Business Research Methods

Probability sampling

Probability sampling is a method in which every unit of the population has a known, non-zero chance of being selected, and selection is by random procedure. It reduces bias, allows sampling error to be estimated, and permits generalisation of the results to the population.

Types of probability sampling

1. Simple random sampling: every unit has an equal chance of selection, using a lottery or a random number table. It is simple and unbiased but needs a complete sampling frame.

2. Systematic sampling: select every $k$th unit after a random start, where the sampling interval is $k = \frac{N}{n}$ For $N = 1000$ and $n = 100$, $k = 10$; if the random start is 4, the sample is the 4th, 14th, 24th units and so on. It is easy but can be biased if the list has a periodic pattern.

3. Stratified random sampling: the population is divided into homogeneous, non-overlapping strata (for example by income or age), and a random sample is drawn from each, either proportionately or disproportionately. It increases precision and ensures every subgroup is represented.

4. Cluster sampling: the population is divided into heterogeneous groups (clusters such as wards or schools); some clusters are randomly selected and all (or some) units within them are studied. It is economical for scattered populations but less precise.

5. Multistage sampling: sampling is done in several stages, for example district, then municipality, then ward, then household, each randomly. It is useful for large national surveys.

6. Area sampling: a form of cluster sampling based on geographical areas.

A research problem is a clearly stated question, difficulty or gap in knowledge that a researcher wants to investigate and resolve through systematic study. It arises from a gap between the existing and the desired situation, and it guides the objectives, hypotheses and design of the research. Example: "Why is employee turnover high in commercial banks of Nepal?"

A theoretical framework is a conceptual model, based on theory and previous literature, that shows the relationships among the independent, dependent, moderating and mediating variables of a study. It explains why the variables are related and provides the basis for developing hypotheses.

A dependent variable is the variable that is affected, explained or predicted by other variables; it is the main outcome the researcher tries to explain. Example: in a study on the effect of advertising on sales, sales is the dependent variable, because it changes in response to advertising (the independent variable).

Qualitative data are non-numerical (words, images, behaviour) that describe meanings and experiences. Main methods of collecting them:

• In-depth interviews: one-to-one, open-ended and probing discussions (structured, semi-structured or unstructured), recorded and transcribed.
• Focus group discussions: a moderator leads 6-10 people in a discussion; group interaction generates ideas and opinions.
• Observation: watching behaviour in its natural setting, as participant or non-participant, with field notes.
• Case study: detailed study of an individual, firm or event using several sources.
• Document and content analysis: reading reports, letters, minutes, media and records.
• Ethnography and diaries/open-ended questionnaires: long immersion in a community, or respondents' own written accounts.
• Projective techniques: word association, sentence completion and role play to reveal hidden feelings.

The data are then coded and analysed thematically. Sampling is usually purposive and the sample is small.

Ethics are the moral principles that guide researchers' conduct. Major ethical issues in business research are:

• Informed consent: participants must be told the purpose, risks and their right to refuse or withdraw.
• Privacy and confidentiality: personal information and company data must be protected, and anonymity maintained.
• Harm to participants: no physical, psychological, social or economic harm.
• Deception: respondents should not be misled about the purpose of the study.
• Honesty and integrity: no fabrication or falsification of data, selective reporting or manipulation of results.
• Plagiarism: all sources must be acknowledged.
• Sponsor-related issues: research should not be biased to suit a client's wishes, and confidentiality of the sponsor must be balanced with the public interest.
• Objectivity and bias: avoiding bias in sampling, questions and interpretation.
• Data storage and misuse: data must be stored safely and used only for the stated purpose.
• Respect for vulnerable groups: extra care with children, the poor and the like.

Following ethical standards raises the credibility of research and protects participants.

Ethical issues in business research can be classified by the party concerned.

1. Ethical issues towards participants

• Informed consent: participation must be voluntary and respondents must know the purpose of the study.
• Privacy and confidentiality: personal information must be protected and anonymity maintained.
• Protection from harm: no physical, psychological or professional harm should come to respondents.
• Deception: respondents should not be misled about the study.

2. Ethical issues towards the client or sponsor

• Confidentiality of the sponsor's information, and honest reporting of findings, even when unfavourable.
• Not accepting a study whose result is pre-determined.

3. Ethical issues of the researcher

• Honesty and objectivity: no fabrication or falsification of data and no selective reporting.
• Plagiarism: acknowledging the work of others through proper citation.
• Competence and integrity in using methods and analysis.

4. Ethical issues towards the society and the research community

• Proper use of results, avoiding misuse that harms society, and sharing methods and findings openly.

In probability sampling every unit of the population has a known, non-zero chance of being selected, so selection is random and results can be generalised and sampling error estimated. The main types are:

1. Simple random sampling Each unit has an equal chance of selection, by lottery method or random number table. Example: selecting 100 customers from a list of 1,000 using random numbers.

• Merit: unbiased, simple in theory. Demerit: needs a complete list; may not represent small subgroups; costly if units are scattered.

2. Systematic sampling Every $k$th unit is selected after a random start, where $k = N/n$. Example: with $N = 1000$ and $n = 100$, $k = 10$; if the start is 7, the sample is 7, 17, 27, ...

• Merit: easy and quick. Demerit: bias if the list has a periodic pattern.

3. Stratified random sampling The population is divided into homogeneous strata (for example, by gender, income or branch) and random samples are drawn from each stratum, proportionally or equally. Example: sampling customers separately from urban and rural branches.

• Merit: ensures representation of every group and increases precision. Demerit: needs prior information about strata.

4. Cluster sampling The population is divided into naturally occurring groups (clusters), some clusters are randomly selected, and all units (or a sample of them) within them are studied. Example: randomly selecting 7 branches out of 30 and surveying all their customers.

• Merit: economical when the population is spread widely. Demerit: higher sampling error than simple random sampling.

5. Multi-stage sampling Sampling is done in several stages, such as district, then municipality, then ward, then household, using random selection at each stage. It is suitable for large national surveys.

Probability sampling gives representativeness and allows statistical inference, but it needs a sampling frame and is more costly and time-consuming than non-probability sampling.

Sampling is the process of selecting a part (sample) of a population so that conclusions about the whole population can be drawn from it. It saves time and cost.

Two probability sampling techniques (each unit has a known, non-zero chance of selection):

(a) Simple random sampling: every unit has an equal chance of selection, using a lottery or random-number table. It is unbiased and simple but needs a complete sampling frame. Example: choosing 100 customers at random from a list of 1,000.

(b) Stratified random sampling: the population is divided into homogeneous groups (strata), such as by gender or income, and a random sample is taken from each, usually in proportion to its size. It ensures representation of every group and increases precision. Example: sampling students from each faculty of a campus.

Ethical concerns in research include:

• Informed consent: participants must know the purpose and agree voluntarily.
• Confidentiality and anonymity of respondents' identity and data.
• No harm (physical, psychological or financial) to participants.
• Honesty: no fabrication, falsification or plagiarism of data or sources.
• Objectivity: avoiding bias and misreporting of findings.
• Proper acknowledgement of sources and sponsors, and no deception.

A theoretical framework is a logically developed and described network of associations among the variables (independent, dependent, moderating and mediating) that are relevant to the research problem. It is based on existing theory and literature and forms the foundation from which hypotheses are derived and tested.

