Research methodology, step by step guide for beginners 3rd edition

This is the book for anyone facing their first research project — students, professionals, or curious readers who want method without jargon. Kumar's step-by-step structure means you can read it cover to cover or dip into the chapter you're stuck on. It turns a daunting assignment into a checklist you can actually follow.

1. 1Chapters 1–2

   Research is a process, not a lightning bolt

   Kumar quotes Kerlinger's (1986) definition of scientific research: 'Scientific research is systematic, controlled, empirical, and critical investigation of hypothetical propositions about the presumed relations among natural phenomena.' He presents the research process as an eight-step model: formulating a research problem, conceptualizing a research design, constructing an instrument for data collection, selecting a sample, writing a research proposal, collecting data, processing and displaying data, and writing a research report. He emphasizes that the steps are not always linear and may require going back and forth.

   The idea

   The eight steps give you a map, but the territory is messier than the map suggests. Kumar's insistence that you'll move back and forth between steps is the honest part — most methodology books pretend research flows neatly forward. It doesn't. You'll refine your problem after designing your instrument, and redesign your instrument after a pilot. The model is a checklist, not a conveyor belt.

   Try this: Write out the eight steps on one page and mark where you currently are — then note which step you'll likely need to revisit.

2. 2Chapter 4

   Your research problem is a question, not a topic

   A research problem should be expressed as a question and must be feasible, clear, significant, and ethical.

   The idea

   Turning a vague interest into a question is the hardest step, and the one beginners skip. The four tests are a filter: if your question isn't feasible with your resources, isn't clear enough to answer with data, isn't significant to anyone, or can't be asked ethically, it needs reworking before you touch a design or a sample.

   Try this: Rewrite your topic as a single question, then score it against the four tests — feasible, clear, significant, ethical — and fix the weakest one.

3. 3Chapter 5

   Name your variables before you name your sample

   An example given is a study on the impact of a health education program (independent variable) on the knowledge of mothers (dependent variable) about child nutrition. The setting is a community health project. Variables are classified into independent, dependent, extraneous, and intervening variables. The independent variable is the cause, and the dependent variable is the effect.

   The idea

   Getting the cause-and-effect labels right sounds trivial until you try to design an instrument or choose a sample — then everything hinges on knowing which variable you're manipulating and which you're measuring. Extraneous and intervening variables are the sneaky ones: they explain why two studies of the same programme get different results.

   Try this: For your own project, write one sentence: 'The effect of [independent] on [dependent], controlling for [extraneous].'

4. 4Chapter 6

   A hypothesis is a prediction you can test

   A hypothesis is defined as 'a tentative prediction or explanation of the relationship between two or more variables.' The book distinguishes between a research hypothesis and a null hypothesis.

   The idea

   The null hypothesis trips up every beginner: you don't set out to prove your idea, you set out to reject the claim that nothing is going on. Framing your prediction as a testable statement about named variables is what separates a research project from an opinion piece.

   Try this: Write your research hypothesis in one sentence, then write its null version alongside it.

5. 5Chapters 7–8

   Your design is the blueprint, not the building

   'The research design is the conceptual structure within which research is conducted; it constitutes the blueprint for the collection, measurement and analysis of data.A longitudinal study example: a researcher follows a cohort of 200 nursing students over three years to measure changes in their attitudes toward community care, with data collected at the start of year 1, year 2, and year 3.

   The idea

   Three sorting axes mean you can describe almost any study in a phrase — 'cross-sectional experimental' or 'longitudinal non-experimental' — and know instantly what it can and can't tell you. The nursing cohort shows the cost of longitudinal work: three years of follow-up buys you change over time, which a single snapshot never can.

   Try this: Classify your planned study on all three axes in one line, then check whether that design can actually answer your question.

6. 6Chapter 9

   Pick your data collection method to fit the question, not your comfort

   Primary data collection methods covered include observation, the interview, and the questionnaire. The choice depends on the study objectives, resources, and skills of the researcher.

   The idea

   Beginners default to questionnaires because they seem easiest. Kumar's three-factor test — objectives, resources, skills — forces a more honest match. If your objective is to understand behaviour in context, a questionnaire will give you self-report, not behaviour.

   Try this: List your three methods against your objectives and resources, and eliminate the one that fits worst.

7. 7Chapter 10

   Attitude scales turn opinions into numbers

   The Likert scale is described as a method where respondents indicate their agreement or disagreement with a series of statements on a five- or seven-point scale. Summing responses across statements yields a score that can be compared across people or groups.

   The idea

   The Likert scale is the workhorse of attitude measurement because it converts a subjective stance into a number you can analyse. The five- versus seven-point choice matters: more points give finer discrimination but may exceed a respondent's ability to distinguish.

