Research Methodology Notes

Research is a step-by-step process of finding answers to questions. It means searching for new knowledge in a systematic way.

Simple words: Research = Re (again) + Search (to look for) = Looking for answers again and again to find truth.

Research Methodology is the science that teaches us how to do research properly. It provides rules and methods to get correct and reliable results.

In Academic (College/University) Context:

• Helps students learn how to find and verify information
• Builds new knowledge and theories
• Develops critical thinking and problem-solving skills
• Helps complete assignments, projects, and PhD thesis

In Professional (Job/Work) Context:

• Helps companies make better decisions based on facts
• Finds solutions to business problems
• Helps create new products and improve existing ones
• Gives competitive advantage over other companies

• 1. Exploration (Exploring): When we know very little about a topic, we explore to understand it better. It answers "What is happening here?"
  Example: "Why are many students failing online exams?" — We explore to find possible reasons.
• 2. Description (Describing): We describe the characteristics of a situation, group, or phenomenon. It answers "What, where, when, how?"
  Example: "65% of college students use mobile phones for studying, 35% use laptops."
• 3. Explanation (Explaining): We explain why something happens. It answers "Why?" and shows cause-effect relationships.
  Example: "Students fail because they don't get personalized feedback on their weak topics."
• 4. Prediction (Predicting): We forecast what will happen in the future based on past patterns.
  Example: "Based on weekly quiz scores, the system predicts which students will fail the final exam."
• 5. Application (Applying): We use research findings to solve real problems.
  Example: "Research shows personalized learning works, so we build an app that provides it to all students."

Basic Research (Pure Research): Research done for knowledge only. No immediate practical use.
Example: "How does the human brain store memories?" — This helps us understand, but doesn't directly solve a problem.

Applied Research: Research done to solve a specific real-world problem.
Example: "How can we design an app that helps students remember what they study?" — This directly solves a problem.

Simple Difference: Basic = Knowledge for knowledge's sake. Applied = Knowledge to solve problems.

Qualitative Research: Deals with words, feelings, opinions. Data is non-numerical. Answers "Why?" and "How?"
Example: Interview 20 students and ask "How do you feel about online exams?" — Answers: "Stressful", "Convenient", "Difficult"

Quantitative Research: Deals with numbers and statistics. Data is numerical. Answers "How many?" and "How much?"
Example: Survey 500 students and ask "Rate online exams from 1 to 5" — Average rating = 3.2

Cross-sectional Research: Data collected at ONE point in time. Like taking a snapshot photo.
Example: Survey 1000 students in March 2024 about their study habits.

Longitudinal Research: Data collected over MULTIPLE time points from the SAME people. Like making a video.
Example: Track the SAME 1000 students for 4 years, measuring their scores every semester.

Which to use? Cross-sectional is faster and cheaper. Longitudinal shows changes over time but takes longer and costs more.

• Step 1: Identify the Problem — Find what needs to be studied. Ask "What problem exists?"
  Example: "Many students fail math exams."
• Step 2: Review Literature — Read what other researchers have already found about this topic.
  Example: Read 30 research papers on why students fail math.
• Step 3: Formulate Research Questions — Convert the problem into specific questions you want to answer.
  Example: "Does daily practice with an AI app improve math scores?"
• Step 4: Design the Research — Plan how you will find answers. Choose methods, participants, tools.
  Example: Take 200 students, give AI app to 100 (group A), no app to 100 (group B).
• Step 5: Collect Data — Gather information using surveys, tests, interviews, etc.
  Example: Conduct math test for both groups after 2 months.
• Step 6: Analyze Data — Use statistics or thematic analysis to find patterns.
  Example: Compare average scores of Group A vs Group B.
• Step 7: Report Findings — Write conclusions, share recommendations.
  Example: "Group A scored 18% higher. Conclusion: AI app helps."

• Technology changes fast: What is new today may be old tomorrow. Research must be on current topics like AI, Blockchain, Cloud Computing.
• You build things: In CS research, you often create working software, algorithms, or systems — not just write theories.
• Focus on efficiency: CS research asks "How fast? How much memory? How accurate?"
• Large amounts of data: CS research often works with big datasets (millions of records, GB/TB of data).

• Experimental Method: Build a system and test its performance. Compare with existing systems.
  Example: Build two search algorithms, test which one is faster.
• Simulation Method: Create a computer model of a real system and run experiments on it.
  Example: Simulate network traffic to test how a new routing protocol performs.
• Survey Method: Ask users about their experience with software.
  Example: "Rate the usability of our new app on a scale of 1-5."
• Case Study Method: Deep study of one specific implementation.
  Example: "How did Company X successfully migrate to cloud computing?"
• Design Science: Create a new IT artifact (software, method, framework) that solves a problem.
  Example: Design a new algorithm to detect fake news on social media.

