Business Research Methods

Ethical Violations in the Scenario:

In the given scenario, undergraduate students collected sensitive medical data regarding fetal abnormalities from ultrasound scan rooms without informing the expectant mothers. This involves severe ethical breaches:

1. Violation of Informed Consent: Participants were not provided with information regarding the research purpose, risks, or benefits, denying them the fundamental right to voluntarily choose whether to participate.
2. Breach of Confidentiality and Privacy: Accessing personal medical diagnostic procedures and records without explicit authorization compromises patient confidentiality and dignity.
3. Lack of Institutional Review Board (IRB) / Ethics Approval: Conducting clinical human-subject research without formal institutional ethical clearance violates established academic and medical protocols (e.g., Nepal Health Research Council guidelines).

Definition of Theoretical Framework:

A theoretical framework is the conceptual foundation and structure that introduces, describes, and establishes the formal theories explaining why a specific research problem exists.

Key Aspects:

• It logically connects existing scientific theories to the research variables under investigation.
• It maps the causal, associative, or directional relationships among independent, dependent, mediating, and moderating variables.
• It provides the theoretical lens through which empirical findings and statistical hypotheses are evaluated and interpreted.

Meaning of Sampling in Research:

Sampling is the scientific and statistical process of selecting a predetermined, representative subset (a sample) of units, individuals, or elements from an entire target population to conduct an investigation.

Purpose:

• To generalize findings and draw valid statistical inferences about the broader population without the prohibitive time, expense, and logistical constraints of a full census.
• To maximize estimation accuracy while minimizing sampling error.

Differences Between Academic and Non-Academic Research Proposals:

Basis	Academic Research Proposal	Non-Academic / Applied Research Proposal
1. Primary Objective	To fulfill university degree requirements (thesis/dissertation) and contribute to theoretical knowledge.	To solve specific operational business problems or inform managerial/policy decisions.
2. Target Audience	Academic advisors, university faculty committees, and scholarly peer reviewers.	Corporate executives, funding agencies, clients, or government bodies.
3. Core Emphasis	Deep theoretical frameworks, extensive literature review, and rigorous methodological defense.	Practical viability, actionable solutions, ROI, cost breakdown, and project timelines.
4. Style & Language	Formal, scholarly, theory-intensive academic tone.	Crisp, pragmatic, executive-oriented business language.

Rejection of Null Hypothesis at 99% Confidence Level:

In hypothesis testing using the p-value approach (significance probability approach):

1. Significance Level (alpha): Confidence Level = 99% implies alpha = 1 - 0.99 = 0.01 (or 1%).

2. Decision Rule:

   • Reject H0 if p-value <= 0.01.
   • Fail to reject (accept) H0 if p-value > 0.01.

Conclusion: The researcher rejects the null hypothesis whenever the calculated p-value is less than or equal to 0.01, indicating that the observed sample evidence has less than a 1% probability of occurring by random chance under H0.

Concept of Mediating Variable with Example:

A mediating variable (or intervening variable) is an intermediate variable that explains the internal mechanism or generative process through which an independent variable (IV) influences a dependent variable (DV).

Independent Variable (IV) -> Mediating Variable (M) -> Dependent Variable (DV)

Practical Example:

• Independent Variable (IV): Employee Training Hours
• Mediating Variable (M): Employee Job Knowledge & Self-Efficacy
• Dependent Variable (DV): Work Performance Quality

Explanation: Training hours do not directly generate higher performance by themselves; training enhances employee knowledge (the mediator), which in turn produces superior performance.

Identification of Error Type:

Analysis of the Judicial Scenario:

• Null Hypothesis (H0): The accused person is innocent.
• Alternative Hypothesis (H1): The accused person is guilty.

Nature of the Error:

The judge sentenced an innocent person to 5 months of imprisonment. This means the judge rejected the null hypothesis (H0) when H0 was actually true.

Conclusion: This is a Type I Error (alpha Error / False Positive).

