Research Methodology

Research report

A research report is a formal, written document that presents the problem, methods, findings and conclusions of a research study, so that readers can understand, evaluate and use or replicate the work. Examples are a thesis, dissertation, journal article and project report.

Format of an academic research report

1. Preliminary pages: title page, certificate/approval, declaration, acknowledgement, abstract, table of contents, lists of tables and figures, abbreviations.
2. Main body:
   • Introduction: background, statement of the problem, objectives, research questions/hypotheses, significance, scope and limitations.
   • Literature review: earlier work and theoretical/conceptual framework.
   • Research methodology: research design, population and sample, data collection tools, data analysis methods, reliability, validity and ethical considerations.
   • Data presentation, analysis and findings (tables, charts, tests).
   • Discussion of the results against the literature.
   • Summary, conclusion and recommendations, with scope for further research.
3. End matter: references/bibliography (APA or another consistent style), appendices (questionnaire, raw data, etc.) and index.

The report should be clear, objective, logically organised and consistently formatted.

Review of literature

A review of literature is a systematic, critical examination of existing scholarly sources (books, journal articles, theses, reports, conference papers, reliable websites) relevant to the research topic. It summarises, evaluates and synthesises what is known, identifies the gaps and shows how the present study fits in. It is more than a list of summaries: it compares authors, shows agreements and contradictions, and judges the quality of evidence.

Importance

1. Identifies the research gap and helps to define the problem clearly.
2. Avoids duplication of earlier work.
3. Provides the theoretical and conceptual framework for the study.
4. Helps to formulate objectives, questions and hypotheses.
5. Guides the choice of methodology: research design, variables, tools, sampling and statistical techniques that others used, and what weaknesses to avoid.
6. Broadens the researcher's knowledge of key concepts, terms and authors in the field.
7. Gives a basis for comparing findings and for interpreting results in the discussion.
8. Justifies the need and significance of the study.
9. Saves time, effort and resources by building on earlier experience.
10. Establishes credibility by acknowledging sources through proper citation and so avoiding plagiarism.
11. Keeps the research current by showing recent developments and trends.

Review of literature

A literature review is a systematic, critical study and summary of existing published and unpublished work (books, journals, theses, reports, conference papers) related to the research topic. It evaluates and synthesises previous findings, theories and methods, and shows the gap that the present study fills.

Importance of a literature review

1. Identifies the research gap and helps to define and refine the problem.
2. Avoids duplication of previous studies.
3. Provides a theoretical and conceptual framework for the study.
4. Helps to develop research questions, objectives and hypotheses.
5. Guides the methodology: designs, variables, instruments, sampling and analysis techniques used by others, and their limitations.
6. Builds the researcher's knowledge of the field, its concepts, terms and key authors.
7. Provides a benchmark to compare and interpret one's own findings.
8. Justifies the significance and originality of the study.
9. Saves time and avoids past mistakes.
10. Shows scholarship and credibility, acknowledging sources through proper citation and avoiding plagiarism.
11. Keeps the study up to date with current trends.

Format of an academic research report

A. Preliminary pages

• Title page (title, author, degree, institution, supervisor, date)
• Certificate/approval sheet and declaration
• Acknowledgement
• Abstract
• Table of contents, list of tables, list of figures, list of abbreviations

B. Main body

1. Chapter 1, Introduction: background of the study, statement of the problem, research questions, objectives, hypotheses, significance, scope and limitations, organisation of the report.
2. Chapter 2, Literature review: theoretical review, empirical review, research gap, conceptual framework.
3. Chapter 3, Research methodology: research design, population, sample and sampling technique, sources and tools of data collection, validity and reliability, methods of data analysis, ethical considerations.
4. Chapter 4, Data presentation, analysis and interpretation: tables, charts, statistical tests, results and discussion.
5. Chapter 5, Summary, conclusion and recommendations: main findings, conclusions, implications and suggestions for future research.