Research Design is a complete plan or blueprint that tells you HOW to conduct your research. It answers questions like: What methods to use? How to collect data? How many people to study? How to analyze results?

Simple words: If research is building a house, research design is the architectural blueprint. Without a blueprint, the house will be weak and may collapse.

• To provide a clear roadmap for the entire research study
• To ensure results are accurate and trustworthy
• To save time, money, and effort by planning ahead
• To help other researchers repeat your study and verify results

• 1. Ensures Accuracy: Good design makes sure you measure what you actually want to measure.
  Example: If you want to measure "learning improvement," you must test students BEFORE and AFTER using the app.
• 2. Reduces Mistakes (Bias): Proper design removes researcher's personal opinions from affecting results.
  Example: Randomly assigning students to groups, not putting all bright students in one group.
• 3. Saves Resources: Planning prevents wasting time and money on wrong methods.
  Example: Deciding sample size of 200 instead of 2000 (saves money) or 20 (too small for accurate results).
• 4. Allows Replication: Other researchers can repeat your study exactly to confirm findings.
  Example: Another college can use your same method to verify if SmartLearn works for their students.
• 5. Increases Credibility: Well-designed research is more likely to be published in good journals and trusted by others.

1. Exploratory Design (Exploring)

• Use when: You know very little about the topic. It's a new area.
• Goal: To explore and understand the problem better.
• Methods: Literature review, expert interviews, focus groups.
• Output: New ideas, hypotheses for future research.
• Example: "Why are students not using online learning platforms? Let's interview 20 students to find possible reasons."

2. Descriptive Design (Describing)

• Use when: You want to describe characteristics of a group or situation.
• Goal: To answer "What?", "How many?", "How much?"
• Methods: Surveys, observations, case studies.
• Output: Percentages, averages, frequencies.
• Example: "Survey 500 students to find out how many use smartphones for studying (72%), laptops (28%)."

3. Experimental Design (Experimenting)

• Use when: You want to prove that one thing CAUSES another thing.
• Goal: To establish cause-effect relationship.
• Key features: Researcher manipulates one variable, random assignment, control group.
• Example: Randomly assign 200 students to Group A (use SmartLearn) and Group B (no app). Compare scores. If Group A scores higher → App CAUSED improvement.

4. Quasi-Experimental Design (Almost Experimental)

• Use when: You cannot randomly assign participants (e.g., existing classrooms).
• Goal: Still try to establish causality, but less strong than true experiment.
• Difference from Experimental: No random assignment.
• Example: Take Class A (already formed) as Experimental group, Class B as Control group. Compare scores. (Problem: Class A might already be brighter students.)

1. Objectives (What do you want to achieve?)

• Clear, specific statements of what your research will accomplish.
• Should be SMART: Specific, Measurable, Achievable, Relevant, Time-bound.
• Example: "To determine if SmartLearn AI app increases math exam scores by at least 10% within 2 months."

2. Hypotheses (What do you predict will happen?)

• A testable prediction about the relationship between variables.
• Null Hypothesis (H₀): NO relationship or NO difference. "SmartLearn does NOT improve scores."
• Alternative Hypothesis (H₁): THERE IS a relationship or difference. "SmartLearn improves scores."
• Research tries to REJECT the null hypothesis.

3. Variables (What are you measuring and manipulating?)

• Independent Variable (IV): What the researcher changes or controls. The "cause".
  Example: Using SmartLearn app (YES/NO).
• Dependent Variable (DV): What is measured as the outcome. The "effect".
  Example: Exam scores after 2 months.
• Control Variables: Things kept the same for all groups to ensure fair comparison.
  Example: Same teacher, same syllabus, same test.

4. Methods of Data Collection (How will you gather information?)

• Surveys/Questionnaires (for opinions, behaviors)
• Tests/Exams (for measuring knowledge or skill)
• Interviews (for deep understanding)
• Observations (for watching behavior)
• Experiments (for testing cause-effect)

5. Sampling Design (Who will be in your study?)

• Population: The entire group you want to study (e.g., all 10,000 students in a college).
• Sample: The subset you actually study (e.g., 200 students).
• Sampling method: How you select these 200 students (random, stratified, etc.).
• Sample size: How many participants you need (calculated using formula).

Validity means: Are you measuring what you INTEND to measure? Does your test actually measure the concept it claims to measure?

