics, treatment assignment or comparison rule, employment
outcomes, and statistical effect estimates.
• Output: the estimated average effect, uncertainty, subgroup differences, and limits to
generalisation.

Version C: mixed methods

• Question: Which factors predict unemployment, and why do graduates with similar qualifications
experience different outcomes?
• Logic: pragmatic; explanatory sequential design.
• Evidence: a large survey identifies patterns; selected follow-up interviews explain surprising or
contrasting cases.
• Integration: quantitative results choose interview participants and topics; both strands are joined
in the final explanation.
What changed?: Not the topic. The question, desired claim, and evidence changed - therefore the
worldview, design, and methods changed.

Page 10
RESEARCH DESIGN STUDY NOTES CHAPTER 1

10. Common confusions - corrected once

Mistake Correct understanding

Qualitative = words; quantitative = Too shallow. The approaches also differ in question style, theory use, design
numbers flexibility, researcher role, analysis, and kind of claim.

Not necessarily. Closed interview schedules can produce quantitative data; open
Interviews make a study qualitative
interpretive interviewing is qualitative.

Sample size affects precision, not causal identification. Causation depends on

A large sample proves causation
design and control of rival explanations.

Mixed methods means using two It requires a reason for mixing and a planned connection or merger between the
tools datasets.

Worldview is a personal opinion It is a defensible set of assumptions that must be visible in the study's questions,
paragraph design, evidence, and claims.

The most complex design is the Strength means fit and rigour. Unnecessary complexity creates more failure points
strongest without improving the answer.

Exam-ready structures
If asked: Define a research approach. Give the definition, name its three components, and explain that
the components must align from assumptions to procedures.
If asked: Compare the three approaches. Compare purpose, question type, role of theory, data,
sampling, analysis, researcher role, and final claim. Do not stop at 'words versus numbers.'
If asked: How should an approach be selected?. Lead with the research problem; then discuss the
researcher's experience and resources, ethical/access constraints, audience expectations, and overall
alignment. End with a short example.
10-mark flow: Definition -> conceptual framework -> relevant comparison -> selection criteria -> one
applied example -> alignment-based conclusion.

Page 11
RESEARCH DESIGN STUDY NOTES CHAPTER 1

11. Application self-test

Answer without looking back. The point is to choose and justify, not merely recall a label.

1. A national sample is used to estimate the percentage of graduates unemployed one year after
graduation. Which approach and design fit best?
2. Researchers want the shared lived experience of repeated job rejection. Which qualitative design
fits?
3. An intervention is randomly offered to half of eligible students. What design is this, and what kind
of claim may it support?
4. A survey reveals an unexpected gender gap; interviews then investigate why it occurred. Name
the mixed design.
5. Interviews are used to create a culturally valid employability scale, which is then tested on 2,000
students. Name the mixed design.
6. A researcher spends a year studying the language, routines, and norms of a placement-cell team.
Which qualitative design fits?
7. A study includes both a survey and interviews but reports them in separate chapters with no
connection. Is it genuine mixed methods? Why?
8. What three practical factors, apart from the problem itself, should influence approach selection?

Answer key with reasons

Q Answer and reason

1 Quantitative survey; it estimates a population pattern from a sample.

2 Phenomenology; it seeks the common essence of a shared lived experience.

True experiment; random assignment can support a causal effect claim if implementation and measurement are
3
sound.

4 Explanatory sequential; quantitative results come first and qualitative evidence explains them.

5 Exploratory sequential; qualitative exploration builds the later quantitative instrument.

6 Ethnography; the focus is a culture-sharing group's patterns in context.

7 No. Both data types were collected, but purposeful integration is missing.

8 Researcher experience/resources, ethical and access feasibility, and the intended audience or disciplinary norms.

Final takeaway: A strong research design is not a collection of impressive labels. It is one coherent chain
in which every choice helps answer the same question.

Page 12