C. End matter

• References/bibliography (in APA or another consistent style)
• Appendices (questionnaire, interview guide, raw data, additional tables)
• Index (if needed)

Definitions

• Research problem: a specific issue, gap, difficulty or contradiction in knowledge that the researcher wants to investigate. Example: low pass rates in mathematics in community schools.
• Research question: the problem stated as a precise question the study will answer. Example: What factors affect the mathematics pass rate in community schools?
• Research objective: a statement of what the study intends to achieve, written with action verbs (to identify, to measure, to compare). Example: To identify the relationship between teaching hours and pass rate.
• Research hypothesis: a tentative, testable statement predicting the relationship between variables. Example: H1: Schools with more teaching hours have higher pass rates. The null hypothesis H0 states there is no relationship.

How they are interrelated

1. The problem is the starting point and justifies the study.
2. It is narrowed down into one or more research questions, which define the scope.
3. Each question is turned into an objective, which states the aim and guides design and data collection.
4. Where the question concerns a relationship, a hypothesis gives a testable answer to it, which is tested using data.
5. The findings answer the questions, meet the objectives, support or reject the hypotheses, and so help solve the problem.

They must be consistent: every objective should relate to a question, and every hypothesis should relate to an objective. The chain is: problem, question, objective, hypothesis, data, conclusion.

Citation and referencing formats

APA (American Psychological Association), MLA (Modern Language Association), Chicago/Turabian, Harvard, IEEE, Vancouver, AMA, and ACM.

APA format (7th edition)

In-text: (Author, Year) or Author (Year). The reference list is alphabetical by author, with hanging indent.

Book: Author, A. A. (Year). Title of book in italics (Edition). Publisher.

Example 1: Kothari, C. R. (2019). Research methodology: Methods and techniques (4th ed.). New Age International.

Report: Author/Organisation. (Year). Title of report (Report No. xx). Publisher.

Example 2: Ministry of Education, Science and Technology. (2022). Flash I report 2078/79. Government of Nepal.

Journal article: Author, A. A., & Author, B. B. (Year). Title of article. Title of Journal in Italics, volume(issue), pages. https://doi.org/xxxx

Example 3: Bhattarai, P., & Sharma, R. (2021). Use of e-learning in Nepali universities. Journal of Education Research, 11(2), 45-60. https://doi.org/10.3126/jer.v11i2.0000

Key points: surname followed by initials; only the first word of the title (and proper nouns) capitalised for books and articles; the journal name and volume in italics; include DOI where available.

Primary data is original data collected first-hand by the researcher for the specific purpose of the study, for example through a survey, interview or experiment. It is fresh, specific and reliable, but costly and time-consuming (in contrast to secondary data from existing sources).

Methods of collecting primary data

1. Observation: systematically watching and recording behaviour or events. It can be participant or non-participant, structured or unstructured, and direct or indirect. It records actual behaviour, but is subject to observer bias and cannot capture opinions.
2. Interview: face-to-face or telephone question-answer sessions. Types: structured, semi-structured and unstructured; individual or focus group. It gives depth and clarification, but is costly and may suffer from interviewer bias.
3. Questionnaire: a set of written questions answered by respondents, mailed, handed out or sent online. It is economical and reaches many people, but has low response rates and no scope for probing.
4. Schedule: a form filled in by the enumerator, who asks the questions. It suits less literate respondents and gives a higher response rate, but costs more.
5. Survey (sample survey/census): collecting data from a sample or the whole population using the above tools.
6. Experiment: manipulating an independent variable under controlled conditions and measuring its effect on the dependent variable; it supports cause-and-effect conclusions.
7. Focus group discussion: a moderated group discussion of 6-10 participants on a topic.
8. Case study, projective techniques and pilot or field tests are further methods in qualitative work.

The choice depends on the objectives, the nature of the population, time, cost and accuracy needed.

Stratified sampling is a probability sampling method in which the population is first divided into homogeneous, non-overlapping groups called strata (by gender, age, income, faculty and so on), and then a random sample is drawn from each stratum. The allocation can be proportionate (sample size proportional to the stratum size) or disproportionate/equal.