• Logic: In statistical hypothesis testing, rejecting a true null hypothesis constitutes a Type I error. In the legal context, convicting an innocent person represents the classic archetype of a Type I error.

Circumstances for Using Observation as a Data Collection Tool:

Observation is justified under the following research conditions:

1. Studying Actual Behavior vs. Self-Reported Claims: When participants tend to give socially desirable, biased, or inaccurate answers in surveys (e.g., studying actual shopper dwell time in supermarkets vs. self-reported time).
2. Subjects Unable to Articulate Themselves: When subjects lack verbal or written communication skills (e.g., infants, patients with cognitive impairments, or non-verbal animals).
3. Complex Physical Processes and Natural Work Environments: When examining machine ergonomics, traffic flows, or physical workplace movements that occur continuously in real-time.

Logical Evaluation of Scales:

The true statement is: “All ordinal scales are nominal scales.”

Mathematical & Logical Justification:

Measurement scales operate in a cumulative hierarchy of properties:

1. Nominal Scale: Possesses only the property of categorization/identity (grouping into discrete classes).
2. Ordinal Scale: Possesses the property of categorization plus order/rank magnitude (A > B > C).

Since every ordinal scale inherently classifies elements into distinct mutually exclusive categories, it fulfills all criteria of a nominal scale.

Conversely, nominal scales are NOT ordinal scales, because nominal classifications (e.g., Gender: Male/Female; Blood Group: A/B/AB/O) have no inherent rank or hierarchical progression.

Role of Literature Review in Academic Research:

1. Identifying Research Gaps: Uncovers unexplored theoretical domains, conflicting empirical findings, and methodological limitations in prior research.
2. Developing Conceptual and Theoretical Frameworks: Guides the selection of tested theories, constructs, and validated measurement models.
3. Preventing Duplication: Ensures the research makes a novel contribution rather than reinventing existing findings.
4. Guiding Methodological Choices: Informs the selection of appropriate research designs, sampling strategies, and statistical tools.
5. Contextualizing Results: Provides the comparative backdrop against which newly collected data can be validated and synthesized.

Research Paradigms and Classification by Reasoning:

1. Concept of Research Paradigms

A research paradigm (originating from Thomas Kuhn, 1962) represents a shared philosophical framework, worldview, or set of beliefs and practices that guides scientific inquiry. It dictates the researcher’s:

• Ontology: The nature of reality (is reality objective and singular, or subjective and multiple?).
• Epistemology: The nature of knowledge and the relationship between the researcher and the known.
• Methodology: The systematic process used to investigate and acquire knowledge.

---

2. Paradigms Classified on the Basis of Reasoning

Deductive Approach (Positivist Paradigm):

Theory -> Hypothesis -> Data Collection -> Confirmation / Rejection

Inductive Approach (Interpretivist Paradigm):

Observations -> Patterns -> Tentative Hypothesis -> New Theory Generation

A. Deductive Reasoning (The Positivist / Quantitative Paradigm)

• Top-Down Logic: Proceeds from the general to the specific (General -> Specific).
• Core Mechanism: Begins with established theoretical axioms, deduces testable causal hypotheses, gathers quantitative empirical data, and statistically confirms or rejects the hypothesis.
• Assumptions: Reality is objective, stable, measurable, and independent of human perception.
• Application: Ideal for large-scale survey research, clinical trials, and econometric modeling.

B. Inductive Reasoning (The Interpretivist / Qualitative Paradigm)

• Bottom-Up Logic: Proceeds from specific empirical observations to broad general theories (Specific -> General).
• Core Mechanism: Begins by observing human experiences and interactions in natural settings, discovers emerging themes and behavioral patterns, and builds new conceptual models.
• Assumptions: Reality is socially constructed, multiple, contextual, and deeply entwined with human consciousness.
• Application: Ideal for phenomenological studies, grounded theory, ethnography, and exploratory case studies.

C. Abductive Reasoning (The Pragmatic Paradigm)

• Synthesizes both approaches by moving back and forth between empirical facts and theory to uncover the most plausible explanation for unexpected organizational phenomena.