The purposes of a literature survey are:

• To identify what is already known and avoid duplication of work.
• To define and sharpen the research problem and objectives.
• To find gaps in existing knowledge.
• To build the theoretical framework and develop hypotheses.
• To select suitable methodology, tools and variables.
• To compare the study's findings with earlier results.

There are four types of measurement scales:

• Nominal scale: classifies into categories without order (gender, religion).
• Ordinal scale: categories with rank order (first, second, third; low, medium, high).
• Interval scale: equal intervals but no true zero (temperature in Celsius, Likert-type ratings).
• Ratio scale: equal intervals with a true zero (income, age, weight).

Methods of observation include:

• Participant and non-participant observation: the observer joins the group or stays outside it.
• Structured and unstructured observation: with or without a pre-set checklist.
• Direct and indirect observation: watching behaviour as it happens or its records and traces.
• Disguised (covert) and undisguised (overt) observation: subjects are unaware or aware.
• Natural (field) and laboratory observation: real-life setting or a controlled one.
• Mechanical observation using cameras, counters and similar devices.

Descriptive statistics are statistical methods used to organise, summarise and present the main features of a data set, without drawing conclusions beyond it. They include measures of central tendency (mean, median, mode), dispersion (range, standard deviation), and tables, charts and graphs.

The essentials of a good research report are:

• Clear, simple and precise language.
• A logical structure: title, abstract, introduction, methodology, findings, conclusion, references.
• Accuracy and objectivity, supported by evidence and data.
• Clearly stated objectives, methods and limitations.
• Proper use of tables, charts and graphs.
• Practical conclusions and recommendations.
• Correct citation and referencing, and a neat, error-free presentation.

APA format for a journal article:

Author, A. A., & Author, B. B. (Year). Title of the article in sentence case. Title of the Journal in Italic Title Case, volume(issue), page range. https://doi.org/xxxx

Example: Shrestha, R., & Karki, S. (2022). Service quality and customer satisfaction in Nepalese commercial banks. Journal of Business and Management Research, 7(2), 45-60.

Scientific method in natural science involves objectivity, controlled experiments, measurement and replication. Applying it to social science (human behaviour and organisations) faces several difficulties:

• Complexity of human behaviour: people are unpredictable, with attitudes, emotions and motives that vary across individuals and situations, so precise laws are hard to build.
• Difficulty in measurement: concepts such as satisfaction, motivation and loyalty are abstract and hard to measure accurately.
• Lack of control and experimentation: variables cannot be isolated or manipulated in a laboratory, and many factors operate together.
• Problem of replication: social situations change with time and place, so results cannot be exactly repeated.
• Researcher bias and subjectivity: the researcher is part of society and may be influenced by values and beliefs.
• Reactivity (Hawthorne effect): subjects behave differently when they know they are being studied.
• Ethical limits: experiments on people may harm or violate privacy.
• Cost, time and non-cooperation: data collection is expensive, and respondents may refuse or give false answers.
• Generalisation problems: findings from one culture or sample may not apply elsewhere.

Despite these difficulties, careful design, sampling, standard instruments and mixed methods let researchers apply scientific principles reasonably well.

An interview is a face-to-face or mediated conversation in which the researcher asks questions to obtain information from respondents. Main types are as follows.

(a) On the basis of structure

• Structured interview: fixed questions in a fixed order, as in a questionnaire. It is easy to compare and analyse, but inflexible.
• Unstructured interview: open conversation guided only by a topic. It gives depth but is hard to analyse.
• Semi-structured interview: a prepared guide with freedom to probe further.

(b) On the basis of the number of participants

• Individual (personal) interview: one respondent at a time.
• Group interview and focus group discussion: a moderator leads 6-10 participants to discuss a topic, which brings out interaction and diverse views.

(c) On the basis of the medium

• Face-to-face (personal) interview: high response and observation of non-verbal cues, but costly.
• Telephone interview: cheap and fast, but short and lacks rapport.
• Online or video interview: reaches distant respondents at low cost.

(d) Other types

• In-depth interview: long, probing sessions on a person's experiences.
• Clinical interview: to find the causes of feelings and behaviour.
• Panel or expert interview: repeated or specialist respondents.

Interviews are suited to qualitative data, and the transcribed responses are analysed by coding and thematic analysis.

Similarities

• Both are systematic ways of investigating a research problem.
• Both begin with a problem and objectives, review literature, collect and analyse data, and draw conclusions.
• Both need a sound design, sampling and ethical care.
• They can be used together (mixed methods).

Differences

A research design is the blueprint for collecting, measuring and analysing data. Business research commonly uses these designs, chosen according to the problem.

• Exploratory research design: used when the problem is vague. It gathers insight through literature review, expert interviews, focus groups and case studies, and is flexible.
• Descriptive research design: describes characteristics of a population, market or phenomenon (who, what, when, where, how many). Surveys and cross-sectional or longitudinal studies belong here. It is the most widely used in business, for example customer profiles and satisfaction surveys.
• Causal (experimental) research design: tests cause-and-effect, for example the effect of price or advertising on sales, using test marketing and controlled experiments.
• Correlational design: examines relationships among variables, for example between service quality and satisfaction.
• Case study design: an in-depth study of one firm or situation.

Often the study moves from exploratory to descriptive to causal. Descriptive and causal designs are the most applicable to business problems, because managers need to describe markets and know what influences results.

Secondary data are data already collected by others (government reports, company records, journals, the Nepal Rastra Bank). Primary data are collected first-hand by the researcher through surveys, interviews, observation and experiments.

Can secondary data substitute for primary data? Partly, but not always.

When it can substitute

• The required information already exists and is reliable, for example population figures from the Census, or bank deposits from NRB reports.
• Time, money or access is limited.
• The study is historical or trend based, such as the NEPSE index over ten years.
• The research is exploratory and the data are only needed for background.

Illustration: A study on the growth of commercial bank deposits in Nepal can be done entirely using NRB annual reports.

When it cannot

• The data may be outdated, not specific to the research objectives, or measured with different definitions.
• Accuracy and reliability may be doubtful.
• Attitudes, opinions, motivation and buyer behaviour of a particular firm's customers are not available elsewhere, so a survey is needed.

Illustration: Measuring customer satisfaction with a new mobile banking app requires a fresh primary survey.

Conclusion: Secondary data are a cheap, fast substitute where they fit the problem, but they cannot replace primary data when the objective needs specific, current information. The best practice is to use secondary data first and collect primary data to fill the gaps.

Sampling is selecting a part of the population to represent the whole.

Probability sampling

Every unit has a known, non-zero chance of selection, so results can be generalised and sampling error measured.

• Simple random sampling: each unit has an equal chance (lottery, random numbers).
• Systematic sampling: every $k$th unit after a random start, where $k = N/n$.
• Stratified sampling: the population is divided into homogeneous strata and random samples are taken from each.
• Cluster sampling: the population is divided into clusters (such as wards), some are selected at random and all or some units in them are studied.
• Multistage sampling: several stages of random selection (district, then municipality, then household).

Non-probability sampling

Selection is based on the researcher's judgement or convenience; the selection chance is unknown, so generalisation is limited.

• Convenience sampling: whoever is easily available.
• Judgement (purposive) sampling: units chosen on the researcher's judgement.
• Quota sampling: fixed quotas by category filled non-randomly.
• Snowball sampling: respondents refer other respondents.