Example: for a student survey in a university with 60% BIT and 40% BBA students, draw 60% of the sample from BIT and 40% from BBA, at random.

When it is more suitable than simple random sampling (SRS)

• The population is heterogeneous and splits naturally into groups which are internally similar (homogeneous within strata, different between strata).
• Representation of every subgroup is needed, including small minority groups that SRS might miss by chance.
• Estimates for each subgroup are wanted separately.
• Higher precision: with the same sample size, stratification reduces the sampling error (variance) compared with SRS.
• A sampling frame for each stratum is available, or administration differs between strata.

If the population is homogeneous, SRS is simpler and as good.

Reliability is the consistency of a measuring instrument: the degree to which it gives the same results when repeated under the same conditions, i.e. is free from random error. A reliable scale gives stable and dependable scores (a reliable scale is not necessarily valid).

Methods of estimating reliability

1. Test-retest method: the same test is given to the same group twice after a time gap, and the two sets of scores are correlated. A high correlation means high stability. Weakness: memory and practice effects, and changes between occasions.
2. Parallel (equivalent) forms method: two equivalent versions of the test are given to the same group, and the scores correlated. It avoids memory effects, but constructing truly equivalent forms is hard.
3. Split-half method: the items are split into two halves (odd-even), and the half-scores are correlated. The Spearman-Brown formula corrects for length: $r_{full}=\frac{2r_{half}}{1+r_{half}}$
4. Internal consistency (Cronbach's alpha, Kuder-Richardson KR-20): measures how closely items relate to each other. An alpha of 0.7 or more is generally acceptable.
5. Inter-rater reliability: the agreement between different observers or judges scoring the same thing (for example Cohen's kappa).

Sampling error vs non-sampling error

Case study

A case study is an in-depth, intensive investigation of a single unit (a person, family, organisation, community, event or programme) within its real-life context, using multiple sources of evidence such as interviews, observation, documents and records. It aims to understand the complex "how" and "why" of a phenomenon and is common in social science, management and education research.

Limitations

• Lack of generalisability: the findings from one or few cases cannot be generalised to a wider population.
• Subjectivity and researcher bias in selecting the case and interpreting the data.
• Time-consuming and costly, as intensive data collection is needed.
• Lack of rigour and control: no control group or random selection; difficult to establish cause and effect.
• Difficulty of replication, so results are hard to verify.
• Depends on the quality of available records and respondents' memory.
• Large volumes of qualitative data are hard to analyse and summarise.
• Ethical issues of confidentiality, since the case is easily identifiable.

Research hypothesis

A hypothesis is a tentative, testable statement about the relationship between two or more variables, or about a population characteristic, which is accepted or rejected after testing with data. Types: null hypothesis ($H_0$, no effect or no relationship) and alternative hypothesis ($H_1$ or $H_a$, there is an effect). Example: $H_0$: there is no difference in the marks of students taught by method A and method B; $H_1$: there is a difference.

A hypothesis must be clear, specific, testable, consistent with known facts and stated in simple terms.

Sources of hypotheses

1. Existing theories: deducing a prediction from a theory.
2. Previous research and literature review: gaps and conflicting findings suggest new hypotheses.
3. Observation and personal experience: noticing patterns in everyday life or work.
4. Analogy: relationships found in one field applied to another.
5. Intuition and creative thinking of the researcher.
6. Culture and society: beliefs, customs and general social patterns.
7. Exploratory or pilot studies and expert discussion.
8. Data from earlier studies and reports, including government records.

Importance of literature reviews

A literature review is a critical survey of the books, journal articles, reports and other sources relevant to a research topic. Its importance:

1. Identifies the research gap by showing what is already known and what is not, so as to define the problem.
2. Avoids duplication of work already done.
3. Gives the theoretical background and builds the conceptual framework of the study.
4. Helps to formulate research questions, objectives and hypotheses.
5. Guides the methodology by showing methods, tools, variables and sampling used by others, and their strengths and weaknesses.
6. Develops the researcher's knowledge of key concepts, terminology and leading scholars.
7. Provides a base for comparison: findings can be discussed against earlier results.
8. Justifies the study and shows its significance and originality.
9. Prevents past mistakes and saves time and resources.
10. Supports arguments with credible evidence and gives proper credit through citations.