Simple example: If you want to measure "intelligence", using a ruler is NOT valid. Using an IQ test IS valid (if it's a good IQ test).

1. Internal Validity

• Question: Is the change in the outcome REALLY caused by your treatment? Or by something else?
• Simple: Did the APP cause improvement? Or did students study extra on their own?
• Threats to Internal Validity: Students studying extra, teacher giving extra help, students already being smarter.
• How to ensure good internal validity: Random assignment, control group, keep everything same except the treatment.

2. External Validity

• Question: Can your findings be applied (generalized) to other people, places, and situations?
• Simple: If SmartLearn works for engineering students in Pune, will it also work for arts students in Mumbai?
• Threats to External Validity: Only testing one type of student, one college, one city.
• How to ensure good external validity: Use diverse sample (different colleges, cities, student types).

3. Construct Validity

• Question: Does your measurement tool actually represent the theoretical concept?
• Simple: Does your "engagement score" actually measure student engagement? Or is it just measuring time spent?
• Example: A student may spend 5 hours on the app but be distracted. High time ≠ high engagement.
• How to ensure good construct validity: Use established measurement tools, multiple indicators of the same concept.

Reliability means: If you repeat the measurement, do you get the SAME result? It is about CONSISTENCY.

Simple example: A weighing scale is reliable if it shows 65kg every time you step on it (even if your actual weight is 70kg).

Primary Data is original data collected directly by the researcher for their specific study. It is first-hand information — collected by YOU for YOUR research.

Secondary Data is data already collected by others (government reports, previous research, company records, internet sources).

Simple difference: Primary = You collect it yourself. Secondary = Someone else already collected it.

1. Surveys / Questionnaires

• Written set of questions given to participants.
• Can be online (Google Forms, SurveyMonkey) or paper-based.
• Good for: Collecting data from many people quickly.
• Example: Send Google Form to 500 students asking about their study habits.

2. Interviews

• One-on-one conversation between researcher and participant.
• Types: Structured (fixed questions), Semi-structured (guide questions), Unstructured (free conversation).
• Good for: Deep understanding, personal stories, feelings.
• Example: Interview 20 students for 30 minutes each about their experience with online learning.

3. Observations

• Watching and recording behavior without interfering.
• Types: Participant (researcher joins the group) vs Non-participant (researcher watches from outside).
• Good for: Seeing what people actually do, not just what they say.
• Example: Sit in a classroom and count how many students use phones during lecture.

4. Experiments

• Researcher manipulates one variable and measures effect on another.
• Done in controlled settings (lab) or real-world (field).
• Good for: Proving cause and effect relationships.
• Example: Give one group a new teaching method, another group old method, compare test scores.

5. Focus Groups

• Group discussion (6-10 people) moderated by researcher.
• Participants discuss a topic together, researcher observes.
• Good for: Exploring attitudes, generating ideas, understanding group opinions.
• Example: Gather 8 students to discuss "What would make online learning better?"

• Population: The ENTIRE group you want to study. Example: All 10,000 students in a university.
• Sample: A smaller group taken from the population that you actually study. Example: 200 students from the university.
• Sampling Frame: The list from which you select your sample. Example: University enrollment database.
• Sampling: The process of selecting a sample from the population.

Why sample? Why not study everyone? Because studying everyone (census) is too expensive, takes too much time, and is often impossible.

Every member of the population has a KNOWN and EQUAL chance of being selected. This is the best method for accurate, generalizable results.

1. Simple Random Sampling

• Every person has equal chance. Like a lottery.
• How to do: Use random number generator, pick names from a hat.
• Pros: Most unbiased, very accurate.
• Cons: Need complete list of all population members.
• Example: Put all 10,000 student names in a computer, randomly select 200 names.

2. Systematic Sampling

• Select every kth person from a list.
• Formula: k = Population size / Sample size. Example: 10000/200 = 50. Select every 50th student.
• Pros: Easier than simple random, still unbiased if list is random.
• Cons: If list has a pattern, sample may be biased.
• Example: From the student roll number list, pick every 50th student.

3. Stratified Sampling

• Divide population into groups (strata) based on a characteristic, then randomly sample from each group.
• Pros: Ensures all groups are represented.
• Cons: Need to know the characteristic for all population members.
• Example: Divide students by branch (CS, IT, Mech, Civil). Take 50 students randomly from each branch. Total 200.

4. Cluster Sampling

• Divide population into clusters (natural groups), randomly select some clusters, study ALL members of selected clusters.
• Pros: Cost-effective for large, spread-out populations.
• Cons: Less accurate than other methods.
• Example: There are 50 colleges. Randomly select 5 colleges. Study ALL students in those 5 colleges.