Probability sampling is more accurate but costly. Non-probability sampling is quick and cheap and is used in exploratory studies.

(a) Major problem identified

Hyatt suffers from brand confusion. Customers cannot differentiate among the Regency Select (business travellers), Place Express (budget) and Grand Hyatt (luxury), so there is no brand loyalty and revenue is threatened. This is worsened by old, poorly maintained buildings, poor service quality ratings, mixed messages and the lack of cooperation from franchise owners. Research problem: "Why are customers unable to differentiate the Hyatt brands, and what factors affect customer loyalty and revenue?"

(b) Theoretical framework

• Dependent variables: brand loyalty (and revenue).
• Independent variables: brand differentiation (clarity), service quality, facility/building condition and maintenance, and franchise owner cooperation.
• Moderating variable (optional): customer segment (business, budget, luxury).

Framework (flow):

Brand differentiation, service quality, facility condition, franchise cooperation $\rightarrow$ Brand loyalty / revenue, with customer segment moderating the relationship.

The logic is that clearer positioning, better service, well-maintained facilities and cooperative franchisees lead to higher loyalty and revenue, and the effect differs among segments.

(c) Hypotheses

• $H_1$: There is a significant positive relationship between brand differentiation and brand loyalty.
• $H_2$: Service quality has a significant positive effect on customer loyalty.
• $H_3$: Facility condition and maintenance have a significant positive effect on customer satisfaction/revenue.
• $H_4$: Franchise owner cooperation has a significant positive effect on brand performance (revenue).
• $H_5$: The effect of service quality on loyalty differs significantly among customer segments.

(d) Most suitable research design

An exploratory design should come first, as the problem is vague and several causes are suspected. It would use interviews with managers and franchisees and focus groups with customers to identify the real issues. This should be followed by a descriptive (cross-sectional survey) design, which is the key design here: a customer survey to measure perception of the three brands, service quality and loyalty, and then correlational/causal analysis to test the hypotheses.

Applied research is research conducted to solve a specific, practical problem and to find solutions that can be put to immediate use. Its findings help decision making in business, industry or government. Example: studying the reasons for falling sales of a company in order to improve its marketing strategy.

Ethnography is a qualitative research method in which the researcher studies a culture, community or group in its natural setting, over a long period, through participant observation, interviews and field notes. The aim is to understand the group's behaviour, values, beliefs and way of life from the members' viewpoint.

A five-point Likert scale measures the degree of agreement with a statement.

Statement: "I am satisfied with the services of my bank."

Grounded theory is a qualitative research approach in which a theory is developed inductively from data systematically collected and analysed (through interviews and observation), rather than testing an existing theory. The researcher uses coding, constant comparison and theoretical sampling until a theory "grounded" in the data emerges.

A telephone interview is a method of collecting data in which the interviewer asks the respondent questions over the telephone. It is quick, inexpensive and reaches geographically scattered respondents, but the interview must be short, there is no visual contact, and people may refuse to respond.

The deductive approach starts with a general theory, from which hypotheses are developed and then tested with empirical data to confirm or reject the theory (general to specific). It is typical of quantitative research and positivism. Example: from the theory that service quality increases satisfaction, test the hypothesis on bank customers.

A moderating variable changes the strength or direction of the relationship between an independent and a dependent variable. Examples:

• In the relationship between training and employee performance, the employee's experience (or age) is a moderator.
• In the relationship between service quality and customer satisfaction, the customer's income level (or gender) is a moderator.

APA referencing of a journal article:

Gautam, T. (2021). Impact of working capital management on profitability of Nepalese manufacturing firms. Nepalese Journal of Management Research, 3(1), 25-38. https://doi.org/10.3126/njmr.v3i1.0000

In-text citation: (Gautam, 2021).

A literature survey is the systematic search and review of existing studies relevant to the research topic. Its steps are:

1. Select and define the topic: identify the research problem and key concepts.
2. Search for literature: use libraries, journals, databases (Google Scholar, JSTOR), reports, books and theses, with keywords.
3. Screen and select sources: choose relevant, recent and credible works.
4. Read and take notes: record the objectives, methods, findings and limitations of each.
5. Analyse and organise: group by theme, variable or chronology; compare agreements and conflicts.
6. Identify the research gap: find what is missing or unresolved.
7. Write the review: present the findings in a critical, coherent way and link them to the present study.
8. Cite and reference: acknowledge all sources in a standard style such as APA.

Measurement scales define how numbers are assigned to characteristics. There are four types.

• Nominal scale: only names or categories; no order and no arithmetic except counting. Example: gender (male/female), marital status, occupation. Statistics: mode, frequency.
• Ordinal scale: categories with a meaningful order, but unequal or unknown gaps. Example: rank in class, satisfaction (low/medium/high). Statistics: median, rank correlation.
• Interval scale: ordered with equal intervals but no true zero. Example: temperature in degrees Celsius, IQ score, a Likert-type score. Statistics: mean, standard deviation.
• Ratio scale: all properties of interval plus a true zero, so ratios are meaningful. Example: income (Rs), age, weight, sales. Statistics: all, including geometric mean and coefficient of variation.

The scales move from lowest (nominal) to highest (ratio) in the amount of information they give.

This statement (attributed to Zora Neale Hurston) means that research is a natural human curiosity made systematic and goal-directed.

• Curiosity: all research begins with questions such as "why" and "how". People wonder why sales fall or why employees leave.
• Formalised: in research the curiosity is organised through a defined problem, objectives, hypotheses, design, sampling, valid instruments, analysis and reporting. Ordinary curiosity is random and casual, but research follows scientific procedure.
• Poking and prying: the researcher probes, observes, interviews and examines data to find out what lies beneath the surface.
• With a purpose: the aim is to add knowledge, solve problems, test theories or help decisions; it is not idle inquisitiveness.

For example, a manager who is simply curious why customers are leaving does nothing systematic. When he designs a survey, collects and analyses data, and recommends retention strategies, it becomes research. Hence the statement is justified: research is curiosity turned into a disciplined, purposeful search for knowledge.

A good management research report should have the following essentials.

• Clear purpose and objectives: the problem and aims are stated precisely.
• Simple, clear language: free of jargon, unambiguous and concise, with correct grammar.
• Logical organisation: preliminary pages (title, certificate, acknowledgement, abstract, contents), main body (introduction, literature review, methodology, analysis and findings, conclusions and recommendations), and end matter (references, appendices).
• Accuracy and objectivity: findings supported by data, not personal opinion.
• Adequate methodology: sampling, data, tools and limitations explained so the study can be replicated.
• Effective presentation of data: well-labelled tables, charts and graphs.
• Practical recommendations: useful and implementable for managers.
• Proper referencing: consistent citation (APA) and avoidance of plagiarism.
• Good appearance: neat layout, uniform formatting and error-free printing.
• Appropriate length and timeliness.

Scientific research

Scientific research is a systematic, controlled, empirical and critical investigation of a problem, guided by theory and hypotheses, in which data are collected and analysed objectively to arrive at reliable and verifiable conclusions. Its features are objectivity, systematic procedure, empirical evidence, replicability and logical reasoning.