Graphical methods of data presentation

Graphs present data visually so that patterns, comparisons and trends are easily understood.

1. Bar chart (simple, multiple, component/stacked): rectangular bars of equal width whose length is proportional to the value; used for categorical or discrete data.
2. Histogram: adjacent rectangles whose area/height represents the frequency of class intervals of continuous data; no gaps between bars.
3. Frequency polygon: a line graph joining the mid-points of the tops of histogram bars; useful for comparing distributions.
4. Frequency curve: a smoothed polygon showing the shape of the distribution.
5. Ogive (cumulative frequency curve): plot of cumulative frequencies against class limits (less-than and more-than ogives); used to find the median and quartiles.
6. Pie chart: a circle divided into sectors with angles proportional to the share of each category: $\text{angle}=\frac{\text{category value}}{\text{total}}\times 360^\circ$
7. Line graph (time series): shows changes of a variable over time.
8. Scatter diagram: points for two variables to show correlation.
9. Pictogram, box plot, stem-and-leaf plot, Lorenz curve and maps (cartograms) for special purposes.

(a) Sampling frame

A sampling frame is the complete list or source of all the units of the target population from which a sample is drawn, for example the electoral roll, a student register or a telephone directory. A good frame is complete, accurate, up to date, free from duplicates and contains each unit once. Errors in the frame (missing or extra units) cause coverage error and bias in results. It is needed especially for probability sampling such as simple random, systematic and stratified sampling.

(b) Non-probability sampling

Here the units are selected by the researcher's judgement or convenience rather than random chance, so not every unit has a known chance of selection. Sampling error cannot be measured, and results cannot be generalised with confidence. It is used when a frame is lacking, or in exploratory and qualitative work, and it is cheap and quick.

Types:

• Convenience sampling: nearest or easily available people.
• Judgemental (purposive) sampling: selected by the researcher's judgement of who is typical or knowledgeable.
• Quota sampling: fixed quotas of defined categories filled non-randomly.
• Snowball sampling: existing participants refer others (useful for hidden populations).

Research design is the overall plan or blueprint for collecting, measuring and analysing data so as to answer the research questions. The main types are as follows.

1. Exploratory research design: used when the problem is vague or little studied. It aims to gain insight, discover ideas and formulate hypotheses. Methods: literature search, expert interviews, focus groups, case studies, pilot studies. It is flexible and usually qualitative.

2. Descriptive research design: describes the characteristics of a population or phenomenon, answering who, what, where, when and how (not why). Examples: surveys, fact-finding, census, correlational and case studies. It needs a clear problem statement and a structured plan, and it often uses large samples.

3. Explanatory (analytical/causal) research design: studies cause-and-effect relationships and explains why something happens by testing hypotheses.

4. Experimental research design: the researcher manipulates one or more independent variables under controlled conditions and observes the effect on the dependent variable. Types:

   • Pre-experimental (one-shot case study, one-group pretest-posttest).
   • True experimental (with random assignment and control group: pretest-posttest control group, Solomon four-group).
   • Quasi-experimental (no random assignment; for example non-equivalent group).

5. Longitudinal design: the same subjects studied over a long period (panel, cohort, trend studies).

6. Cross-sectional design: data collected from different subjects at a single point in time.

7. Case study design: an in-depth study of one or few units.

8. Historical research design: past events investigated using records and documents.

9. Comparative design: compares two or more groups, cases or countries.

The choice depends on the objectives, the state of existing knowledge, time and resources.

Research questionnaire

A questionnaire is a structured set of written questions, arranged in a logical order, given to respondents to answer in writing so as to collect primary data for a research study. It may contain open-ended questions (free answers), closed-ended questions (fixed options, rating scales, yes/no) or both, and can be delivered by post, in person, or online (Google Forms and so on).