Inferential Statistics & Parametric vs. Non-Parametric Tests:

1. Definition of Inferential Statistics

Inferential statistics encompasses mathematical methods that allow researchers to make generalizations, probabilistic predictions, and decisions about an entire population based on findings obtained from a representative sample. Unlike descriptive statistics (which merely summarize collected data), inferential statistics test hypotheses and calculate confidence intervals.

---

2. Comparative Matrix: Parametric vs. Non-Parametric Tests

Parameter	Parametric Tests	Non-Parametric Tests
1. Distributional Assumption	Assumes the population data follows a specific distribution (typically normal Gaussian distribution).	Distribution-free; makes no assumptions regarding underlying population normality.
2. Measurement Scale	Requires continuous quantitative data measured on Interval or Ratio scales.	Accommodates discrete data measured on Nominal or Ordinal scales (as well as non-normal continuous data).
3. Central Measure Focus	Focuses on comparisons of Means and standard deviations.	Focuses on comparisons of Medians, ranks, or frequency distributions.
4. Sample Size Sensitivity	Requires relatively moderate-to-large sample sizes (N >= 30) to maintain validity.	Highly robust and reliable with small sample sizes (N < 30).
5. Statistical Power	High statistical power when assumptions are met; more capable of detecting true effects.	Slightly lower statistical power; requires slightly larger samples to achieve equivalent power.
6. Common Tests	- Independent samples t-test<br>- Paired samples t-test<br>- One-Way and Two-Way ANOVA<br>- Pearson’s correlation (r)	- Mann-Whitney U test<br>- Wilcoxon Signed-Rank test<br>- Kruskal-Wallis H test<br>- Spearman’s rank correlation (rho)<br>- Chi-Square test (chi-square)

The Scientific Research Process:

The scientific research process is a rigorous, multistep, cyclical process undertaken systematically to investigate business phenomena and establish verifiable knowledge.

1. Structural Flowchart of the Research Process

1. Identification & Formulation of Research Problem (Pinpointing the research gap, ambiguity, or managerial dilemma that requires investigation) v
2. Extensive Review of Relevant Literature (Synthesizing existing empirical findings to establish the research backdrop and prevent duplication) v
3. Development of Theoretical Framework & Testable Hypotheses (Specifying dependent and independent variables and positing directional relationships) v
4. Preparation of Research Design (Formulating the master plan: exploratory, descriptive, or causal) v
5. Sampling Design & Measurement Instrumentation (Selecting sampling method and designing validated questionnaires with scale checks) v
6. Fieldwork & Empirical Data Collection (Administering surveys, structured interviews, or systematic observations) v
7. Data Processing, Statistical Analysis & Hypothesis Testing (Coding data, verifying reliability/validity, and executing parametric/non-parametric tests) v
8. Interpretation, Conclusion & Research Report Preparation (Synthesizing empirical findings into actionable managerial recommendations)

Reliability Testing in Research & Comparative Analysis:

1. Purpose of Reliability Testing

A researcher performs a reliability test to establish the internal consistency, stability, and dependability of a research instrument. It ensures that if the measurement process is repeated under identical conditions across independent trials, it will produce identical, consistent, and error-free results. It eliminates random measurement noise.

---

2. Comparison: Test-Retest vs. Equivalent Forms Reliability

Parameter	Test-Retest Reliability	Equivalent / Parallel Forms Reliability
Core Concept	Evaluates the temporal stability of an instrument over time using a single questionnaire.	Evaluates the equivalence across two independently constructed versions of an instrument measuring the identical construct.
Administration	The exact same test is administered to the same group of subjects at two distinct points in time (T1 and T2).	Two distinct forms (Form A and Form B) are administered to the same group either simultaneously or with a brief interval.
Primary Weakness	Susceptible to memory/recall bias, maturation, and historical events occurring between sessions.	Challenging and resource-intensive to construct two genuinely parallel forms with identical difficulty and variance.
Statistical Metric	Pearson product-moment correlation coefficient (r) between Scores at T1 and Scores at T2.	Correlation coefficient (r) between scores obtained on Form A and Form B.
Practical Example	Administering an Employee Motivation Questionnaire to 50 bank tellers on Shrawan 1, and readministering the exact same questionnaire to the same 50 tellers on Shrawan 28.	Creating two different versions of a 50-item Business Statistics aptitude test (Form X and Form Y) with different numerical figures but identical difficulty, administered on the same day.

