Business Research Methods

Basic vs. Applied Research:

• Basic (Fundamental/Pure) Research: Conducted primarily to expand the theoretical frontiers of human knowledge and develop fundamental laws/principles, without any immediate practical commercial problem-solving application.
• Applied Research: Conducted directly to solve a specific, practical, and immediate problem faced by managers, corporations, or policy makers.

Moderating vs. Mediating Variable:

• Moderating Variable: Alters the direction or strength of the relationship between an independent ($X$) and dependent variable ($Y$). Example: Digital training ($X$) increases employee productivity ($Y$), but Age/Tech-savviness moderates this impact.
• Mediating (Intervening) Variable: Explains the internal psychological or procedural mechanism through which $X$ influences $Y$. Example: High compensation ($X$) enhances employee retention ($Y$) through Job Satisfaction (mediator).

Cronbach’s Alpha ($\alpha$) Thresholds:

Cronbach’s Alpha measures internal consistency across multi-item survey instruments:

• $\alpha \ge 0.90$: Excellent reliability.
• $0.80 \le \alpha < 0.90$: Good reliability.
• $\alpha \ge 0.70$: Minimum universally accepted threshold for scientific exploratory and descriptive business research.
• $\alpha < 0.60$: Poor/unacceptable reliability; requires item deletion.

Type I Error vs. Type II Error:

• Type I Error ($\alpha$ - Producer’s Risk / False Positive): Rejecting the null hypothesis ($H_0$) when it is actually true in reality. The probability of making this error is the significance level $\alpha$ (typically $5\%$).
• Type II Error ($\beta$ - Consumer’s Risk / False Negative): Failing to reject (accepting) the null hypothesis ($H_0$) when it is actually false in reality. Power of the test is defined as $(1 - \beta)$.

APA 7th Edition Journal Reference Format:

Author, A. A., & Author, B. B. (Year). Title of the article. Title of the Periodical, Volume(Issue), pages–pages. https://doi.org/10.xxxx/xxxx

Concrete Example: Adhikari, D. R., & Gautam, D. K. (2020). Human resource management practices and performance in Nepalese organizations. Journal of Management Research, 18(2), 115–130. https://doi.org/10.5430/jmr.v18n2p115

Stages of Scientific Business Research Process:

1. Broad Problem Identification: Observing symptoms in organizational operations (e.g. falling app user retention).
2. Preliminary Literature Review & Theoretical Exploration: Reviewing peer-reviewed journals, trade publications, and existing conceptual models.
3. Problem Statement & Research Questions: Formulating clear, testable research questions and specific objectives.
4. Theoretical / Conceptual Framework Development: Mapping Independent, Dependent, Mediating, and Moderating variables and deriving hypotheses ($H_1, H_2$).
5. Research Design: Choosing between exploratory, descriptive, or causal research designs; selecting cross-sectional vs longitudinal approaches.
6. Measurement & Instrumentation: Operationalizing abstract constructs into multi-item Likert scales, pre-testing for face and construct validity.
7. Sampling Design: Defining target population, choosing sampling method (e.g. stratified random), and calculating required sample size ($n$).
8. Data Collection: Administering digital/paper surveys, structured interviews, or experiment trials.
9. Data Cleansing & Statistical Analysis: Coding data, running descriptive statistics, reliability tests, and hypothesis tests ($t$-test, ANOVA, Regression).
10. Interpretation, Report Writing & Academic Presentation: Drawing actionable managerial insights, acknowledging limitations, and drafting according to APA 7th standards.

Probability vs. Non-Probability Sampling:

• Probability Sampling: Every element in the target population has a known, non-zero probability of selection. Enables objective statistical generalization and computation of margin of error. Includes Simple Random, Systematic, Stratified, and Cluster sampling.
• Non-Probability Sampling: Selection relies on subjective researcher judgment or convenience. Elements do not have equal chance; findings cannot be generalized to the broader population. Includes Convenience, Purposive, Quota, and Snowball sampling.

When Stratified Random Sampling is Preferred:

1. Heterogeneous Subgroups: When the population consists of distinct sub-populations (strata) that differ systematically on the research variable (e.g. bank customer sizes: Retail, SME, Corporate).
2. Guaranteed Minority Representation: Simple Random Sampling might fail to pick enough minority or small-stratum respondents by pure chance. Stratified sampling guarantees representative proportional or disproportionate quotas for each subgroup.
3. Higher Statistical Precision: Reduces within-stratum variance, yielding lower standard errors and narrower confidence intervals for the entire population estimate.

(a) Cochran’s Sample Size Formula:

For large populations:

Given:

• $Z = 1.96$ (for 95% confidence)
• $p = 0.5$ (maximum variability assumption)
• $e = 0.05$ (5% margin of error)

---

(b) Taro Yamane’s Formula:

Given $N = 500,000$ and $e = 0.05$:

Reason for Difference:

• Cochran’s formula is derived from the standard normal binomial distribution and explicitly incorporates the confidence level ($Z$).
• Yamane’s formula is a simplified approximation formula for finite populations that assumes a fixed 95% confidence level with slightly more conservative margins, yielding $n = 400$.

Four Levels of Measurement Scales:

Step 1: Hypotheses

• $H_0$: Payment method preference is independent of gender.
• $H_1$: Payment method preference is dependent on gender.