Requisites of a good questionnaire

1. Clear objectives: every question should serve a research objective.
2. Simple, clear language, with no technical jargon, ambiguity or double-barrelled questions (two questions in one).
3. Short and well organised: reasonable length, logical order from easy and general to difficult and sensitive questions.
4. Neutral wording: no leading or loaded questions that suggest an answer.
5. Specific and precise questions that respondents can answer (they must have the information and be willing).
6. Proper types of questions and adequate, mutually exclusive and exhaustive response options.
7. Clear instructions and a polite introductory covering letter stating the purpose and assuring confidentiality.
8. Attractive, neat layout, with sufficient space and good formatting.
9. Avoid sensitive or embarrassing questions where possible, or ask them with tact.
10. Pre-testing (pilot study) to detect errors, check reliability and validity, and improve it.
11. Ease of tabulation and analysis, with proper coding.
12. Consistency checks (cross-check questions) and an appropriate language for the respondents.

Interview method

An interview is a method of collecting primary data by direct verbal question and answer between the interviewer and the respondent. Types:

1. Structured (standardised) interview: fixed questions in a fixed order, asked in the same way to everyone. It is easy to compare and analyse, but not flexible.
2. Unstructured (non-directive) interview: no fixed questions; the conversation is open and flexible with only a topic guide. It gives depth, but is hard to analyse and compare.
3. Semi-structured interview: prepared key questions, with freedom to probe and add questions.
4. Personal (face-to-face) interview: gives rapport and observation of non-verbal cues, but is costly.
5. Telephone interview: quick and cheap and wide in reach, but short, with no visual cues.
6. Focus group interview: a moderator leads a group of about 6-10 people in a discussion, bringing out group opinions and interaction.
7. In-depth and clinical interview: a long, probing session on feelings and motives (clinical interviews are used in psychology).
8. Panel (repeated) interview: the same respondents interviewed repeatedly over time.
9. Online/video interviews (Zoom and similar) are now also used.

Levels (scales) of measurement

Stevens classified measurement into four levels, in increasing order of power:

1. Nominal scale: numbers or labels only name categories, with no order. Example: gender, religion, blood group, district. Permissible statistics: frequency, mode, chi-square.
2. Ordinal scale: categories have a meaningful order or rank, but the gaps between ranks are not equal or known. Example: class rank, satisfaction (low, medium, high), Likert items. Statistics: median, percentiles, rank correlation.
3. Interval scale: ordered with equal intervals between values, but no true (absolute) zero, so ratios are meaningless. Example: temperature in Celsius, IQ score, calendar years. Statistics: mean, standard deviation, Pearson correlation, t-test.
4. Ratio scale: has equal intervals and a true zero, so ratios make sense. Example: height, weight, income, age, distance. All statistical measures are permissible, including geometric mean and coefficient of variation.

Each higher level has all the properties of the lower ones, and data can be converted down (ratio to ordinal) but not up.

Cluster sampling

In cluster sampling the population is divided into naturally occurring groups called clusters (villages, wards, schools, households), a random sample of clusters is chosen, and then all units in the selected clusters (one-stage) or a random sample of units within them (two-stage) are studied. The clusters should be heterogeneous inside (each a small image of the population) and similar to each other.

Example: to survey Kathmandu secondary students, randomly select 10 schools and survey the students there.

When it is more suitable than simple random sampling (SRS)

• The population is large and geographically scattered, so listing and visiting randomly spread individuals is costly and slow.
• No complete sampling frame of individual units exists, but a list of clusters does.
• Cost and time must be reduced: travel and administration are concentrated in few areas.
• Units within clusters are diverse (heterogeneous), and clusters resemble each other, so precision is acceptable.
• Practical and administrative convenience matter more than the loss of a little precision.

Its drawback is a higher sampling error than SRS for the same sample size, if units within a cluster are alike.

Inductive vs deductive theory

Many studies use both, in a cycle.

