Social Research Methods & Ethics: Designs, Ethics, and Analysis

Core Concepts & Definitions

What is Social Research?

Social research is a systematic, scientific investigation of social phenomena to understand human behavior, relationships, and institutions. It aims to:

• Describe social realities (e.g., consumer preferences in Kathmandu).
• Explain causes (e.g., why Pathao drivers leave the platform).
• Predict trends (e.g., NEPSE stock volatility).
• Prescribe solutions (e.g., improving NTC’s customer service).

Three Ways of Doing Sociological Research (Visualized Below):

mindmap
  root((Social Research Methods))
    Qualitative
      Ethnography["Fieldwork (e.g., observing Daraz warehouse operations)"]
      Interviews["In-depth (e.g., interviewing Nabil Bank loan officers)"]
      Focus Groups["Group discussions (e.g., eSewa user experience testing)"]
    Quantitative
      Surveys["Structured questionnaires (e.g., NTC customer satisfaction polls)"]
      Experiments["Controlled tests (e.g., A/B testing WhatsApp ad designs)"]
      Statistical Analysis["Data crunching (e.g., analyzing NEPSE trading patterns)"]
    Mixed Methods
      Triangulation["Combining surveys + interviews (e.g., studying Pathao driver turnover)"]
      Sequential["Phases (e.g., qualitative first, then quantitative validation)"]

Research Design: The Blueprint

Research design is the framework for collecting and analyzing data. It ensures validity, reliability, and objectivity.

Key Types of Research Designs

Social Network Analysis (SNA): Mapping Connections

Social Network Analysis (SNA) studies relationships and flows between people/organizations. In business, it reveals:

• Influence: Who are the key opinion leaders in Kathmandu’s startup scene?
• Resource flow: How does information spread in a company like Chaudhary Group?
• Collaboration: Which Daraz suppliers are most interconnected?

How SNA Works in Business

graph TD
  A["SNA Goal: Understand Relationships"] --> B["Step 1: Define Nodes"]
  B --> C["Nodes = Actors (e.g., employees, customers, suppliers)"]
  A --> D["Step 2: Define Ties"]
  D --> E["Ties = Connections (e.g., emails, transactions, friendships)"]
  A --> F["Step 3: Map the Network"]
  F --> G["Visualize with tools like Gephi or UCINET"]
  A --> H["Step 4: Analyze Metrics"]
  H --> I["Centrality (who’s most connected?)"]
  H --> J["Density (how clustered is the network?)"]
  H --> K["Clusters (who works together?)"]
  A --> L["Step 5: Apply Insights"]
  L --> M["Example: Nabil Bank identifies high-value customer clusters for targeted loans."]

Real-World Example: eSewa’s Trust Network eSewa uses SNA to:

1. Identify fraud rings by detecting unusual transaction patterns.
2. Boost user trust by highlighting well-connected, verified merchants.
3. Optimize recommendations (e.g., "Users like you also bought...").

Research Ethics: The Moral Compass

Research ethics ensures studies are conducted fairly, transparently, and without harm. Core principles:

1. Informed Consent: Participants must know the study’s purpose, risks, and right to withdraw.
   • Example: Before interviewing NTC employees, explain how data will be used (e.g., "to improve service, not for layoffs").
2. Confidentiality: Protect identities (e.g., anonymizing survey responses).
3. Anonymity: Ensure no one can link data to individuals (e.g., coding names as "Respondent 1").
4. Avoiding Harm: No physical/psychological damage (e.g., don’t ask Pathao drivers about unsafe working conditions without support resources).
5. Honesty/Transparency: Disclose funding sources, conflicts of interest (e.g., "This study is funded by Daraz to improve delivery routes").

Ethical Dilemmas in Business Research

mindmap
  root((Ethical Dilemmas))
    Deception["Lying to participants (e.g., 'This is a market research study' when it’s for a new WhatsApp feature)."]
    Coercion["Pressuring employees to join a study (e.g., 'Sign this or your bonus is at risk')."]
    Privacy["Using hidden cameras in Daraz warehouses to study worker efficiency."]
    Conflict of Interest["A Nabil Bank study claims loans reduce poverty, but the bank profits from high interest."]
    Data Misuse["Selling Khalti’s user data to a third party without consent."]

Case Study: Ncell’s Customer Data Scandal In 2021, Ncell faced backlash for:

• Violating anonymity: Sharing customer call logs with a telecom rival.
• Lack of consent: Using data for targeted ads without opt-in.
• Outcome: Fines and a PR crisis; now Ncell must get explicit consent for data use.

Data Collection Methods: Tools of the Trade

1. Primary Data (Collected firsthand)

• Surveys: Structured (e.g., Google Forms for NEPSE investors) or unstructured (e.g., open-ended questions for Daraz sellers).
• Interviews: Structured (fixed questions) vs. unstructured (exploratory, e.g., interviewing a Chaudhary Group CEO).
• Observation: Participant (e.g., working as a Pathao driver for a week) or non-participant (e.g., watching Kathmandu traffic from a café).
• Experiments: Field experiments (e.g., testing two WhatsApp payment interfaces in Pokhara vs. Chitwan).

2. Secondary Data (Existing sources)

• Government reports: NTC’s annual traffic surveys.
• Company data: Daraz’s sales records, Nabil Bank’s loan defaults.
• Academic journals: Studies on consumer behavior in Nepal (e.g., Journal of Nepalese Business Studies).

Comparison Table: Primary vs. Secondary Data

Data Analysis: Turning Data into Insights

Qualitative Analysis

• Thematic Analysis: Identifying patterns in interview transcripts (e.g., "Why do Daraz sellers prefer cash-on-delivery?").
  • Steps:
    1. Transcribe interviews.
    2. Code themes (e.g., "trust issues," "logistics delays").
    3. Count frequencies (e.g., "80% mentioned delivery delays").
• Content Analysis: Studying texts (e.g., analyzing NTC’s social media posts for customer complaints).