Steps of the scientific research process

1. Identifying and defining the problem: choose a topic, review preliminary sources and state the problem and objectives clearly.
2. Review of literature: study previous work to identify gaps, theories and variables.
3. Formulating the theoretical framework and hypotheses: specify the relationships between the variables and state testable hypotheses.
4. Research design: decide the type of study (exploratory, descriptive, causal), the unit of analysis, time horizon and the methods.
5. Sampling design: define the population, sampling frame, technique and sample size.
6. Data collection: construct and pretest the instruments (questionnaire, interview guide), then collect primary or secondary data in the field.
7. Data processing and analysis: edit, code, tabulate and analyse using statistical tools (descriptive and inferential) or thematic analysis.
8. Hypothesis testing and interpretation: accept or reject the hypotheses and explain the results in relation to theory.
9. Report writing and presentation: present the findings, conclusions and recommendations in a standard report, with references.

The process is iterative; findings may send the researcher back to revise earlier steps.

Research design

A research design is the overall plan or blueprint that specifies how the data will be collected, measured and analysed so as to answer the research questions with the least cost, time and error. It covers the type of study, sampling, data collection methods, instruments and analysis plan.

Types of quantitative research designs

1. Descriptive design: describes the characteristics of a population or phenomenon (who, what, where, when, how). The researcher does not manipulate variables.

• Cross-sectional: data are collected once from a sample at one point in time (a customer survey).
• Longitudinal: the same sample or variables are observed over a period (panel or trend studies).

2. Correlational design: measures the degree and direction of association between two or more variables using correlation or regression, without claiming cause. Example: relation between advertising spending and sales.

3. Causal-comparative (ex post facto) design: examines possible causes after the event by comparing groups that already differ, without manipulating the independent variable.

4. Experimental design: the researcher manipulates one or more independent variables and controls others to establish cause and effect. It uses random assignment to treatment and control groups.

• True experimental: randomisation and a control group (pretest-posttest control group design).
• Quasi-experimental: no random assignment (non-equivalent groups).
• Pre-experimental: weak designs such as the one-shot case study.

5. Survey design: a form of descriptive design that collects data from a large sample by questionnaires or structured interviews to measure opinions, attitudes and behaviour.

Choice: descriptive designs suit describing a situation, correlational designs suit finding relationships, and experimental designs suit testing causes. Quantitative designs give numerical, generalisable results and test hypotheses statistically.

Research proposal

A research proposal is a detailed, formal plan submitted before the study begins, which describes what the researcher intends to study, why, and how. It is used to get approval or funding and acts as a roadmap for the research.

Typical contents: title, introduction and background, statement of the problem, objectives, research questions/hypotheses, significance, review of literature, theoretical framework, methodology (design, population and sample, data collection, analysis), limitations, work plan/schedule, budget and references.

Procedure for writing the research report

1. Plan and organise: decide the purpose, audience, scope and outline of the report.
2. Collect and arrange materials: gather the analysed data, tables, notes and references.
3. Prepare the outline of chapters:
   • Preliminary pages: title page, certificate, acknowledgement, abstract, table of contents, lists of tables and figures.
   • Main text: Chapter 1 introduction (background, problem, objectives, significance, limitations); Chapter 2 literature review and framework; Chapter 3 methodology; Chapter 4 data presentation, analysis and findings; Chapter 5 summary, conclusion and recommendations.
   • End matter: references (APA) and appendices (questionnaire, extra tables).
4. Write the first draft: in clear, objective, simple language and in the proper tense and person.
5. Present data effectively: use well-titled tables, charts and graphs with interpretation.
6. Revise and edit: check logic, accuracy, grammar, consistency and citations; avoid plagiarism.
7. Proofread and finalise: format, paginate, print and bind or submit the final report.
8. Present or submit: share with the supervisor, sponsor or audience and take feedback.

(a) Five probable research questions

1. What is the level of service quality performance (tangibility, reliability, responsiveness, assurance, empathy) of the banks in Bharatpur?
2. What is the level of customer satisfaction with the banking services in Bharatpur?
3. What is the gap between customers' expectations and perceptions of service quality?
4. Is there a significant relationship between the dimensions of service quality and customer satisfaction?
5. Which dimension of service quality has the greatest impact on customer satisfaction?

(b) Why inferential statistics was used

Descriptive statistics only summarise the data of the 421 respondents. Inferential statistics were needed because:

• The study used a sample, and the researcher wanted to generalise the findings to all bank customers of Bharatpur.
• Hypotheses (alternative against null) had to be tested, for example whether the gap between perception and expectation is significant, using the t-test, correlation or regression.
• They show the relationship between service quality and satisfaction (bivariate analysis) and its strength and significance.
• They allow the researcher to decide whether results arose by chance or reflect a real effect at a stated significance level (for example 5%).

(c) Theoretical framework

• Independent variables: service quality dimensions: tangibility, reliability, responsiveness, assurance and empathy.
• Dependent variable: customer satisfaction.
• Moderating variables: customer characteristics such as age, gender, income or education level, and type of bank.

Diagram (flow):

Tangibility, Reliability, Responsiveness, Assurance, Empathy $\rightarrow$ Customer satisfaction

Moderators (age, gender, income, bank type) act on the arrow between the independent and dependent variables.

(d) The study was quantitative in nature

• The data are numerical and measured, collected using a structured questionnaire.
• The sample is large (421 respondents) and drawn using probability random sampling, allowing generalisation.
• Statistical analysis was used: descriptive and inferential statistics, including correlation.
• It tested hypotheses (alternative against null) and measured relationships between variables, which is deductive reasoning.
• It measured the gap between perception and expectation, and the strength of correlation (for example low positive in empathy), as numbers.
• The researcher is objective and detached from the respondents, as in the positivist tradition.

Hence the design is quantitative (descriptive and correlational, cross-sectional survey).

Research is a systematic and objective process of collecting, analysing and interpreting data to solve a problem, answer a question or add to existing knowledge. In the words of Redman and Mory, it is a "systematised effort to gain new knowledge".

Convenience sampling is a non-probability sampling technique in which the researcher selects respondents who are easiest to reach, available and willing to participate, rather than choosing them by random selection.

For example, a researcher studying customer satisfaction may simply survey the customers who happen to be present in a bank branch on a given day. It is quick, cheap and simple, but it may be biased and the results cannot be generalised safely to the whole population.

A semantic differential scale measures attitude towards an object using a series of bipolar adjectives placed at the two ends of a 5- or 7-point scale. The respondent marks the point that best reflects his or her feeling.

Example: Rate the service of ABC Bank.

A face-to-face interview is conducted in the following steps.

1. Decide the objective and prepare the interview schedule or guide with clear, ordered questions.
2. Select the respondents and fix a convenient time and place by appointment.
3. Introduce yourself and the purpose of the study, assure confidentiality and obtain consent.
4. Build rapport, ask the questions in a neutral way without leading the respondent, and listen carefully.
5. Probe where answers are incomplete, and record the responses by notes or with permission by audio recording.
6. Close politely by thanking the respondent, then review and organise the notes immediately.

Data can be presented visually using diagrams such as bar diagrams, pie charts or line graphs.

Example: A pie chart can show the share of customers by satisfaction level in a survey of 100 customers: Highly satisfied 40%, Satisfied 35%, Neutral 15%, Dissatisfied 10%. Each slice is drawn in proportion to its percentage (for example, the 40% slice has a central angle of $0.40 \times 360^\circ = 144^\circ$), so the reader can compare the categories at a glance.

