10  Steps of Research

10.1 The Nine-Step Research Process

The research process is a logically sequenced but iterative workflow. The classical sequence has nine steps; failure to complete any step properly weakens every subsequent step. Most-repeated PYQ patterns: (a) sequence ordering — arrange the given steps in correct order, (b) first step / last step identification, and (c) matching activities (literature review, hypothesis, sampling) to their step.

TipNine Steps of the Research Process
  1. Identification of the research problem
  2. Review of related literature
  3. Formulation of objectives and hypotheses
  4. Research design
  5. Sampling — defining population, selecting sample
  6. Data collection
  7. Data analysis
  8. Interpretation and conclusion
  9. Reporting and publication

flowchart TB
  S1[1 Problem<br/>Identification] --> S2[2 Literature<br/>Review]
  S2 --> S3[3 Objectives &<br/>Hypotheses]
  S3 --> S4[4 Research<br/>Design]
  S4 --> S5[5 Sampling]
  S5 --> S6[6 Data<br/>Collection]
  S6 --> S7[7 Data<br/>Analysis]
  S7 --> S8[8 Interpretation<br/>& Conclusion]
  S8 --> S9[9 Reporting &<br/>Publication]
  S9 -. feedback .-> S1
    classDef default fill:#003366,color:#ffffff,stroke:#ffcc00,stroke-width:3px,rx:10px,ry:10px;

10.2 Step 1 — Identification of the Research Problem

10.2.1 What a Research Problem Is

A research problem is a clearly-stated, unresolved question worth investigating. It is the first and most consequential step — a poorly defined problem cannot be rescued by a strong design later.

10.2.2 Sources of Research Problems

TipWhere Research Problems Come From
  • Theory — gaps, contradictions, untested implications.
  • Practice — recurring practical challenges (clinical, classroom, industrial).
  • Literature — explicit “directions for future research”.
  • Replication — verifying contested findings.
  • Personal experience — observation, intuition, anomaly.
  • Policy / social need — public-interest priorities.
  • Funding-agency calls — DST, ICSSR, UGC, ICMR, CSIR.

10.2.3 Characteristics of a Good Research Problem

TipSeven Marks of a Good Problem
  1. Researchable — answerable with available methods.
  2. Significant — adds to theory or practice.
  3. Original — not a duplicate.
  4. Clear and unambiguous.
  5. Feasible — within time, budget, ethics.
  6. Ethical — does not harm subjects.
  7. Specific — narrow enough to investigate deeply.

10.2.4 Defining vs Delimiting the Problem

TipDefine vs Delimit
  • Defining — write the problem in a clear, testable statement.
  • Delimiting — set explicit boundaries (population, geography, time, variables).
  • Limitations — what the design cannot control (acknowledged in the final report).

A research problem is usually written as a declarative statement (“To examine the impact of …”) and, optionally, as a research question (“What is the effect of …?”).

10.3 Step 2 — Review of Related Literature

10.3.1 Purposes of a Literature Review

TipSeven Purposes of Literature Review
  1. Map existing knowledge.
  2. Identify gaps and contradictions.
  3. Avoid duplication.
  4. Refine the research problem.
  5. Suggest theoretical framework.
  6. Guide method and instrument choice.
  7. Establish the researcher’s credibility.

10.3.2 Sources

TipTwo Categories of Source
Source Examples
Primary Original research articles, theses, conference papers, lab reports
Secondary Textbooks, review articles, meta-analyses, encyclopaedias, abstracts

10.3.3 Major Databases and Repositories

TipDatabases the Researcher Must Know
  • Scopus, Web of Science — indexing services with citation metrics.
  • Google Scholar — broad academic search.
  • PubMed — biomedical.
  • ERIC — education.
  • JSTOR — humanities, social sciences.
  • SciHub / Anna’s Archive — controversial, not recommended.
  • Indian repositories — Shodhganga (theses), e-ShodhSindhu (journals), e-PG Pathshala (PG content), NDLI (digital library).
  • OER — DOAJ, OpenDOAR, OER Commons.