Quantitative Analysis

• Descriptive Statistics: Mean, median, mode (e.g., average loan size at Nabil Bank).
• Inferential Statistics: Hypothesis testing (e.g., "Does WhatsApp’s new UI increase user retention?").
• Software Tools: SPSS, R, Excel, or free tools like Jamovi.

Worked Example: Analyzing Pathao Driver Turnover Problem: Pathao wants to reduce driver attrition. Data Collected:

• Surveys from 500 drivers (qualitative: open-ended; quantitative: Likert scale).
• Secondary data: Driver earnings vs. fuel costs (from Pathao’s internal reports). Analysis:

1. Qualitative: Themes emerged:
   • "Low pay" (45% of responses).
   • "Unpredictable orders" (30%).
   • "Poor app support" (25%).
2. Quantitative:
   • Drivers earning <Rs. 25,000/month had 60% higher turnover.
   • Correlation: Fuel price spikes → 20% drop in active drivers. Recommendation: Pathao introduced a fuel subsidy and guaranteed minimum earnings.

In the Real World

1. eSewa’s Trust Surveys

   • Idea Used: Mixed-methods research (quantitative surveys + qualitative interviews).
   • How: eSewa sends post-transaction emails with a 5-question Likert scale (quantitative) and an open-ended "What could we improve?" (qualitative). They also conduct monthly focus groups with merchants.
   • Impact: Reduced fraud by 30% after identifying that users trusted merchants with >4.5-star ratings.

2. Daraz’s Supplier Network Analysis

   • Idea Used: Social Network Analysis (SNA).
   • How: Daraz maps supplier connections to:
     • Identify key players (e.g., a textile supplier connected to 20% of sellers).
     • Detect bottlenecks (e.g., a single logistics hub causing delays).
   • Impact: Optimized delivery routes, reducing costs by 15%.

3. NTC’s Traffic Congestion Study

   • Idea Used: Experimental design + secondary data.
   • How: NTC rerouted buses in Thapathali for 3 months (experimental) and compared congestion data (secondary: Google Maps traffic reports) before/after.
   • Impact: Proved that reducing bus stops by 20% cut delays by 40%, leading to a city-wide policy change.

4. Nabil Bank’s Loan Default Prediction

   • Idea Used: Quantitative analysis + ethical data use.
   • How: Nabil Bank analyzed 10 years of loan data (income, employment type, credit history) to predict defaults. They ensured ethical compliance by:
     • Anonymizing customer data.
     • Getting explicit consent for data use.
   • Impact: Reduced bad loans by 25% by targeting high-risk borrowers with stricter terms.

5. Pathao’s Driver Feedback System

   • Idea Used: Ethnography + surveys.
   • How: Pathao sent undercover researchers to ride with drivers for a week (ethnography) and supplemented with a driver satisfaction survey (quantitative).
   • Ethical Note: Drivers were informed post-study and compensated for their time.

Exam Tip: How to Score Full Marks

This unit is conceptual + applied, so exams test:

1. Definitions: Know the exact wording (e.g., "Research design is a plan for collecting and analyzing data to answer a research question").
2. Differentiation: Compare terms like:
   • Stratification vs. Inequality (see Unit 3 notes).
   • Primary vs. Secondary Data (table above).
3. Case Analysis: For scenarios like GlobalMart (past exam), use this 5-step framework:
   • Identify the research method (e.g., "GlobalMart used a survey to gather customer data").
   • Spot ethical issues (e.g., "Lack of informed consent if customers were forced to answer").
   • Suggest improvements (e.g., "Use mixed methods to combine surveys with interviews").
   • Link to theory (e.g., "This aligns with quantitative research principles").
   • Real-world tie: "Like Daraz’s annual reports, GlobalMart should ensure anonymity to avoid bias."
4. Visuals: If asked to "draw a diagram," sketch a simple flowchart (e.g., research process) or network map (e.g., eSewa’s payment flows). Use Mermaid syntax if allowed.
5. Ethics: Always mention 3 ethical principles (e.g., "This study violates confidentiality, informed consent, and avoiding harm").

Common Pitfalls to Avoid:

• ❌ Vague answers: Instead of "Research is important," write "Research design ensures reliability (consistent results) and validity (measuring what’s intended), critical for NTC’s traffic studies to avoid flawed policies."
• ❌ Ignoring ethics: Even if not asked, always mention ethical concerns in case studies.
• ❌ Overcomplicating: Exams reward precision. For example:
  • ❌ "Social network analysis is complex."
  • ✅ "Social network analysis identifies central nodes (e.g., top Daraz suppliers) and clusters (e.g., groups of Kathmandu-based merchants), optimizing resource allocation."

flowchart TD
  A["Exam Question: Analyze GlobalMart’s research methods."] --> B["Step 1: Identify Methods"]
  B --> C["Quantitative: Surveys (customer feedback forms)"]
  B --> D["Qualitative: Interviews (with merchants)"]
  A --> E["Step 2: Ethical Issues"]
  E --> F["No informed consent (customers forced to fill forms)."]
  E --> G["Data privacy risk (storing names without anonymization)."]
  A --> H["Step 3: Recommendations"]
  H --> I["Use mixed methods: Surveys + ethnography."]
  H --> J["Ensure anonymity: Code respondents as R1, R2..."]
  H --> K["Pilot test: Check survey questions for bias."]
  A --> L["Step 4: Real-World Link"]
  L --> M["Like Daraz’s feedback system, which combines surveys with focus groups."]