Research Design & Methodology: Types, Steps & Applications

---

Core Concepts: What is Research Design?

Research design is the blueprint of a study—it outlines how data will be collected, analyzed, and interpreted to answer research questions. A well-structured design ensures validity, reliability, and objectivity.

Key Characteristics of a Good Research Design

Why does it matter?

• Poor design → wasted resources (e.g., a survey with leading questions).
• Strong design → actionable insights (e.g., Daraz’s A/B testing for ad effectiveness).

---

Types of Research Design

Research designs are categorized based on purpose, timing, and control. The three primary types are:

---

Step-by-Step Research Design Process

Worked Example: Ncell’s Customer Retention Study

1. Problem: High churn rate among prepaid users.
2. Objective: Identify reasons for switching to competitors (e.g., NTC).
3. Design: Descriptive + Causal
   • Method: Online survey (quantitative) + focus groups (qualitative).
   • Sampling: Stratified (by age, region, usage).
   • Tool: Structured questionnaire + open-ended questions.
4. Findings: Poor network in Kathmandu’s busy routes (e.g., Thapathali) was a top complaint.
5. Action: Targeted tower upgrades in high-churn zones.

---

Methodology: Qualitative vs. Quantitative

When to Use Which?

• Qualitative: Explore why (e.g., "Why do users abandon carts on Daraz?").
• Quantitative: Measure how much (e.g., "What % of users abandon carts?").
• Mixed Methods: Combine both (e.g., survey + interviews for deeper insights).

---

Sampling Techniques: How to Select Participants

Sampling ensures representativeness and generalizability. Common methods:

1. Probability Sampling (Random selection → unbiased)
   • Simple Random: Every user has equal chance (e.g., NEPSE’s investor surveys).
   • Stratified: Divide population into subgroups (e.g., Pathao’s riders by income tier).
2. Non-Probability Sampling (Purposeful selection → faster but biased)
   • Convenience: Easy access (e.g., mall intercept surveys).
   • Snowball: Participants refer others (e.g., niche product testers).

Worked Example: Kathmandu Traffic Study

• Problem: Congestion on Ring Road during peak hours.
• Sampling: Stratified random (divide by time slots: 7–9 AM, 5–7 PM).
• Tool: GPS trackers in sample taxis + driver interviews.
• Finding: 60% of delays caused by unregulated auto-rickshaws.

---

Ethical Considerations in Research Design

Ethics ensure trust, transparency, and participant safety. Key principles:

• Informed Consent: Participants know the study’s purpose (e.g., eSewa’s user agreement disclosures).
• Anonymity/Confidentiality: Protect identities (e.g., Khalti’s transaction data encryption).
• Avoiding Harm: No misleading questions (e.g., "Do you hate our service?").
• Voluntary Participation: No coercion (e.g., offering incentives without pressure).

---

In the Real World

1. eSewa’s Customer Feedback Loop

   • Design: Mixed-methods (post-transaction surveys + qualitative interviews).
   • Why it works: Identified that mobile OTP failures (due to poor network in rural areas) caused 30% cart abandonment. Led to SMS fallback options.

2. Daraz’s A/B Testing for Ads

   • Design: Causal experimental (test ad copy A vs. B).
   • Impact: Found that humor in ads increased CTR by 22% in Nepal’s youth demographic.

3. NTC’s Fiber Optic Expansion Plan

   • Design: Descriptive + Predictive (survey current usage + model future demand).
   • Insight: Predicted 30% growth in Kathmandu Valley by 2025, justifying new towers in Lalitpur.

---

Exam Tip

1. Define Clearly: For questions like "Define research design," structure your answer as:

   "Research design is a framework outlining the methods and procedures to collect, analyze, and interpret data for solving a research problem. It includes exploratory, descriptive, and causal types, each serving distinct purposes (e.g., exploratory for eSewa’s initial wallet feedback)."

2. Compare with Examples: When asked to "evaluate factors affecting research decisions," use a table (like above) and tie each factor to a real case (e.g., sampling bias → Pathao’s skewed rider data from only Kathmandu).

3. Process Questions: For "explain the research design process," use the flowchart and add one local example (e.g., "Like Ncell’s churn study, step 5 would involve stratified sampling by region").

4. Ethics Shortcuts: Memorize the 4 ethical pillars (consent, anonymity, no harm, voluntary) and link them to Nepali contexts (e.g., "Khalti must ensure transaction data is anonymous to comply with Nepal’s data protection laws").

---

Key Formula to Remember: $$\text{Validity}=\frac{\text{Relevance of Data}}{\text{Systematic Error}}$$ (A high-validity design ensures data directly answers the research question without bias.)