Purpose and Typology of Validity in Research:

1. Purpose of Validity Testing

Validity measures the degree of accuracy with which a research instrument measures the specific theoretical construct it is intended to measure. While reliability asks “Are we measuring consistently?”, validity asks “Are we measuring the right thing?”. Validity ensures findings reflect authentic phenomena rather than systematic errors.

---

2. Major Types of Validity

A. Content Validity:

Evaluates whether the test items adequately cover the entire operational domain of the construct.

• Face Validity: Subjective evaluation by laypersons or respondents that the test looks relevant.
• Logical / Expert Validity: Rigorous verification by subject-matter experts confirming full curricular coverage.

B. Criterion-Related Validity:

Compares instrument scores against an established external benchmark or performance standard.

• Concurrent Validity: Assessed when test scores correlate with an existing, validated standard measured at the same time (e.g., a new quick test correlating with an established clinical test).
• Predictive Validity: Assessed when scores successfully forecast future behavior or performance (e.g., GMAT scores predicting MBA academic GPA).

C. Construct Validity:

Confirms that the instrument accurately captures the abstract psychological or managerial construct.

• Convergent Validity: Demonstrated when scores correlate strongly with other tests measuring similar concepts (Average Variance Extracted > 0.50).
• Discriminant Validity: Demonstrated when scores do not correlate with tests measuring theoretically distinct concepts (Fornell-Larcker criterion).

Classification of Sampling Techniques:

Sampling techniques are bifurcated into two primary classifications based on whether selection involves known, non-zero probabilistic chance: Probability Sampling and Non-Probability Sampling.

---

1. Classification Overview

• Probability Sampling: Simple Random, Systematic Sampling, Stratified Sampling, Cluster Sampling.
• Non-Probability Sampling: Convenience Sampling, Judgmental / Purposive Sampling, Quota Sampling, Snowball Sampling.

---

2. Comprehensive Breakdown

A. Probability Sampling (Objective & Statistically Generalizable)

1. Simple Random Sampling: Every member of the population has an equal and independent chance of selection (e.g., lottery method, random number generators).
2. Systematic Sampling: Elements are selected from an ordered sampling frame at a fixed periodic interval (k = N/n) starting from a random origin.
3. Stratified Random Sampling: The heterogeneous population is partitioned into homogeneous sub-groups (strata based on age, income, gender), and random samples are drawn proportionally from each stratum.
4. Cluster Sampling: The population is divided into heterogeneous geographical clusters; entire clusters are randomly chosen and fully investigated.

B. Non-Probability Sampling (Subjective & Exploratory)

1. Convenience Sampling: Elements are selected purely based on ease of accessibility and proximity to the researcher (e.g., surveying people exiting a mall).
2. Purposive / Judgmental Sampling: Elements are chosen deliberately based on the specialized expertise or characteristics needed to answer the research question.
3. Quota Sampling: The researcher sets non-random demographic quotas (e.g., 50 males, 50 females) and fills them conveniently.
4. Snowball Sampling: Existing research subjects recruit future subjects from among their acquaintances (used for hidden or rare populations such as luxury collectors).