Identification and formulation of good research questions

Methods of identification (sources)

1. Reviewing literature: gaps, contradictions and suggestions for future research in earlier studies.
2. Practical experience and observation of problems at work or in society.
3. Discussion with experts, supervisors, colleagues and practitioners.
4. Existing theories: testing or extending them in a new context.
5. Government and organisational reports, media and current issues.
6. Brainstorming and exploratory/pilot studies, and personal interest.

Formulating a good question

1. Start with a broad topic and narrow it to a specific, manageable problem.
2. State it as a clear question, in simple, unambiguous language.
3. Make it focused and researchable with available data, time and resources.
4. Make it specific and measurable: variables, population, place and period are clear.
5. Ensure it is relevant and significant, original, and of interest to the field.
6. Ensure it is feasible and ethical.
7. Make it open to analysis, not answered by a simple yes or no.
8. Check it with the literature and supervisor, and refine it.

Example: "Does use of mobile banking affect customer satisfaction in Kathmandu commercial banks?"

Descriptive vs inferential statistics

Types of research proposal

A research proposal is a written plan that sets out what is to be studied, why, and how, in order to get approval or funding. Its main types are:

1. Academic (student) proposal: prepared by a student for a thesis or dissertation, to be approved by the supervisor or department.
2. Solicited proposal: written in response to a formal request for proposals (RFP) or call, by an agency or funder with set requirements.
3. Unsolicited proposal: sent on the researcher's own initiative to an agency or organisation that did not ask for it, to persuade it to fund or accept the idea.
4. Informal (or internal) proposal: a short proposal, often within an organisation, usually a few pages, for small projects.
5. Formal proposal: a detailed, structured document with title page, abstract, introduction, literature review, objectives, methodology, timeline, budget and references.
6. Renewal or continuation proposal: requests continued support for an ongoing project.
7. Supplemental proposal: requests additional funds for an existing project.

By approach, proposals may also be quantitative, qualitative or mixed-method.

(a) Plagiarism in research

Plagiarism is presenting another person's words, ideas, data or work as one's own without proper acknowledgement. It is academic dishonesty and a serious ethical offence.

Forms: copying text directly, paraphrasing without citation, self-plagiarism (reusing one's own earlier work), patchwork or mosaic copying, buying or copying a whole paper, and fabricating or misattributing sources.

Consequences include rejection or withdrawal of the paper, cancellation of a degree, loss of reputation and legal action.

Prevention: cite every source, use quotation marks for exact words, paraphrase properly, keep a reference list, use plagiarism-checking software (Turnitin, Urkund) and follow the institution's policy.

(b) Stratified random sampling

The population is divided into homogeneous groups (strata) by a characteristic such as gender, age or faculty, and a simple random sample is drawn from each stratum, usually in proportion to its size. It ensures representation of every subgroup, reduces sampling error compared with simple random sampling, and permits subgroup analysis. It needs prior information on the strata and a complete frame.

Example: a sample of 100 students from 600 BIT and 400 BBA students takes 60 from BIT and 40 from BBA at random.

Research design

A research design is the overall plan or blueprint that specifies how the research will be conducted: the strategy for selecting subjects, collecting data and analysing it in order to answer the research questions and test the hypotheses accurately, efficiently and economically. It links the research problem to the data and the conclusions.

Importance

1. Gives direction and structure to the whole study, making it systematic.
2. Saves time, cost and effort by planning in advance and avoiding wasteful activity.
3. Ensures that the data collected is relevant to the objectives and that the right data are gathered.
4. Increases reliability and validity by controlling bias and extraneous variables.
5. Enables a suitable choice of methods, sampling, tools and statistical techniques.
6. Allows the research to be replicated and the findings to be evaluated.
7. Helps to anticipate problems and allocate resources.
8. Supports valid conclusions and generalisation.