10.3.4 Types of Literature Review

TipFive Types of Literature Review
Type What it does
Narrative / Traditional Critical overview of selected works
Systematic Pre-registered protocol; transparent inclusion/exclusion (e.g., PRISMA guidelines)
Meta-analysis Statistical synthesis of quantitative findings
Scoping Maps the breadth of a field
Integrative Combines diverse research types, including theory and methods

10.3.5 Reference Management

Use a reference manager from day one: Mendeley, Zotero, EndNote, RefWorks, JabRef. Common citation styles: APA, MLA, Chicago, Harvard, Vancouver, IEEE.

10.4 Step 3 — Formulation of Objectives and Hypotheses

10.4.1 Research Objectives

TipSMART Research Objectives

Specific · Measurable · Achievable · Relevant · Time-bound. Distinguish general (overall) from specific objectives.

10.4.2 Hypothesis — Definition and Properties

A hypothesis is a tentative, testable statement about the relationship between two or more variables. Best definitions: Goode and Hatt — “a proposition which can be put to a test to determine its validity”; Kerlinger — “a conjectural statement of the relation between two or more variables”.

TipSix Qualities of a Good Hypothesis
  1. Testable / Falsifiable — measurable variables, refutable in principle (Popper).
  2. Specific — clear conditions and predictions.
  3. Empirically grounded — connects to theory or prior evidence.
  4. Parsimonious — simplest plausible explanation.
  5. Consistent with knowledge.
  6. States a relationship between variables.

10.4.3 Types of Hypotheses

TipFive Hypothesis Pairs / Types
Type What it says
Research / Working / Alternative (H₁) Anticipates a relationship or effect
Null (H₀) States no relationship; the default to be disproved
Directional Specifies direction (e.g., A > B)
Non-directional States only that a difference exists
Statistical Cast in inferential-statistics language

Other useful distinctions: descriptive, relational, causal; simple vs complex; a-priori vs ad-hoc.

10.4.4 Sources of Hypotheses

Theory, prior literature, personal observation, analogy, intuition, replication, and folk wisdom that needs testing.

10.4.5 Testing a Hypothesis — Type I and Type II Error

TipType I and Type II Error
H₀ is TRUE H₀ is FALSE
Reject H₀ Type I error (α) Correct
Fail to reject H₀ Correct Type II error (β)

Power = 1 − β — the probability of correctly rejecting a false H₀. Conventional α = 0.05; power ≥ 0.80.

10.5 Step 4 — Research Design

10.5.1 What a Design Specifies

TipFive Elements a Design Must Specify
  1. Method — experimental, descriptive, historical, qualitative, quantitative, mixed.
  2. Setting — lab, field, online, archive.
  3. Subjects — participants, materials, units.
  4. Procedure — sequence, timing, instructions.
  5. Measurement — variables, scales, instruments.

10.5.2 Major Design Families

Covered in detail in Topic 8: Pre-experimental, True experimental, Quasi-experimental, Descriptive (survey, case, correlational, comparative), Historical, Qualitative (phenomenology, ethnography, grounded theory, case study, narrative), Mixed-methods (Convergent Parallel, Explanatory Sequential, Exploratory Sequential, Embedded).

10.5.3 Design Quality — Validity Layers

TipFour Validity Layers (Shadish, Cook & Campbell, 2002)
Validity Question it answers
Statistical conclusion validity Are the inferences about covariation correct?
Internal validity Did the IV cause the DV?
Construct validity Do the measures capture the intended construct?
External validity Do the findings generalise beyond this study?