Not every member has a chance to be selected. Selection is based on convenience or researcher judgment. Used when random sampling is not possible.

1. Convenience Sampling

• Select whoever is easily available.
• Pros: Very fast, cheap, easy.
• Cons: High bias, cannot generalize results.
• Example: Survey students in your own classroom because they are right there.

2. Purposive / Judgmental Sampling

• Researcher deliberately selects participants with specific characteristics.
• Pros: Good for qualitative research, expert opinions.
• Cons: Researcher bias can affect selection.
• Example: Only interview students who failed the exam to understand why they failed.

3. Snowball Sampling

• Existing participants recruit more participants from their network.
• Pros: Can reach hidden or hard-to-find populations.
• Cons: Sample may be biased (similar people refer similar people).
• Example: Find one student who uses drugs, ask them to refer other students who use drugs.

4. Quota Sampling

• Researcher ensures specific proportions are met, but selection within quotas is not random.
• Pros: Ensures representation of subgroups.
• Cons: Within-quota selection can be biased.
• Example: Ensure 50% male, 50% female in the sample. But you choose which males and females conveniently.

Factors that affect sample size:

• Population size (larger population → larger sample needed, but not proportionally)
• Margin of error (smaller error → larger sample needed)
• Confidence level (95% common, 99% needs larger sample)
• Population variability (more diversity → larger sample needed)
• Budget and time available (less money/time → smaller sample)

Simple Sample Size Formula (for known population):

Descriptive Statistics describes your sample data (mean, percentage, standard deviation).

Inferential Statistics allows you to draw conclusions about the entire POPULATION based on your SAMPLE data. It helps you make predictions and test hypotheses.

Simple example: You survey 500 students (sample) and find 70% like online learning. Inferential statistics helps you conclude that 68-72% of ALL students (population) like online learning.

Hypothesis testing is a formal process to decide if your results are real or just happened by chance.

Two Types of Hypotheses:

• Null Hypothesis (H₀): "No difference" or "No relationship". This is what we assume is true.
• Alternative Hypothesis (H₁): "There IS a difference" or "There IS a relationship". This is what we want to prove.

The p-value (Very Important):

• p-value = Probability that the results happened by chance (if H₀ is true).
• If p-value < 0.05: Results are "statistically significant". We reject H₀ and accept H₁.
• If p-value > 0.05: Results are NOT significant. We fail to reject H₀.
• Simple memory: p < 0.05 = "Proven!" p > 0.05 = "Not proven, maybe by chance."

Confidence Intervals:

• A range of values within which the true population value likely falls.
• 95% Confidence Interval = If you repeat the study 100 times, 95 times the true value will be in this range.
• Example: "The average score improvement is between 8.5 and 12.5 points (95% CI)."

1. t-test (For comparing TWO groups)

2. ANOVA (For comparing THREE or MORE groups)

3. Chi-Square Test (For categorical data — Yes/No, Pass/Fail)

Qualitative Data Analysis is the process of examining non-numerical data (interview transcripts, open-ended survey responses, observation notes, images) to find patterns, themes, and meanings.

Simple words: You read through text data, find common ideas, and group them into themes to understand what people are saying.

Step 1: Open Coding (Initial Coding)

• Read through ALL your data carefully.
• Assign a short label (code) to each meaningful sentence or paragraph.
• Don't worry about grouping yet — just label everything.
• Example: Student says: "The app was confusing at first but after a week I loved how it showed my weak topics."
  Codes: "confusing at first", "loved weak topics", "took one week to learn"

Step 2: Axial Coding (Grouping Codes into Categories)

• Look at all your codes and find which ones are similar.
• Group similar codes together into broader categories.
• Give each category a name that describes the group.
• Example:
  Codes: "confusing at first", "hard to navigate", "didn't understand buttons"
  ↓ Group into Category: "Usability Issues"
  
  Codes: "loved weak topics", "recommendations helped", "personalized practice"
  ↓ Group into Category: "Perceived Value"

Step 3: Selective Coding (Identifying Core Theme)

• Look at all your categories and find ONE central theme that connects everything.
• This is the main story or main finding of your research.
• Example: Core Theme = "Students accept technology when perceived benefits outweigh initial learning difficulties."
  This theme connects both "Usability Issues" and "Perceived Value" categories.

Most research theses follow the IMRaD structure: Introduction, Methods, Results, and Discussion.