---

Step 2: Calculate Expected Frequencies ($E = \frac{\text{Row Total} \times \text{Column Total}}{\text{Grand Total}}$)

• Male / Digital Wallet: $E_{11} = \frac{100 \times 120}{200} = 60$
• Male / COD: $E_{12} = \frac{100 \times 80}{200} = 40$
• Female / Digital Wallet: $E_{21} = \frac{100 \times 120}{200} = 60$
• Female / COD: $E_{22} = \frac{100 \times 80}{200} = 40$

---

Step 3: Compute Chi-Square Statistic ($\chi^2 = \sum \frac{(O - E)^2}{E}$)

---

Step 4: Decision & Interpretation

• Degrees of freedom ($df$) = $(r - 1)(c - 1) = (2 - 1)(2 - 1) = 1$
• Critical value $\chi^2_{0.05, 1} = 3.841$
• Comparison: Calculated $\chi^2$ ($8.334$) is strictly greater than the critical value ($3.841$).
• Conclusion: We reject the null hypothesis ($H_0$) at the 5% significance level. There is statistically significant evidence that payment method preference is dependent on consumer gender (males exhibit a significantly higher preference for digital wallets).

Ethical Principles in Business Research:

1. Informed Consent:
   • Participants must be voluntarily informed regarding the explicit purpose, institutional affiliation, expected duration, and potential risks of the study.
   • Freedom to withdraw at any time without penalty or loss of benefit.
2. Confidentiality and Anonymity:
   • Personally Identifiable Information (PII such as phone numbers, citizenship numbers, corporate trade secrets) must be de-identified and encrypted.
   • Reporting data only in aggregate statistical clusters rather than individual identifiable records.
3. Plagiarism and Scholarly Integrity:
   • Strictly forbidding verbatim duplication of ideas, data, or texts without proper quotation and attribution.
   • Zero tolerance for data fabrication, data falsification, or selective cherry-picking of data points to force statistical significance ($p < 0.05$).
   • Mandatory similarity index reporting using tools such as Turnitin.

Step 1: Summary Calculations Table ($n = 5$)

Means:

---

Step 2: (a) Karl Pearson’s Correlation Coefficient ($r$)

Numerator:

Denominator:

Interpretation: There is an almost perfect positive linear correlation ($r = 0.99$) between digital advertising spend and online sales revenue.

---

Step 3: (b) Regression Equation of $Y$ on $X$ ($Y = a + bX$)

Slope coefficient ($b$):

Intercept ($a$):

Regression Equation:

---

Step 4: (c) Prediction for Advertising Spend $X = 12$

If $X = 12$ (i.e. NPR 12,000):

The expected online sales revenue is NPR 1,590,000.

1. Problem Statement:

Commercial banks in Nepal are investing heavily in AI-driven technologies (e.g. conversational chatbots, robotic process automation for KYC screening, algorithmic credit scoring, automated AML fraud detection). However, there is an acute lack of empirical evidence demonstrating whether these high capital outlays translate into tangible reductions in cost-to-income ratios, error mitigation, and loan turnaround acceleration, particularly under Nepal’s regulatory infrastructure.

---

2. Research Objectives:

1. To assess the current level of AI adoption across Class ‘A’ commercial banks in Nepal.
2. To evaluate the relationship between AI-driven automated processing and banking cost efficiency.
3. To determine the impact of algorithmic fraud detection on customer transaction security.
4. To examine the moderating effect of employee digital competence on the relationship between AI implementation and overall bank operational efficiency.

---

3. Conceptual Framework & Variables:

• Independent Variables ($IV$):
  • Robotic Process Automation ($RPA$) in Back-Office Operations
  • Algorithmic Credit Risk Assessment
  • AI Customer Service Chatbots
• Dependent Variable ($DV$):
  • Operational Efficiency (Cost reduction, transaction speed, error reduction)
• Moderating Variable ($MV$):
  • Employee Digital Competence / Tech Savviness
• Control Variables:
  • Bank Size (Total Assets) and Bank Age

---

4. Hypotheses:

• $H_1$: There is a statistically significant positive relationship between RPA adoption and banking operational efficiency.
• $H_2$: Algorithmic credit assessment significantly shortens loan processing turnaround times.
• $H_3$: Employee digital competence positively moderates the effect of AI systems on bank productivity.

---

5. Research Methodology:

• Research Design: Explanatory and descriptive cross-sectional survey design utilizing quantitative methods.
• Population & Unit of Analysis: IT managers, operations officers, and compliance executives across all 20 Class ‘A’ Commercial Banks in Nepal ($N \approx 600$).
• Sampling Technique & Size: Purposive and stratified sampling selecting head-office operations personnel. Sample size determined via Yamane ($n = 240$ respondents).
• Instrumentation: 5-point Likert scale questionnaire adapted from verified UTAUT and Technology-Organization-Environment (TOE) frameworks. Pre-tested for face validity and Cronbach’s Alpha ($\alpha > 0.80$).
• Data Analysis Plan: Data screened using SPSS/SmartPLS; statistical tests include descriptive means/SD, Pearson correlation, and Multiple Linear Regression / Structural Equation Modeling (PLS-SEM) to test path coefficients ($\beta$) at 5% significance level.