   Try this: Draft three Likert statements for your own study and test whether the five-point response options feel natural to a colleague.

8. 8Chapter 11

   An instrument is only as good as its validity and reliability

   'Validity is the degree to which an instrument measures what it is supposed to measure,' and 'Reliability is the degree to which an instrument consistently measures whatever it is measuring.' An instrument can be reliable without being valid — consistently measuring the wrong thing.

   The idea

   The distinction is easy to state and hard to apply. A bathroom scale that's always five kilos heavy is reliable but not valid. Kumar's pairing of the two definitions in one chapter signals that you must check both before trusting any data your instrument produces.

   Try this: For your main instrument, write one sentence each: what it's supposed to measure, and how you'll check it measures it consistently.

9. 9Chapter 12

   Sampling: give everyone a known, non-zero chance

   Probability sampling designs include simple random sampling, stratified random sampling, and cluster sampling. In probability sampling, each element in the population has a known, non-zero chance of being selected. An example of stratified random sampling: a researcher studying the attitudes of 150 students in a school with 1,000 students divides the population into strata by year level, then selects a proportionate random sample from each stratum.

   The idea

   Stratified sampling earns its extra effort when subgroups matter — year levels in a school, for instance — because it guarantees each group is represented proportionally rather than left to chance. The 'known, non-zero chance' rule is what lets you generalise from sample to population with quantifiable confidence.

   Try this: Identify the subgroups in your population that must be represented, and decide whether stratification is worth the extra setup.

10. 10Chapter 14

    Ethics is a design problem, not a form you sign

    Core ethical principles include informed consent, privacy, confidentiality, and protection from harm. 'It is unethical to collect information without the knowledge of participants, and their expressed willingness and informed consent.' An ethical dilemma example: a researcher studying drug use among teenagers must balance the need for parental consent against the risk that seeking consent will bias the sample by excluding the very teenagers the study aims to reach.

    The idea

    The drug-use example shows ethics as a genuine trade-off, not a checkbox. Requiring parental consent protects minors but can silence the population most at risk. Kumar doesn't resolve the dilemma for you — he shows that ethical design means anticipating these conflicts before you're in the field.

    Try this: List your four ethical principles and write down the one most likely to conflict with your sampling plan.

11. 11Chapter 15

    Clean data before you analyse it

    Data processing involves editing (checking for completeness, accuracy, and uniformity) and coding (assigning numbers to categories to facilitate computer analysis). Skip the editing and your errors propagate silently through every table and statistic downstream.

    The idea

    Editing and coding are unglamorous but they determine whether your analysis means anything. Completeness, accuracy, uniformity — three checks that catch the survey returned half-blank, the date entered as 1978 instead of 1987, and the 'male' coded as M in one batch and 1 in another.

    Try this: Before analysing, run one edit pass on your raw data checking only for completeness — count missing values per variable.

12. 12Chapter 16

    Tables and figures: structure first, decoration second

    Tables should have a title, a stub (row headings), a box head (column headings), and a body. Figures include bar charts, histograms, frequency polygons, and scattergrams. A specific example of a frequency distribution table shows the ages of 50 participants in a study, grouped into class intervals of 20-24, 25-29, 30-34, 35-39, and 40-44 years, with corresponding frequencies of 5, 15, 18, 8, and 4.

    The idea

    The anatomy of a table — stub, box head, body — is the kind of detail that separates a readable report from a confusing one. The age distribution example shows how a well-built table does the analysis for the reader: the peak at 30–34 is visible before anyone computes a mean.

    Try this: Take one table from your draft and check it has all four parts — title, stub, box head, body — before you format it.

13. 13Chapters 13, 17, 18

    The report is the product; the proposal was the promise

    'A research proposal is a plan, a roadmap, that tells your reader how you intend to undertake your study.' The core components listed are the problem, objectives, study design, sampling, data collection, and analysis plan. 'The research report is the final product of your research endeavour.' The standard structure is: introduction, literature review, methodology, results, discussion, and conclusions. The book distinguishes between 'research as a process' and 'research as a product.' The process focuses on the methodology, while the product is the final report.

    The idea

    The proposal and the report are the same skeleton at two moments: one promises, the other delivers. Kumar's process distinction is a useful check — if your report reads as a diary of what you did rather than an argument about what you found, you've confused the two.

    Try this: Draft your report's six section headings now, before your data is analysed, and note what each must deliver.

First-time researchers — undergraduate and postgraduate students, professionals asked to conduct a study, anyone facing a dissertation or a workplace research task. Also useful as a reference to dip into chapter by chapter when stuck on a specific step.

FAQ