Loading...
Loading...
Research methodology is the systematic framework that guides how a researcher designs, collects, analyzes, and interprets data to answer a research question. For PhD students in India, understanding research methodology is mandatory — it forms the backbone of the entire thesis and determines the validity of your contributions.
| Type | Definition | Example | |------|-----------|---------| | Basic/Pure Research | Expanding knowledge without immediate application | Proving a new mathematical theorem | | Applied Research | Solving a specific practical problem | Improving network security algorithm | | Action Research | Improving practice in a specific setting | Teaching methodology improvement | | Developmental Research | Designing new systems/products | Developing a new deep learning model |
Quantitative Research:
Qualitative Research:
Mixed Methods:
A research design is the blueprint of the study. It addresses:
Experimental Design:
Quasi-Experimental:
Descriptive Design:
Correlational Design:
Population → Sampling Frame → Sample → Data Collection
Probability Sampling (each unit has known chance):
├── Simple Random Sampling — lottery method
├── Stratified Sampling — divide into strata, then random
├── Cluster Sampling — select groups/clusters
└── Systematic Sampling — every nth element
Non-Probability Sampling:
├── Purposive/Judgmental — researcher selects based on criteria
├── Convenience — readily available participants
├── Snowball — existing participants recruit more
└── Quota — proportional representation without randomization
Sample Size Determination:
n = (Z² × p × q) / e²
A literature review is NOT just a summary of papers. It:
Step 1 — Define Search Strategy:
Boolean Operators:
- AND: narrows search (machine learning AND cybersecurity)
- OR: broadens search (CNN OR convolutional neural network)
- NOT: excludes terms (NLP NOT speech recognition)
Databases to use:
- IEEE Xplore (CS/ECE research)
- ACM Digital Library (Computer Science)
- Google Scholar (all domains, free)
- Scopus / Web of Science (paid, institutional access)
- PubMed (medical/biomedical)
- SSRN (economics, social science)
Step 2 — Inclusion/Exclusion Criteria:
Step 3 — PRISMA Framework:
Identification → Screening → Eligibility → Included
(All records) (Remove (Full-text (Final set
duplicates, review) for synthesis)
titles)
Step 4 — Synthesis (NOT Summary):
| Criterion | What to Check | |-----------|---------------| | Relevance | Does it directly address your research problem? | | Authority | Who are the authors? What are their credentials? | | Accuracy | Is the methodology sound? Are findings reproducible? | | Currency | Is the information recent enough for your field? | | Objectivity | Is there potential bias? Who funded the research? |
Types of Questions:
Questionnaire Design Principles:
Reliability Testing:
| Type | Control | Depth | When to Use | |------|---------|-------|-------------| | Structured | High | Low | Comparative studies | | Semi-structured | Medium | Medium | Exploratory + validation | | Unstructured | Low | High | Deep exploration, pilot studies | | Focus Group | Medium | Medium-High | Group dynamics, consensus |
import pandas as pd
import numpy as np
data = pd.read_csv('research_data.csv')
# Measures of Central Tendency
print("Mean:", data['score'].mean())
print("Median:", data['score'].median())
print("Mode:", data['score'].mode()[0])
# Measures of Dispersion
print("Std Dev:", data['score'].std())
print("Variance:", data['score'].var())
print("Range:", data['score'].max() - data['score'].min())
print("IQR:", data['score'].quantile(0.75) - data['score'].quantile(0.25))
Research Question Type → Appropriate Statistical Test
Comparing 2 groups (parametric):
└── Independent t-test (unpaired)
└── Paired t-test (same subjects, before/after)
Comparing 3+ groups (parametric):
└── One-way ANOVA → Post-hoc: Tukey HSD, Bonferroni
Non-parametric alternatives:
└── Mann-Whitney U (instead of t-test)
└── Kruskal-Wallis (instead of ANOVA)
└── Wilcoxon signed-rank (instead of paired t-test)
Relationship between variables:
└── Pearson's r (both continuous, normal)
└── Spearman's rho (ordinal or non-normal)
└── Chi-Square (both categorical)
Prediction:
└── Simple/Multiple Linear Regression
└── Logistic Regression (binary outcome)
└── SEM - Structural Equation Modeling
Important: Statistical significance ≠ Practical significance
Front Matter:
├── Title Page
├── Declaration
├── Certificate (Guide's)
├── Acknowledgments
├── Abstract (300–500 words)
├── Table of Contents
├── List of Tables
├── List of Figures