Case Study Analysis: Customer Service Experience in Nepalese Restaurant Sector (Limbu & Kutal, 2022)

---

Part (a): Conceptual Research Framework and Variable Definitions

1. Diagrammatic Research Framework:

• Independent Variables (IV):
  1. Social Interaction: Frontline employee courtesy, promptness, attentiveness, and warmth.
  2. Core Service Quality: Food taste, culinary presentation, portion size, and unexpected delightful touches (e.g., complimentary snacks).
  3. Physical Environment (Servicescape): Ambiance, music, lighting, cleanliness, seating comfort, and interior aesthetics.
• Moderating Variable:
  • Customer Co-Production: Active participation of the customer in specifying preferences and dining pacing.
• Dependent Variable (DV):
  • Customer Service Experience: Subjective cognitive and emotional evaluation resulting in Favorable (delight) vs. Unfavorable (dissatisfaction) outcomes.
• Consequent Outcome:
  • Customer Loyalty, Positive Word-of-Mouth, and Revisit Intention.

2. Explanation of Variables:

• Social Interaction: Represents interpersonal contact between service personnel and diners. Highly friendly staff mitigate delays, while rude or inattentive staff trigger unfavorable memories.
• Core Service Quality: The fundamental reason for visiting a restaurant. Exceptional food creates satisfaction, but unexpected delight (e.g., an unbilled chef special) cements favorable recall.
• Physical Environment: The tangible surroundings creating atmospheric comfort.
• Customer Service Experience: The holistic experiential process where favorable experiences foster long-term loyalty.

---

Part (b): Statement of the Problem

In the burgeoning Nepalese hospitality sector, particularly within urban centers like the Kathmandu Valley, customer dining preferences have shifted from basic caloric sustenance toward experiential consumption (Pradhan, 2020). Despite intensified competition and substantial capital investments in interior design, restaurant operators face high customer churn and volatile customer loyalty (Shrestha & Maharjan, 2021).

While Western service management literature emphasizes physical ambiance and mechanized service standards, the Nepalese dining context relies heavily on personalized relational dynamics and informal service gestures (Limbu & Kutal, 2022). Existing domestic studies have primarily relied on rigid, quantitative closed-ended surveys that fail to capture the nuanced, qualitative ‘voice of the customer’ during critical service incidents. Consequently, restaurant managers lack empirical clarity regarding the relative weight of social interactions versus culinary quality in generating favorable versus unfavorable customer narratives. This study addresses this empirical and contextual gap by examining the primary drivers of customer service experiences in Nepalese restaurants.

---

Part (c): Research Questions, Objectives, and Null Hypotheses

1. Research Questions (RQ):

• RQ1: Does social interaction between frontline employees and patrons significantly impact the favorability of customer service experiences in Nepalese restaurants?
• RQ2: Does core service quality (food and beverage delivery) significantly influence favorable customer service evaluations?

2. Research Objectives (RO):

• RO1: To evaluate the effect of employee social interaction on customer service experience outcomes in Nepalese dining establishments.
• RO2: To determine the influence of core service offerings on the formation of favorable customer experiences.

3. Null Hypotheses (H0):

• H01: There is no significant relationship between frontline employee social interactions and the favorability of customer service experiences in Nepalese restaurants.
• H02: Core service quality has no significant effect on the generation of favorable customer service experiences in Nepalese restaurants.

---

Part (d): Research Design Justification

To execute this research, an Explanatory Mixed-Methods Design (or Exploratory Sequential Design) utilizing the Critical Incident Technique (CIT) is employed:

1. Qualitative Phase (Critical Incident Technique):
   • Conduct in-depth qualitative interviews asking patrons to recall specific memorable incidents (either highly favorable or highly unfavorable) during their recent restaurant visits.
   • Justification: CIT allows customers to articulate experiences in their own words, capturing raw emotions, unexpected delightful gestures (e.g., free dessert), or service breakdowns.
2. Quantitative Phase (Survey Verification):
   • A structured survey administered to a larger stratified sample (N = 300) across casual, fine-dining, and traditional Nepalese restaurants using 5-point Likert scales.
   • Data Analysis: Structural Equation Modeling (SEM) or Multiple Regression to statistically test the hypothesized paths (H01 and H02).
3. Epistemological Fit: This design balances inductive contextual richness with deductive statistical generalizability, offering direct managerial utility to Nepalese restaurateurs.