10.6 Step 5 — Sampling

10.6.1 Key Terms

TipSampling Vocabulary
  • Population — the full set the researcher wishes to generalise to.
  • Target population — the conceptual full set.
  • Accessible population — the set actually reachable.
  • Sampling frame — the operational list of units.
  • Sampling unit — individual or cluster.
  • Sample — the subset actually studied.
  • Parameter vs Statistic — population value vs sample estimate.
  • Sampling error — random difference between sample and population value.
  • Non-sampling error — bias from instrument, non-response, coverage.

10.6.2 Probability vs Non-Probability Sampling

TipSampling Methods at a Glance
Probability Non-Probability
Simple Random Sampling (SRS) Convenience
Stratified Random Sampling Purposive / Judgemental
Systematic Sampling Quota
Cluster Sampling Snowball
Multi-stage Sampling Voluntary / Self-selected
Probability-proportional-to-size (PPS) —

Probability sampling permits statistical generalisation; non-probability sampling does not.

10.6.3 Sample Size

Drivers: desired confidence level (usually 95 %), margin of error (often 5 %), population variability, expected effect size, design effect (DEFF) for cluster/stratified, subgroup analyses. Cochran’s formula is a standard starting point:

\[n_0 = \frac{Z^2 \cdot p (1-p)}{e^2}\]

where Z = 1.96 for 95 % CI, p = expected proportion (use 0.5 for max), e = margin of error.

10.7 Step 6 — Data Collection

10.7.1 Tools by Type of Data

TipTools and Methods of Data Collection
Quantitative Qualitative
Questionnaire In-depth interview
Structured interview Focus group discussion
Test / inventory Participant observation
Rating scale (Likert, Thurstone, Guttman, semantic differential) Field notes
Observation schedule Document analysis
Existing dataset Photo / video / audio data

10.7.2 Standardising the Instrument

Before main data collection, every instrument must be piloted, validated (content, criterion, construct) and reliability-checked (test-retest, parallel-form, split-half, KR-20/21, Cronbach’s α).

10.7.3 Ethics During Collection

Informed consent · Anonymity & confidentiality · Right to withdraw · Minimal risk · IRB / IEC approval · Vulnerable groups protection. (Full treatment in Topic 12.)

10.8 Step 7 — Data Analysis

10.8.1 Preparing the Data

TipData Preparation Sequence

Editing → Coding → Classification → Tabulation → Visualisation → Analysis

10.8.2 Choosing the Analytic Technique

TipPicking an Analysis
  • One variable, descriptive: frequency, mean, SD, percentiles, distribution shape.
  • Two variables, association: Pearson’s r, Spearman’s ρ, χ², t-test, ANOVA, Mann-Whitney, Wilcoxon, Kruskal-Wallis.
  • Predicting outcome: linear regression, logistic regression, multiple regression.
  • Reducing dimensions: factor analysis, PCA.
  • Modelling: SEM, multilevel modelling, time series, machine-learning models.
  • Qualitative: thematic analysis (Braun & Clarke 2006), content analysis, narrative analysis, discourse analysis, IPA, grounded theory’s constant comparison.

10.8.3 Software

Quantitative: SPSS, R, Python, SAS, Stata, JASP, jamovi, MATLAB. Qualitative: NVivo, ATLAS.ti, MAXQDA, Dedoose, QDA Miner.

10.9 Step 8 — Interpretation and Conclusion

10.9.1 What Interpretation Adds

Analysis produces numbers, themes, models. Interpretation explains what the findings mean in relation to theory, prior research, and the original problem.

TipWhat Interpretation Must Cover
  1. Compare findings with hypotheses and prior literature.
  2. Explain unexpected results.
  3. Discuss practical and theoretical implications.
  4. Acknowledge limitations.
  5. Suggest directions for future research.

10.9.2 Common Inferential Errors

Correlation-causation confusion · Over-generalisation beyond sample · Ignoring effect-size · p-hacking · HARKing (hypothesising after results are known) · Confirmation bias · Survivorship bias.