• Title Page: Title of research, your name, institution, supervisor name, date, degree statement.
• Abstract: 250-300 word summary of ENTIRE research. Includes: problem, methods, key findings, conclusion. Write this LAST.
• Chapter 1 — Introduction: Background of the problem, problem statement, research questions, objectives, significance of study, scope and limitations.
• Chapter 2 — Literature Review: Summary of previous research on your topic, theoretical framework, identification of research gap (what is missing that your study will fill).
• Chapter 3 — Methodology: Research design (experimental/survey/etc.), population and sample, data collection instruments (surveys, tests), procedures step by step, data analysis plan.
• Chapter 4 — Results: Present your findings with tables, charts, graphs. NO interpretation here — just the facts. Show statistical results (t-test, ANOVA, etc.).
• Chapter 5 — Discussion: Interpret what your results mean. Compare with previous research from Chapter 2. Explain unexpected findings. Discuss implications.
• Chapter 6 — Conclusion: Summary of key findings, limitations of your study, recommendations for practice, suggestions for future research.
• References: Complete list of ALL sources cited in your thesis. Follow APA/MLA/IEEE style consistently.
• Appendices: Survey questionnaire, interview questions, consent forms, raw data tables, additional materials.

APA Style (American Psychological Association) — 7th Edition

• Used for: Education, Psychology, Social Sciences, Business
• In-text: (Author, Year) → (Kumar, 2024)
• Reference: Author, A. A. (Year). Title of book. Publisher.
• Example - Book: Kumar, R. (2024). Research Methodology (5th ed.). SAGE Publications.
• Example - Journal: Sharma, P., & Gupta, R. (2023). Adaptive learning in colleges. Journal of Educational Technology, 18(2), 112-128.

MLA Style (Modern Language Association) — 9th Edition

• Used for: Humanities, Literature, Arts
• In-text: (Author page) → (Kumar 45)
• Reference: Author, First Name. Title of Book. Publisher, Year.
• Example: Kumar, Ranjit. Research Methodology. SAGE Publications, 2024.

IEEE Style (Institute of Electrical and Electronics Engineers)

• Used for: Computer Science, Engineering, Technology
• In-text: [Number] → [1]
• Reference: [1] A. Author, "Title of article," Journal, vol., no., pp., year.
• Example: [1] R. Kumar and P. Sharma, "Adaptive learning systems," IEEE Trans. on Education, vol. 45, no. 3, pp. 234-245, 2024.

Clarity: Write simply so anyone can understand. Avoid fancy words. Short sentences are better.

Objectivity: Present facts, not opinions. Don't say "The results were amazing" — say "The results showed a 15% improvement."

Precision: Be specific. Say "42.5% of students" not "many students."

Active Voice: "We conducted the survey" (active) NOT "The survey was conducted" (passive).

Avoiding Plagiarism (Very Important — Can Get You Expelled):

• Plagiarism = Using someone else's words or ideas without giving credit.
• Always use quotation marks ("...") for direct quotes AND cite the source.
• When paraphrasing (rewriting in your own words), you still MUST cite the source.
• Cite everything that is not common knowledge or your own original idea.
• Use plagiarism checker tools (Turnitin, Grammarly) before submission.
• Golden Rule: When in doubt, cite it!

Tables: Best for exact numbers. Use when reader needs precise values.

Bar Charts: Best for comparing categories. Good for showing frequencies, percentages across groups.

Line Graphs: Best for showing trends over time (e.g., scores across weeks 1-8).

Pie Charts: Best for showing parts of a whole. Limit to 5-6 categories maximum.

Scatter Plots: Best for showing relationship between two variables (e.g., study hours vs exam score).

Ethics = Doing the right thing in research. Unethical research can get you expelled, fired, or sued.

• 1. Informed Consent: Participants must KNOW what they are agreeing to. They must sign a consent form. They can withdraw ANYTIME without penalty.
• 2. Confidentiality & Privacy: Protect participant identities. Use codes (Student #47) not names. Store data securely (password-protected). Don't share identifiable information.
• 3. Avoid Harm: No physical harm. No psychological harm (don't embarrass, stress, or upset participants). If any risk exists, you must tell participants beforehand.
• 4. No Fabrication or Falsification (This is serious fraud): NEVER make up data. NEVER change results to fit your hypothesis. This will destroy your career if discovered.
• 5. No Plagiarism: Never copy others' work without credit. Cite all sources.
• 6. No Conflict of Interest: Disclose any funding sources. Disclose any personal relationships that could bias your research.
• 7. Ethics Committee Approval (IRB): Most universities require you to get approval from the Institutional Review Board BEFORE starting research with human participants.