└── List of Abbreviations
Main Body:
├── Chapter 1: Introduction
│ ├── Background and Motivation
│ ├── Problem Statement
│ ├── Research Objectives
│ ├── Research Questions/Hypotheses
│ ├── Scope and Limitations
│ └── Thesis Organization
│
├── Chapter 2: Literature Review
│ ├── Theoretical Background
│ ├── Review of Related Work
│ ├── Critical Analysis
│ └── Research Gap Identification
│
├── Chapter 3: Research Methodology
│ ├── Research Design
│ ├── Data Collection Methods
│ ├── Sampling Strategy
│ ├── Tools and Instruments
│ └── Validity and Reliability
│
├── Chapter 4: Proposed System / Framework / Model
│ (Domain-specific — implementation details)
│
├── Chapter 5: Results and Discussion
│ ├── Data Presentation
│ ├── Statistical Analysis Results
│ ├── Interpretation
│ └── Comparison with Existing Work
│
└── Chapter 6: Conclusion and Future Work
├── Summary of Contributions
├── Implications
├── Limitations
└── Future Research Directions
Back Matter:
├── References (APA / IEEE / MLA / Vancouver)
├── Appendices
└── Index (optional)
A good abstract has 5 elements in ≤ 300 words:
| Style | Used In | Format | |-------|---------|--------| | APA 7th | Social sciences, psychology | Author, Year. Title. Journal, Vol(Issue), pages. | | IEEE | CS, ECE, Engineering | [1] A. Author, "Title," Journal, vol. X, pp. Y–Z, Year. | | MLA | Humanities | Author. "Title." Journal, vol., no., Year, pp. | | Vancouver | Medical, health sciences | Author. Title. Journal. Year;Vol(Issue):pages. | | Harvard | Business, management | Author (Year) Title. Journal, Vol(Issue), pp. |
UGC-CARE (Current list required for PhD):
International Indexing:
| Index | Prestige | Access | Cost | |-------|---------|--------|------| | SCI/SCI-E | Highest | Mostly paid | High | | Scopus | High | Paid | Medium-High | | ESCI | Medium-High | Mostly paid | Medium | | DOAJ | Medium | Open access | Free | | Google Scholar | Low barrier | Free | Free |
Beware Predatory Journals:
Abstract (200–250 words)
↓
I. Introduction
↓
II. Related Work / Literature Review
↓
III. Proposed Methodology / System Design
↓
IV. Experimental Setup / Implementation
↓
V. Results and Discussion
↓
VI. Conclusion and Future Work
↓
References
Responding to Reviewers:
Q: Why did you choose this research topic?
Answer: State the problem it solves, its relevance to current industry/society, and why existing solutions are inadequate. Show personal motivation + academic gap.
Q: What is the research gap your work addresses?
Answer: After reviewing X papers in [domain], I found that [specific limitation]. My work addresses this by [your contribution].
Q: What are the limitations of your study?
Answer: Be honest — mention sample size, geographical scope, time period, computational resources. Follow with how future work can address these.
Q: How is your work different from [cite related work]?
Answer: Compare on: methodology, dataset, results, scope, novelty. Show your improvement metrics.
Q: What would you do differently if you were to start again?
Answer: Shows self-reflection. Mention better methodology choices, larger dataset, different statistical approach.
Complete doctoral research guide covering research design, literature review, data collection, thesis writing, publication strategy, and viva preparation for PhD students.
48 pages · 1.8 MB · Updated 2025-03-01
Quantitative research collects numerical data analyzed statistically — surveys, experiments, measurements. Qualitative research collects non-numerical data — interviews, case studies, observations. Mixed methods combine both. PhD students choose based on research objective and nature of the problem.
UGC 2022 regulations require at least one research paper published in a UGC-CARE listed journal before thesis submission. Most universities additionally require 2–3 publications. Some IITs/NITs require 2 SCI/SCOPUS indexed papers.
A hypothesis is a testable statement about the relationship between variables. Null hypothesis (H0) states no relationship; Alternative hypothesis (H1) states a relationship exists. Example: H0: There is no significant difference in ML model accuracy between Algorithm A and B. H1: Algorithm A has significantly higher accuracy than Algorithm B.
DBMS Complete Notes — B.Tech CS Sem 4
Database Management Systems
Compiler Design — Complete Notes CS Sem 6
Compiler Design
Machine Learning Complete Notes — B.Tech CS Sem 6
Machine Learning
Digital Electronics — Complete Notes IT Sem 1
Digital Electronics
Analog Electronics — Complete Notes ECE Sem 3
Analog Electronics
Your feedback helps us improve notes and tutorials.