10.10 Step 9 — Reporting and Publication

10.10.1 The Standard Research Report Structure (IMRaD)

TipIMRaD — The Universal Article Structure

Introduction → Methods → Results → and → Discussion (with Conclusion, References, Appendices, Abstract on top).

10.10.2 Outlets

TipWhere Research Gets Published
  • Peer-reviewed journals (UGC-CARE, Scopus, WoS, ABDC).
  • Conferences — proceedings, posters.
  • Books and chapters — Springer, Elsevier, OUP, Routledge, Sage.
  • Theses — Shodhganga deposit mandatory for Indian PhDs.
  • Working papers and preprints — SSRN, arXiv, bioRxiv, PsyArXiv.
  • Public scholarship — policy briefs, blogs, newspaper op-eds.

10.10.3 Reporting Standards

Use the right reporting standard for the study type: CONSORT (trials), STROBE (observational), PRISMA (systematic reviews), COREQ / SRQR (qualitative), GRADE (evidence quality), MIAME (microarrays). Honest reporting includes declaring funding, conflicts of interest, ethics approval, data availability.

10.10.4 Citation, Plagiarism, Predatory Journals — Brief Pointers

Plagiarism, conflicts of interest, predatory journals, data fabrication are full Topic-12 (Research Ethics) territory. Important here: always cite primary sources, use a reference manager, run Turnitin / iThenticate, avoid predatory journals (Beall’s list legacy).

10.11 Practice Questions

Q 01 Sequence Easy

The FIRST step in the research process is:

  • AReview of related literature
  • BIdentification of the research problem
  • CFormulation of hypothesis
  • DData collection
View solution
Correct Option: B
Identifying the research problem is always step 1 — every other step depends on it.
Q 02 Sequence Medium

Arrange the following steps in correct order:

(i) Sampling   (ii) Literature review   (iii) Hypothesis formulation   (iv) Problem identification

  • A(iv) → (ii) → (iii) → (i)
  • B(ii) → (iv) → (i) → (iii)
  • C(iv) → (iii) → (ii) → (i)
  • D(i) → (ii) → (iii) → (iv)
View solution
Correct Option: A
Problem → Literature → Hypothesis → Sampling.
Q 03 Problem Easy

Which is NOT a characteristic of a good research problem?

  • AResearchable
  • BOriginal
  • CVague and broad
  • DFeasible
View solution
Correct Option: C
A good problem is clear and specific, not vague.
Q 04 Hypothesis Medium

"A hypothesis is a proposition which can be put to a test to determine its validity." This definition is by:

  • AKerlinger
  • BGoode and Hatt
  • CC.R. Kothari
  • DKarl Popper
View solution
Correct Option: B
Goode and Hatt, Methods in Social Research (1952).
Q 05 Null Medium

The null hypothesis (H₀) states that:

  • AThere IS a difference / relationship
  • BThere is NO difference / relationship
  • CThe relationship is positive
  • DThe relationship is negative
View solution
Correct Option: B
H₀ is the "no effect" default; statistical inference tries to reject H₀.
Q 06 Type I / II Hard

Rejecting a TRUE null hypothesis is called:

  • AType I error (α)
  • BType II error (β)
  • CSampling error
  • DPower error
View solution
Correct Option: A
Type I (α) = false positive (rejecting true H₀). Type II (β) = false negative (failing to reject false H₀). Conventionally α = 0.05, power = 1 − β ≥ 0.80.
Q 07 Lit Review Medium

A systematic review that statistically pools the quantitative findings of multiple studies is called:

  • ANarrative review
  • BScoping review
  • CMeta-analysis
  • DAnnotated bibliography
View solution
Correct Option: C
Meta-analysis = statistical synthesis of effect sizes across studies (Glass, 1976).
Q 08 Lit Review Medium

The PRISMA guidelines are used in:

  • ARandomised controlled trials
  • BObservational studies
  • CSystematic reviews and meta-analyses
  • DQualitative case studies
View solution
Correct Option: C
PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses). CONSORT for trials; STROBE for observational; COREQ/SRQR for qualitative.
Q 09 Sampling Easy