Key features of a good research design

• Clear statement of the problem, objectives and hypotheses.
• Objectivity and freedom from bias; reliable and valid measurement.
• Flexibility to deal with different aspects of the problem, where appropriate.
• Appropriateness to the research question (exploratory, descriptive, experimental and so on).
• Adequate control of extraneous variables, and a proper sampling plan.
• Generalisability (external validity) and accuracy of results.
• Economy: feasible within the time, cost and available data.
• Ethical soundness, with consent and confidentiality.
• Specification of the data sources, collection techniques and analysis procedure, in a transparent and replicable way.

Questionnaire

A questionnaire is a structured set of written questions, in a planned order, given to respondents to answer in writing, so as to collect primary data for the research objectives. It can use open-ended, closed-ended (multiple choice, yes/no, rating or Likert scale) or mixed questions, and be delivered by post, personally or online.

Characteristics of a well-structured questionnaire

1. Relevant to the objectives: each question serves a purpose.
2. Simple, clear and unambiguous wording, with no jargon or double-barrelled questions.
3. Logical sequence, from simple and non-sensitive to complex and personal questions; related questions grouped together.
4. Neutral and unbiased: no leading or loaded questions.
5. Reasonable length, to avoid respondent fatigue.
6. Appropriate response options that are exhaustive and mutually exclusive.
7. Clear instructions and a covering letter explaining the purpose and ensuring confidentiality.
8. Neat and attractive layout with adequate space.
9. Answerable questions: respondents should have the knowledge and be willing to respond; sensitive matters handled with tact.
10. Pre-tested through a pilot study, and checked for reliability and validity.
11. Easy to code, tabulate and analyse.

Characteristics of scientific research methods

1. Systematic and logical: follows an orderly series of steps (problem, hypothesis, design, data, analysis, conclusion).
2. Empirical: based on observable and measurable evidence collected from the real world, not on opinion or belief.
3. Objective: free from the researcher's personal bias and values; conclusions come from facts.
4. Verifiable and replicable: other researchers can repeat the study and get similar results.
5. Reliable and valid: measurements are consistent and measure what they claim to measure.
6. Controlled: extraneous variables are controlled or accounted for to establish relationships.
7. Testable hypotheses: statements that can be tested and are open to rejection (falsifiable).
8. Generalisable and predictive: results apply beyond the sample and allow prediction.
9. Cumulative: builds on earlier knowledge and adds to the theory.
10. Ethical, precise and self-correcting: uses exact definitions, reports honestly and is open to revision in light of new evidence.

Research hypothesis

A research hypothesis is a tentative, testable statement about the expected relationship between two or more variables (or about a population parameter), put forward for investigation and to be accepted or rejected on the basis of data. It is usually written as a null hypothesis $H_0$ (no relationship or difference) and an alternative hypothesis $H_1$ (a relationship or difference exists).

Example: $H_1$: Students who use e-learning score higher than those who do not.

Criteria of a good hypothesis

1. Clear and precise in wording, with clearly defined variables.
2. Testable and falsifiable with available methods and data.
3. Specific rather than vague or general.
4. States the expected relationship between variables (direction, where relevant).
5. Consistent with known facts and theory, and based on prior knowledge or literature.
6. Simple, with as few assumptions as possible, and not self-contradictory.
7. Relevant to the research problem and objectives.
8. Within the scope of available techniques, time and resources (feasible).
9. Capable of explaining the facts and giving predictions.

Process of identifying and formulating a research question

1. Select a broad area of interest from experience, courses, current issues or discussions with experts.
2. Review the literature to learn what is known, find gaps, contradictions and unanswered questions.
3. Identify the research problem: a specific gap or difficulty that deserves study.
4. Narrow the topic to a manageable, focused problem in terms of variables, population, place and time.
5. Formulate the question in clear, simple and unambiguous language, as an interrogative sentence. Example: "What is the effect of mobile banking on customer satisfaction in Nepali commercial banks?"
6. Evaluate the question using criteria: clear, focused, researchable, feasible (time, cost, data), significant, original, ethical.
7. Consult supervisors and experts and refine the question.
8. Derive objectives and hypotheses from it, which guide the design.

The process is iterative: the question may be revised as understanding grows.