Grounded theory is a qualitative research approach in which a theory is developed inductively from the data collected, rather than testing a theory formed beforehand. The researcher collects data (mainly through interviews and observation), codes it into concepts and categories, compares them continuously, and allows a theory to "emerge" that is grounded in the data. It was introduced by Glaser and Strauss (1967). It is useful when little is known about a phenomenon.

The null hypothesis ($H_0$) is a statement of no difference, no effect or no relationship between variables. It is the hypothesis that is tested statistically, and the researcher either rejects it or fails to reject it on the basis of sample evidence.

Example: $H_0$: There is no significant relationship between service quality and customer satisfaction in ABC Bank. The alternative hypothesis $H_1$ states that there is a significant relationship.

Two disadvantages of secondary data:

1. May not fit the research objective: The data were collected for another purpose, so the variables, units, definitions or time period may not match the present problem.
2. Questionable accuracy and being outdated: The researcher has no control over how the data were collected, so errors or bias may exist, and the data may be old and no longer relevant.

APA format for a book: Author surname, Initials. (Year). Title of the book (edition). Publisher.

Example:

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

In the text it is cited as (Kothari, 2019).

Validity and reliability are the two key criteria for judging the quality of a measuring instrument.

A weighing scale that always shows 2 kg extra is reliable (consistent) but not valid (not accurate).

The internet is now the main source for a literature search. The process is as follows.

1. Define the topic and keywords: List the main concepts, synonyms and related terms (for example, "customer satisfaction", "banking", "Nepal").
2. Use scholarly databases and search engines: Use Google Scholar, JSTOR, ProQuest, EBSCO, ScienceDirect, Emerald, NepJOL (Nepal Journals Online) and university digital libraries, rather than general web pages only.
3. Use search techniques: Combine keywords with Boolean operators (AND, OR, NOT), use quotation marks for exact phrases, and filter by year, subject and document type.
4. Screen the results: Read the title and abstract first, then select the relevant and credible sources (peer-reviewed journals, books, reports).
5. Follow citations: Use the reference lists of good papers and "cited by" links to find more studies.
6. Evaluate the sources: Check the author, publisher, date, and reliability, and avoid unverified websites.
7. Record and manage references: Save the full bibliographic details using tools such as Mendeley or Zotero, and cite in APA style.

Deductive approach: Reasoning moves from general to specific. The researcher starts with an existing theory, develops hypotheses from it, collects data to test them, and then confirms or rejects the theory. It is common in quantitative research.

Example: Theory says higher service quality increases customer satisfaction. The researcher hypothesises that service quality is positively related to satisfaction at ABC Bank, surveys customers and tests the hypothesis with correlation and regression.

Inductive approach: Reasoning moves from specific to general. The researcher starts by collecting data and observing patterns, and then builds a general theory or conclusion from them. It is common in qualitative research.

Example: A researcher interviews bank customers, finds that many mention friendly staff and quick complaint handling as reasons for loyalty, and from these observations develops a theory about how customer loyalty forms.

This statement means that clearly defining a research problem is the most important step of research, and once the problem is stated properly, much of the work needed to solve it is done.

Why a well-stated problem is half solved:

• Gives direction: It sets the objectives, scope and boundaries of the study, so the researcher does not wander.
• Guides the design: The research design, variables, hypotheses, sample and data collection method follow from the problem statement.
• Saves time, cost and effort: A clear problem avoids collecting irrelevant data.
• Helps in literature review: The researcher knows what to search for.
• Helps analysis and conclusion: Results can be judged against the stated objectives.
• Avoids vague or wrong solutions: Solving a wrongly stated problem gives useless results.

Example: "Why are sales falling?" is vague. "What is the effect of price, promotion and product quality on the sales of Brand X in Kathmandu during 2024?" is specific, and it immediately shows which variables to measure and how.

Hence a precisely defined problem, as stated by Einstein, is often more essential than its solution.

Scientific research has the following main features.

1. Systematic: It follows an organised sequence of steps from problem identification to reporting.
2. Objective: Findings are based on facts and evidence, free from personal bias or opinion.
3. Empirical: Conclusions rest on observation and data collected from the real world.
4. Logical: It uses deductive and inductive reasoning to reach conclusions.
5. Replicable: Other researchers can repeat the study under the same conditions and get similar results.
6. Verifiable and testable: Hypotheses can be tested and results checked.
7. Generalisable: Results from a sample can be applied to the wider population, within limits.
8. Controlled: Extraneous variables are controlled or accounted for to find true relationships.
9. Reliable and valid: Instruments are accurate and consistent.
10. Cumulative and ethical: It builds on earlier knowledge and follows ethical norms.

Quantitative research designs collect numerical data and use statistical analysis to describe variables, test hypotheses and examine relationships. They are mainly of the following types.

1. Descriptive research design It describes the characteristics of a population or phenomenon (who, what, when, where, how much) through surveys. Example: the profile and opinions of customers of a bank. It can be cross-sectional (one point in time) or longitudinal (over time).

• Strength: simple, gives a broad picture, can use large samples.
• Limitation: cannot explain causes.

2. Correlational research design It examines the degree and direction of relationship between two or more variables, using correlation and regression, without manipulating them. Example: the relationship between service quality and customer satisfaction.

• Strength: shows association and allows prediction.
• Limitation: correlation does not prove causation.

3. Causal-comparative (ex post facto) design It compares groups that differ in some already existing condition to find possible causes. Example: comparing productivity of trained and untrained employees.

• Limitation: no control over the independent variable, so causes are only suggested.

4. Experimental research design The researcher manipulates an independent variable, controls other variables and measures the effect on the dependent variable, usually with experimental and control groups and random assignment. Example: testing the effect of a discount on sales in two groups of stores.

• Strength: strongest evidence of cause and effect.
• Limitation: costly, artificial, and sometimes ethically or practically difficult in business settings.

5. Quasi-experimental design It is similar to the experimental design but without random assignment of subjects, used where randomisation is impossible. It is more practical but has weaker internal validity.

Critical evaluation: Quantitative designs offer objectivity, precision, generalisability and replicability, and hypotheses can be tested statistically. However, they may miss deeper meanings and context, depend on good instruments and sampling, and can give a superficial picture of human behaviour. The choice of design should therefore match the research question: descriptive for "what", correlational for "relationship" and experimental for "cause and effect". Mixed methods are often used to cover the weaknesses.

Primary data are first-hand data collected afresh by the researcher for the specific purpose of the study. The main methods are as follows.

1. Observation The researcher watches and records behaviour, events or conditions directly. It may be participant or non-participant, structured or unstructured, and disguised or undisguised.

• Merits: records actual behaviour, not what people say; independent of respondents' willingness.
• Demerits: time-consuming, observer bias, cannot reveal motives or attitudes.

2. Interview Questions are asked orally, face to face or by telephone, and may be structured, semi-structured or unstructured.

• Merits: high response rate, flexibility, probing is possible, non-verbal cues observed.
• Demerits: costly, time-consuming, interviewer bias, and respondents may be reluctant on sensitive issues.

3. Questionnaire / survey A set of written questions (open or closed, often Likert scales) is given to respondents personally, by mail or online.

• Merits: economical for large samples, uniform questions, anonymity, convenient for respondents.
• Demerits: low response rate, cannot clarify misunderstanding, no probing, illiterate respondents cannot answer.