In stratified random sampling, the population is FIRST divided into:

  • ARandom units
  • BHomogeneous strata
  • CHeterogeneous clusters
  • DQuotas
View solution
Correct Option: B
Stratified = population divided into homogeneous strata (e.g., gender, region), then random sampling within each stratum. Cluster = heterogeneous clusters, random selection of clusters.
Q 10 Sampling Medium

A researcher selects every 10th name from a college register. This is:

  • ASimple random sampling
  • BSystematic sampling
  • CCluster sampling
  • DSnowball sampling
View solution
Correct Option: B
Every k-th element from an ordered list = Systematic sampling.
Q 11 Reliability Medium

Cronbach's α primarily assesses:

  • AValidity of a test
  • BInternal consistency reliability of a test
  • CDiscriminant power of items
  • DTest-retest reliability
View solution
Correct Option: B
Cronbach's α = average inter-item correlation; an estimate of internal consistency. Conventionally α ≥ 0.70 acceptable.
Q 12 Likert Medium

Likert scales typically generate data on which scale of measurement?

  • ANominal
  • BOrdinal (often treated as interval)
  • CInterval
  • DRatio
View solution
Correct Option: B
Likert items yield ordinal data; with multiple items, totals are commonly treated as interval for parametric analysis.
Q 13 Analysis Medium

The sequence of qualitative data preparation in grounded theory is:

  • ASelective → Axial → Open coding
  • BOpen → Axial → Selective coding
  • CAxial → Open → Selective coding
  • DOpen → Selective → Axial coding
View solution
Correct Option: B
Glaser & Strauss coding sequence: Open → Axial → Selective (each refines into higher-order categories).
Q 14 Interpretation Hard

"Hypothesising After the Results are Known" — fabricating a hypothesis to match the data afterwards — is called:

  • Ap-hacking
  • BHARKing
  • CCherry-picking
  • DConfirmation bias
View solution
Correct Option: B
HARKing (Hypothesising After the Results are Known) — coined by Norbert Kerr (1998). A widely-flagged research-integrity issue.
Q 15 IMRaD Easy

"IMRaD" is the standard structure of a research article. It stands for:

  • AIdea → Method → Research → Display
  • BIntroduction → Methods → Results → and Discussion
  • CInference → Measure → Replicate → and Document
  • DIssue → Method → Reasoning → and Decision
View solution
Correct Option: B
IMRaD = Introduction, Methods, Results, and Discussion (with Abstract, References, Appendices).
Q 16 Indian Repositories Medium

In India, doctoral theses are deposited in the national repository called:

  • AVIDWAN
  • Be-PG Pathshala
  • CShodhganga
  • De-ShodhSindhu
View solution
Correct Option: C
Shodhganga (INFLIBNET) is the national theses repository. Synopses are at Shodhgangotri; expert database at VIDWAN; PG content at e-PG Pathshala; journal consortium at e-ShodhSindhu.
Q 17 Step Function Medium

Match each activity to its research step:

(i) Running Cronbach's α on the pilot questionnaire (a) Sampling
(ii) Cochran's formula calculation (b) Data collection prep
(iii) PRISMA flow diagram (c) Reporting
(iv) IMRaD structure (d) Literature review
  • A(i)-b, (ii)-a, (iii)-d, (iv)-c
  • B(i)-a, (ii)-b, (iii)-c, (iv)-d
  • C(i)-c, (ii)-d, (iii)-a, (iv)-b
  • D(i)-d, (ii)-c, (iii)-b, (iv)-a
View solution
Correct Option: A
Cronbach's α → data-collection prep; Cochran's formula → sampling; PRISMA → literature review; IMRaD → reporting.
Q 18 Sampling Error Hard

Which of the following is a NON-SAMPLING error?