Reference management software (Zotero, Mendeley)

These tools help researchers collect, organise, cite and share references. Assistance they provide:

1. Collecting references: import papers, books and web pages with one click from browsers, databases and PDFs (metadata and DOI are filled in automatically).
2. Organising a library: folders, tags, notes and search, and storing PDFs with annotations and highlights.
3. Automatic citation and bibliography: plug-ins for MS Word and Google Docs insert in-text citations and generate the reference list in styles such as APA, MLA, IEEE, Harvard or Chicago, and the style can be changed in seconds.
4. Saves time and reduces errors in formatting, and keeps citations consistent, which helps avoid plagiarism.
5. Removes duplicates and keeps data clean.
6. Collaboration: shared groups and libraries for team projects.
7. Cloud sync across devices, with backup and access anywhere.
8. Discovery: Mendeley recommends related papers and has a research network.
9. Export/import to BibTeX and other formats.

Zotero is free and open source; Mendeley, owned by Elsevier, offers limited free storage.

Reliability

Reliability is the consistency or stability of a measuring instrument: the extent to which it gives the same results when repeated under similar conditions.

Example: a weighing machine that shows 60 kg for the same person on repeated weighings is reliable. A test given twice to the same students that produces nearly the same scores (high test-retest correlation) is reliable.

Validity

Validity is the extent to which an instrument measures what it is intended to measure, and the accuracy of the conclusions drawn. Types: content, criterion-related (predictive and concurrent), construct, and face validity.

Example: an arithmetic test containing only English comprehension passages is not a valid measure of mathematical ability; a mathematics exam covering the whole syllabus is valid for measuring mathematical knowledge. A scale that always shows 5 kg too much is reliable (consistent) but not valid (not accurate).

A valid instrument must be reliable, but a reliable instrument is not necessarily valid.

Descriptive vs inferential statistics

Methods of presenting data

1. Frequency table

A table that arranges data into classes or categories and shows the number of observations (frequency) in each; it may also show relative frequency (%) and cumulative frequency. Steps: decide the classes, tally the observations, count the frequencies and add totals. It should have a number, title, clear headings, units and source. It is compact and the basis for the graphs.

Example:

2. Bar chart

Rectangular bars of equal width, with gaps between them, whose heights (or lengths) are proportional to the frequency or value of each category. It can be vertical or horizontal, and simple, multiple (grouped) or component (stacked). It is best for comparing categories of qualitative or discrete data, with the axes labelled and scaled from zero.

3. Pie chart

A circle divided into sectors whose angles are proportional to each category's share of the total: $\text{angle}=\frac{\text{frequency}}{\text{total}}\times 360^\circ$ For the table above, A gets $\frac{10}{40}\times360^\circ=90^\circ$, B gets $180^\circ$ and C gets $90^\circ$. It is best for showing parts of a whole with few categories, and is labelled with percentages.

(a) Stratified sampling

In stratified sampling the population is divided into homogeneous, non-overlapping subgroups called strata (for example by gender, age, faculty or region), and a random sample is drawn from each stratum. Allocation can be proportionate (sample size in each stratum proportional to its size) or disproportionate.

Merits: ensures representation of all subgroups, gives greater precision (smaller sampling error) than simple random sampling for a heterogeneous population, and allows separate estimates for each stratum. Limitations: needs prior information about the population and a complete list for each stratum, and is more complex and costly.

Example: from 600 BIT and 400 BBA students, a sample of 100 takes 60 and 40 respectively, at random.

(b) Plagiarism

Plagiarism is the use of another person's words, ideas, data or creative work without proper acknowledgement, presenting it as one's own. Forms include direct copying, paraphrasing without citation, self-plagiarism, mosaic plagiarism and false citation or fabrication.

It is academic dishonesty with consequences such as rejection of the work, loss of the degree, damaged reputation and legal action for copyright violation. It can be avoided by citing all sources, using quotation marks for exact words, paraphrasing properly, listing references and checking with software such as Turnitin.