4. Focus group discussion A moderator leads a small group (6-10 people) in a discussion on a topic.

• Merits: rich insights, group interaction produces ideas quickly.
• Demerits: dominance by some members, difficult to generalise, moderator bias.

5. Experiment Variables are manipulated under controlled conditions to observe cause and effect.

• Merits: establishes causality.
• Demerits: costly, artificial, limited use in business.

6. Projective techniques and case study Word association, sentence completion and similar tests reveal hidden attitudes; a case study examines a single unit in depth.

Choice of method depends on the objective, time, cost, nature of the population and accuracy needed. Using more than one method (triangulation) improves reliability.

A research report communicates the research process and findings. It is written in the following sequence of steps.

1. Analyse the subject and audience: Decide the purpose, the readers (academics, management, sponsor) and the style and length needed.
2. Collect and organise the material: Arrange the findings, tables, figures, notes and references according to the objectives.
3. Prepare an outline: Draft a logical plan of chapters and headings.
4. Write the first draft: Write each part clearly, objectively and in a simple, formal style, in the past tense.
5. Revise and edit: Check logic, accuracy, grammar, consistency, tables, figures, citations and length; get feedback.
6. Prepare references, appendices and finalise: List all sources in a consistent style (such as APA), add appendices (questionnaire, raw data), and proofread.
7. Format and submit: Bind and submit the final report, with a presentation if needed.

Typical format of the report:

• Preliminary pages: title page, certificate/approval, declaration, acknowledgement, table of contents, list of tables and figures, abstract.
• Main body:
  • Chapter 1 Introduction: background, problem statement, objectives, hypotheses, significance, limitations.
  • Chapter 2 Literature review and conceptual framework.
  • Chapter 3 Research methodology: design, population and sample, data collection, tools of analysis.
  • Chapter 4 Data presentation and analysis: tables, figures, interpretation, hypothesis testing.
  • Chapter 5 Summary, conclusions and recommendations.
• End matter: references/bibliography and appendices.

A good report is clear, concise, accurate, objective and well-organised, and properly acknowledges sources.

(a) Five probable research questions

1. Does service quality have a significant effect on customer satisfaction in ABC Bank of Nepal?
2. What is the relationship between service features and customer satisfaction?
3. How does customer complaint handling influence customer satisfaction?
4. Which of the three factors (service quality, service features, complaint handling) has the strongest influence on customer satisfaction?
5. What is the level of customer satisfaction at ABC Bank, and how does it affect customers' future behavioural intention?

(b) Why the correlational research design was used

A correlational design examines the existence, direction and strength of the relationship between two or more variables without manipulating them. It was used here because:

• The objective was to find the relationship of service quality, service features and complaint handling with customer satisfaction, and of satisfaction with future behavioural intention.
• These variables cannot be controlled or manipulated by the researcher; they exist naturally among the bank's customers, so an experiment is not feasible.
• Correlation and regression analysis on Likert-scale data from SPSS could show statistical significance, direction (positive) and strength of the relationships.
• It helps to predict satisfaction from the factors and gives guidance to management on which factors to improve. Its limitation is that it shows association, not proof of cause and effect.

(c) Theoretical framework

Independent variables: service quality, service features, complaint handling. Dependent variable: future behavioural intention, with customer satisfaction as the mediating link. (If a moderating variable is wanted, a demographic factor such as age or gender of customers can be used.)

Independent variables                                       Dependent variable

 Service quality      ----\
 Service features     ------>  Customer satisfaction  ----->  Future behavioural intention
 Complaint handling   ----/              ^
                                         |
                       Moderating variable (customer age / gender / experience)

Here customer satisfaction is the outcome of the three factors and also determines future behavioural intention. The moderating variable affects the strength of the relationship between the independent variables and satisfaction.

(d) The study is quantitative

• Numerical data: A five-point Likert scale questionnaire converted opinions into numbers.
• Large sample and structured instrument: 310 questionnaires were distributed and 278 were returned (89.7% response rate, since $278/310 = 0.897$).
• Statistical analysis: Data were analysed using SPSS with correlation and regression to test significance.
• Deductive and objective: Predetermined relationships between variables were tested, and the findings are reported as statistics (highly significant, positive relationships).
• Generalisation: Results from the sample are used to draw conclusions about the customers of the bank. The case itself states that the design was quantitative in nature, which supports this.

Two objectives of research:

1. To discover new facts or increase knowledge about a phenomenon (exploration and explanation).
2. To solve practical problems and support decision-making by testing hypotheses and finding relationships between variables.

Reliability means the consistency or stability of a measuring instrument: it gives the same results when the measurement is repeated under similar conditions. It is checked by test-retest, split-half, parallel forms or Cronbach's alpha.

A mediating variable explains how or why an independent variable affects a dependent variable; it lies in the causal chain between them (IV → mediator → DV).

Example: Training (IV) improves job satisfaction (mediator), which increases employee performance (DV).

A Type I error occurs when a true null hypothesis is rejected (a false positive). Its probability is the significance level $\alpha$.

Example: Concluding that a new training programme improves sales when it actually has no effect.

A face-to-face interview is a method of data collection in which the interviewer meets the respondent in person, asks questions and records the answers. It allows probing, observation of body language and a high response rate.

Example: A researcher visits customers at a bank branch to ask about their satisfaction with its services.

Applied research is research done to solve a specific, practical problem or to improve a decision, product or policy, using existing theory in a real situation.

Example: A study to find how to reduce employee turnover in a bank.

A literature review is essential because it:

• Shows what is already known and avoids duplicating earlier studies.
• Helps identify the research gap and define the problem.
• Provides theories, concepts and variables for the framework and hypotheses.
• Guides the choice of methods, instruments and data analysis.
• Supports the interpretation and comparison of the findings.

A cohort study is a longitudinal observational study in which a group of people (cohort) sharing a common characteristic, such as birth year or exposure, is followed over time to observe outcomes. It is used to find how exposure affects later results.

Example: Following the graduates of 2020 for ten years to see how their career growth differs.

Data are raw facts, figures, observations or measurements collected for research, which become information after processing.

Sources of data:

(a) Primary data are collected first-hand by the researcher for the specific purpose. Sources and methods: survey (questionnaire), interview, observation, experiment, focus group. They are accurate and up to date but expensive and time-consuming.

(b) Secondary data were already collected by others for another purpose. Sources: published reports, government statistics, journals, books, company records, websites, the central bureau of statistics. They are cheap and quick but may be outdated or not fit the exact need.

A study often uses both: secondary data for background and primary data for specific findings.

Purposes of a research proposal:

1. Presents the problem, objectives and significance of the study to reviewers.
2. Shows the researcher's understanding of the topic and literature.
3. Describes the methodology, sampling, data collection and analysis plan.
4. Helps in getting approval, funding or permission.
5. Gives a plan with timetable and budget for execution.
6. Serves as a guide for the researcher and a basis for evaluation.

Research design is the overall plan or blueprint that specifies how data will be collected, measured and analysed to answer the research questions. It ensures that evidence collected answers the problem accurately and efficiently.

Two types:

(a) Descriptive research design: describes characteristics of a population or phenomenon, such as who, what, where and how much. It uses surveys, observation and case studies, and does not manipulate variables. Example: describing the profile and satisfaction of bank customers.