  • AA random difference between sample mean and population mean
  • BBias from leading question wording in a survey
  • CVariance reduced by increasing sample size
  • DConfidence-interval width
View solution
Correct Option: B
Non-sampling error = instrument bias, non-response, coverage failure, data-entry mistakes. Larger samples do NOT cure non-sampling error.
Q 19 Population Medium

A numerical characteristic of a POPULATION (such as μ or σ) is called:

  • AStatistic
  • BParameter
  • CEstimate
  • DVariable
View solution
Correct Option: B
Parameter = population value. Statistic = sample value, used to estimate the parameter.
Q 20 Sequence Hard

Arrange the FULL 9-step research process in correct order:

(i) Sampling   (ii) Problem identification   (iii) Hypothesis   (iv) Data collection   (v) Literature review   (vi) Reporting   (vii) Research design   (viii) Interpretation   (ix) Data analysis

  • A(ii) → (v) → (iii) → (vii) → (i) → (iv) → (ix) → (viii) → (vi)
  • B(ii) → (iii) → (v) → (vii) → (i) → (iv) → (ix) → (viii) → (vi)
  • C(v) → (ii) → (iii) → (vii) → (iv) → (i) → (ix) → (viii) → (vi)
  • D(ii) → (v) → (vii) → (iii) → (i) → (iv) → (ix) → (viii) → (vi)
View solution
Correct Option: A
Problem → Literature → Hypothesis → Design → Sampling → Collection → Analysis → Interpretation → Reporting.

10.12 Quick Recall

ImportantQuick recall
  • 9 steps: Problem → Literature → Hypothesis → Design → Sampling → Collection → Analysis → Interpretation → Reporting.
  • Good research problem: Researchable, Significant, Original, Clear, Feasible, Ethical, Specific.
  • Lit-review purposes: map, gap, avoid duplication, refine problem, framework, method guide, credibility.
  • Lit-review types: Narrative · Systematic (PRISMA) · Meta-analysis (Glass 1976) · Scoping · Integrative.
  • Hypothesis definitions: Goode & Hatt — “proposition testable for validity”; Kerlinger — “conjectural statement of relation between two or more variables”.
  • Hypothesis types: Research/H₁ · Null/H₀ · Directional · Non-directional · Statistical.
  • Type I (α) = false positive · Type II (β) = false negative · Power = 1 − β ≥ 0.80 · α = 0.05.
  • 4 validity layers (Shadish, Cook & Campbell 2002): Statistical conclusion · Internal · Construct · External.
  • Sampling vocabulary: Population · Target · Accessible · Frame · Unit · Parameter vs Statistic · Sampling error vs Non-sampling error.
  • Probability: SRS · Stratified · Systematic · Cluster · Multi-stage · PPS. Non-probability: Convenience · Purposive · Quota · Snowball · Voluntary.
  • Cochran’s formula: n₀ = Z²·p(1-p)/e²; Z=1.96 for 95% CI.
  • Reliability: test-retest, parallel-form, split-half, KR-20/21, Cronbach’s α ≥ 0.70.
  • Validity: content, construct, criterion (concurrent & predictive), face.
  • Quantitative software: SPSS · R · Python · SAS · Stata · JASP · jamovi. Qualitative: NVivo · ATLAS.ti · MAXQDA · Dedoose.
  • Inferential errors: correlation-causation · over-generalisation · p-hacking · HARKing · confirmation/survivorship bias.
  • IMRaD: Introduction · Methods · Results · and Discussion (plus Abstract, References, Appendices).
  • Reporting standards: CONSORT (trials) · STROBE (observational) · PRISMA (reviews) · COREQ/SRQR (qualitative) · GRADE (evidence).
  • Indian repositories: Shodhganga (theses) · Shodhgangotri (synopses) · VIDWAN (experts) · e-ShodhSindhu (journals) · e-PG Pathshala (PG content) · NDLI.