(b) Causal (experimental) research design: examines cause-and-effect relationships by manipulating independent variables and controlling other factors, with experimental and control groups. Example: testing the effect of a price discount on sales in two groups of stores.

(Exploratory design is another type, used to gain insight when the problem is unclear.)

A research interview is a planned conversation in which the researcher asks questions and the respondent gives information relevant to the research problem.

Personal (face-to-face) interview: the interviewer meets the respondent directly.

• Advantages: high response rate, probing and clarification, observation of non-verbal cues.
• Limitations: costly, time-consuming, risk of interviewer bias, and limited geographic reach.

Telephone interview: questions are asked over the phone.

• Advantages: cheap and fast, wide coverage, easier supervision.
• Limitations: shorter length, no visual cues, lower trust, and only people with phones can be reached.

A research report is a formal written document that presents the problem, methods, findings, conclusions and recommendations of a research study to readers.

Contents of a report format:

1. Preliminary pages: title page, certificate or declaration, acknowledgement, abstract, table of contents, list of tables and figures, abbreviations.
2. Main body:
   • Introduction: background, problem statement, objectives, hypotheses, significance, limitations.
   • Review of literature and theoretical framework.
   • Research methodology: design, population and sample, data collection, tools, analysis.
   • Data presentation and analysis, with interpretation of results.
   • Summary, findings, conclusions and recommendations.
3. End matter: references (bibliography), appendices (questionnaire, tables), glossary and index.

Scientific research is a systematic, objective and logical search for knowledge. Its main features:

1. Empirical: based on observed and measurable evidence rather than opinion.
2. Systematic and logical: follows ordered steps, from problem to conclusion.
3. Objective: free from personal bias and values; findings can be checked by others.
4. Replicable (verifiable): the same study repeated gives similar results.
5. Reliable and valid: uses tested instruments that measure consistently and accurately.
6. Generalisable: findings from a sample can be applied to the wider population.
7. Theory-based and analytical: relies on concepts, hypotheses and tests, and aims at explanation, prediction and control.
8. Cumulative and ethical: builds on previous knowledge and follows ethical rules of honesty and confidentiality.

A theoretical framework is a structure of concepts, theories and variables, with their expected relationships, that guides the study. Its needs:

1. Foundation: gives the study a base in established theory and connects it with prior literature.
2. Identifies variables: clarifies the dependent, independent, mediating and moderating variables.
3. Develops hypotheses: the expected relationships lead to testable hypotheses.
4. Guides design: helps in choosing measurement, sampling and analysis methods.
5. Explains findings: provides a basis for interpreting results and comparing with other studies.
6. Adds focus and clarity: avoids aimless study and improves credibility, and helps in generalisation.

Data analysis is the process of examining, organising, summarising and interpreting collected data, using statistical or logical techniques, to find patterns, test hypotheses and reach conclusions.

Steps in data processing:

1. Editing: checking questionnaires for errors, omissions, inconsistencies and illegible entries.
2. Coding: assigning numerical symbols to responses so they can be entered into a computer.
3. Classification: grouping data into classes by common characteristics (by attribute or class interval).
4. Tabulation: arranging classified data in tables or cross-tabs for comparison.
5. Data entry and cleaning: entering data into software (SPSS, Excel) and checking for outliers and missing values.
6. Analysis and interpretation: applying descriptive statistics (mean, percentage), correlation, regression, t-test, chi-square or others, then explaining the results in relation to the objectives.
7. Presentation: displaying results with graphs and charts.

(a) Research questions

1. What is the impact of legal, ethical, economic and philanthropic CSR on customer satisfaction of Nepal Telecom?
2. Which dimension of CSR has the greatest effect on the satisfaction of NTC customers?

(b) Research design

A descriptive and causal (explanatory) cross-sectional survey design was used. It is descriptive because it surveyed 200 subscribers with a questionnaire in one period (January-March 2017) to describe their perceptions. It is causal/explanatory because it tests the effect of the four CSR dimensions (independent) on customer satisfaction (dependent) using statistical analysis, without manipulating the variables.

(c) Theoretical framework

Independent variables                    Dependent variable
 Legal CSR          ──┐
 Ethical CSR        ──┤
 Economic CSR       ──┼──────────►   Customer satisfaction
 Philanthropic CSR  ──┘                       ▲
                                              │
                          Moderating variable (e.g. age, gender, education)

The moderating variable (such as age, gender or usage) is shown as affecting the strength of the relationship, since the study was done among college subscribers.

(d) Research hypotheses

• $H_1$: Legal CSR has a significant positive effect on customer satisfaction.
• $H_2$: Ethical CSR has a significant positive effect on customer satisfaction.
• $H_3$: Economic CSR has a significant positive effect on customer satisfaction.
• $H_4$: Philanthropic CSR has a significant positive effect on customer satisfaction.

Criteria for a literature search through the Internet: relevance to the topic and keywords used; authority and credibility of the author or publisher; accuracy and reliability of the content; currency (date of publication); objectivity and absence of bias; coverage of the topic; and use of scholarly sources such as journals and databases rather than unverified websites.

Categorical data (qualitative data) are data placed into groups or categories according to a quality or attribute rather than a numerical value. They are either nominal (no order) or ordinal (ordered). Examples: gender (male, female), marital status, religion, brand preference, or satisfaction level (low, medium, high).

Thematic analysis is a qualitative analysis method of identifying, organising and interpreting patterns (themes) in textual data such as interview transcripts or open-ended responses. The researcher reads the data, codes it, groups codes into themes, reviews the themes, and reports them to answer the research question.

Features of computer-assisted interviews (CATI/CAPI):

• Questions appear on screen in a pre-programmed order, with automatic skip patterns.
• Answers are entered directly into the computer, so data entry and coding errors are reduced.
• Built-in checks detect inconsistent or invalid responses at once.
• Data are available immediately, saving time and cost.
• Questions and answer options can be randomised to reduce bias.
• Consistency across interviewers and easy handling of large samples.

Importance of a case study:

• It gives an in-depth, detailed understanding of a single person, organisation or event in its real-life context.
• It uses multiple sources of evidence (interviews, documents, observation).
• It helps in generating hypotheses and theories.
• It is useful for studying rare or complex problems and for exploring "how" and "why" questions.
• It links theory with practice and provides practical lessons for decision making.

Experimental research is appropriate when the researcher wants to establish a cause-and-effect relationship, can manipulate one or more independent variables, can control extraneous variables, and can assign subjects randomly to treatment and control groups. Examples are testing the effect of a price change, advertisement or training programme on sales or performance.

Cooper, D. R., & Schindler, P. S. (2006). Business research methods. New Delhi: Tata McGraw-Hill.

A research proposal includes:

• Title of the study
• Introduction and background
• Statement of the problem
• Research objectives and research questions
• Hypotheses
• Significance of the study
• Literature review and conceptual framework
• Research methodology (design, population and sample, data collection and analysis)
• Limitations and delimitations
• Time schedule and budget
• References and appendices

Even a study done purely out of academic interest raises practical and ethical problems.

Practical problems

• Defining and measuring "achievement motivation" is hard; valid and reliable instruments are needed.
• Many factors (family, teaching, peers, personality, finance) interact, so isolating the contributing factors is difficult.
• Obtaining access to students, a representative sample and cooperation takes time.
• Limited time, funds and the lecturer's own workload; non-response and incomplete questionnaires.
• Analysis and interpretation of motivation data can be complex.

Ethical problems

• Informed consent: students must know the purpose and agree voluntarily, especially because the lecturer holds power over them and they may feel pressured.
• Confidentiality and anonymity of responses and grades must be protected.
• No harm: students should not feel labelled or stressed by questions about weak motivation or results.
• Honesty in reporting: the lecturer's own bias must not distort findings or results.
• Permission from the institution and proper use of student records.

Hypothesis: A hypothesis is a tentative, testable statement about the relationship between two or more variables, proposed as an explanation to be tested by research. Example: "Advertising expenditure has a positive effect on sales."

Types of hypothesis

1. Null hypothesis ($H_0$): states that there is no relationship or no difference. Example: $H_0$: there is no significant relationship between advertising and sales.
2. Alternative hypothesis ($H_1$ or $H_a$): states that a relationship or difference exists. Example: $H_1$: there is a significant relationship between advertising and sales.
3. Directional (one-tailed) hypothesis: predicts the direction of the relationship. Example: male employees are more satisfied than female employees.
4. Non-directional (two-tailed) hypothesis: states a difference or relationship without a direction. Example: employee satisfaction differs between males and females.
5. Simple hypothesis: one independent and one dependent variable. Example: training increases productivity.
6. Complex hypothesis: two or more independent and/or dependent variables. Example: training and incentives increase productivity and job satisfaction.
7. Descriptive and relational/causal hypotheses: the first describes a variable or a situation, the second states association or cause-effect.

Developmental research design: also called longitudinal or growth research, it studies how variables or groups change over time. Data are collected from the same people (panel) or different groups at different times, as in studying the growth of consumers' brand loyalty over several years. A cross-sectional version compares different age or stage groups at one time.

Correlational research design: examines the degree and direction of the relationship between two or more variables without manipulating them. The relationship is measured by a correlation coefficient $r$ ($-1 \le r \le +1$). It shows association, not cause. Example: the relationship between employees' training hours and job performance; a high positive $r$ shows the two move together.

Causal (explanatory/experimental) research design: aims to establish cause-and-effect. The researcher manipulates the independent variable, controls other variables and observes the effect on the dependent variable, usually using control and experimental groups. Example: a firm cuts the price of a product in one set of stores and keeps it unchanged in another, then compares the sales to see whether the price change caused the increase.

Difference between sampling error and non-sampling error

Reducing sampling error in questionnaire surveys

• Increase the sample size.
• Use probability sampling (random, stratified, cluster) rather than convenience sampling.
• Use an accurate and complete sampling frame.
• Use stratification so that all important groups are represented.
• Follow up non-respondents to improve response rate.
• Choose an appropriate sampling design and procedure and use pilot testing.

Research report: A research report is a formal, written presentation of the problem studied, the methods used, the findings, the conclusions and the recommendations of a research study, prepared for a specific audience such as a university, sponsor or manager.

Characteristics of a good research report

• Clarity: language is simple, precise and free from ambiguity.
• Accuracy: facts, figures and references are correct and supported by data.
• Completeness: it covers the objectives, methodology, findings and conclusions.
• Logical organisation: well-structured sections (introduction, literature, method, results, conclusion) with headings.
• Objectivity: impartial, free from the researcher's bias.
• Conciseness: avoids unnecessary detail and repetition.
• Proper use of tables, charts and diagrams for easy understanding.
• Proper documentation: citations, references and appendices in a standard style such as APA.
• Practical recommendations that are relevant and workable.
• Good presentation: correct grammar, formatting and layout.

Measurement scale: a measurement scale is a system of assigning numbers or symbols to the characteristics of objects or people according to rules, which determines the type of data and the statistics that can be used.

Types of measurement scales

1. Nominal scale: classifies items into categories with no order; numbers are only labels. Examples: gender, religion, brand name. Statistics: mode, frequency, chi-square.
2. Ordinal scale: categories are ranked in order, but the distance between ranks is unknown. Examples: rank in class, satisfaction (poor, fair, good). Statistics: median, percentile, rank correlation.
3. Interval scale: ordered with equal distances between values but no true zero. Examples: temperature in Celsius, attitude scores such as the Likert scale. Statistics: mean, standard deviation, correlation.
4. Ratio scale: has all the properties of the interval scale plus a true zero, so ratios are meaningful. Examples: income, age, sales, weight. All statistical operations are possible.

The scales increase in power from nominal to ratio.

(a) Title, objectives and statement of the problem

Title: "Effect of Advertising Budget, Product Packaging and Public Relations Activities on Sales of a Multinational Corporation."

Objectives

1. To examine the effect of advertising budget, product packaging and public relations activities on sales.
2. To identify which of the three factors has the greatest influence on sales.

Statement of the problem: The corporation spends money on advertising, packaging and public relations, but it is not clear how much each of these actually contributes to sales. Without this knowledge, resources may be wasted on ineffective activities and sales growth may be lost. This study tries to find the relationship and relative influence of these promotional factors on sales.

(b) Research methodology and hypotheses

• Research design: descriptive and causal (explanatory) design, quantitative.
• Population and sample: the population is the company's branches or sales territories (or customers) in the market; a sample of, say, 200 is chosen through stratified random sampling.
• Data sources: primary data from a structured questionnaire (five-point Likert scale) on advertising, packaging and PR; secondary data on advertising budget and sales from company records and reports.
• Instrument and pre-testing: the questionnaire is pilot tested; reliability is checked with Cronbach's alpha and validity through expert review.
• Data collection: surveys, personal visits and online forms, with consent and confidentiality.
• Analysis: descriptive statistics, correlation and multiple regression.

Hypotheses

• $H_1$: Advertising budget has a significant positive effect on sales.
• $H_2$: Product packaging has a significant positive effect on sales. (Public relations activities may be tested in a third hypothesis in the same way.)

(c) Conceptual framework

 Independent variables                Dependent variable

 Advertising budget      ----\
 Product packaging       -----+---->        Sales
 Public relations        ----/                ^
                                              |
 Other (moderating/control) variables: price, product quality,
 distribution, competition, economic conditions, brand image

Terms: the independent variables are advertising budget (money spent on promotion), product packaging (design, material, labelling) and public relations (sponsorship, press, community and image activities). The dependent variable is sales (revenue or units sold). Other variables such as price, quality, distribution, competition and economic conditions are included as control or moderating variables, because they also affect sales.

(d) Data analysis process and tools; limitations

Process: editing the data for errors and completeness; coding and entering in SPSS or Excel; tabulating; checking reliability (Cronbach's alpha) and normality; analysing; interpreting; and presenting the results.

Tools:

• Descriptive statistics: mean, standard deviation, frequencies.
• Pearson correlation to see the relationships between the variables.
• Multiple regression: $Sales = \beta_0 + \beta_1 (Advertising) + \beta_2 (Packaging) + \beta_3 (PR) + \varepsilon$
• $t$-test, $F$-test, $R^2$ and ANOVA to test the hypotheses and the model.
• Tables and charts for presentation.

Limitations

1. The study is limited to one corporation and a limited sample, so results may not be generalised.
2. Other factors such as price, competition and the economy cannot be fully controlled.
3. Time and budget constraints, and possible bias in the respondents' answers.