Research Design: Types, Models & Applications

TAKEAWAYS:

• Research design determines the structure and methodology of a study, ensuring validity and reliability.
• Exploratory, descriptive, and causal designs serve distinct purposes in business research.
• Experimental vs. non-experimental designs differ in control and manipulation of variables.
• Cross-sectional and longitudinal designs impact data collection scope and timeframe.
• Real-world applications include customer satisfaction surveys (descriptive), A/B testing (experimental), and market trend analysis (longitudinal).

1. Definition and Purpose of Research Design

Research design is a blueprint that outlines how a study will be conducted, including:

• Research questions to be answered.
• Variables to be measured.
• Data collection methods (surveys, experiments, observations).
• Sampling techniques (random, stratified, etc.).
• Data analysis approaches (quantitative, qualitative, or mixed).

A well-structured design ensures validity (measuring what is intended) and reliability (consistent results).

2. Types of Research Design

Research designs are classified based on purpose, timeframe, and control over variables.

A. Based on Purpose

WORKED EXAMPLE: A company like Nabil Bank wants to know if offering cashback on credit card transactions increases customer spending.

• Design: Causal (Experimental)
• Method: Randomly select 1000 customers, split into two groups:
  • Group A: Offer 5% cashback.
  • Group B: No cashback (control group).
• Measure: Compare spending over 3 months.
• Conclusion: If Group A spends significantly more, cashback is effective.

B. Based on Timeframe

REAL-WORLD APPLICATION:

• Nepal Electricity Authority (NEA) uses longitudinal design to track electricity consumption trends across seasons.
• Pathao uses cross-sectional surveys to gauge driver satisfaction in a single month.

C. Based on Control Over Variables

WORKED EXAMPLE (Experimental Design): A startup like Khalti wants to test if sending SMS reminders increases app usage.

• Independent Variable (IV): SMS reminders (yes/no).
• Dependent Variable (DV): App usage frequency.
• Method:
  • Randomly assign 500 users to Group A (SMS reminders) and 500 to Group B (no SMS).
  • Measure app logins over 1 month.
• Result: If Group A logs in 30% more, SMS reminders work.

WORKED EXAMPLE (Non-experimental Design): Nepal’s Central Bureau of Statistics (CBS) studies the impact of education on income without intervening.

• Variables:
  • IV: Years of education.
  • DV: Annual income.
• Method: Survey 10,000 individuals, analyze correlation.
• Limitation: Cannot prove causation (e.g., smarter people may earn more, not just education).

3. Key Research Design Models

A. Survey Research Design

• Definition: Uses questionnaires or interviews to collect data.
• Types:
  • Cross-sectional survey (one-time data).
  • Longitudinal survey (repeated data).
• Example:
  • Daraz conducts customer satisfaction surveys after purchases to improve service.

graph TD
    A["Survey Research"] --> B["1. Define Objectives"]
    A --> C["2. Choose Survey Type"]
    C --> C1["Cross-sectional"]
    C --> C2["Longitudinal"]
    A --> D["3. Design Questionnaire"]
    D --> D1["Close-ended (MCQ, Likert scale)"]
    D --> D2["Open-ended (essay, comments)"]
    A --> E["4. Sample Selection"]
    A --> F["5. Data Collection"]
    A --> G["6. Analysis"]

B. Experimental Research Design

• Definition: Manipulates variables in a controlled setting.
• Types:
  • Pre-experimental (weak control, e.g., one-group post-test).
  • True experimental (random assignment, control group).
  • Quasi-experimental (no random assignment, e.g., natural groups).
• Example:
  • Nepal Investment Bank tests if offering 0% interest loans increases home loan applications.
    • Group A: 0% interest for 6 months.
    • Group B: Normal interest rate.
    • Result: If Group A applies 40% more, the policy is effective.

C. Case Study Design

• Definition: In-depth analysis of a single entity (company, event, individual).
• Example:
  • Studying Nepal’s COVID-19 vaccine rollout to understand challenges.
  • Chaudhary Group’s expansion strategy in the FMCG sector.

mindmap
  root((Case Study Design))
    Types
      Single Case
      Multiple Cases
    Steps
      Define Research Question
      Select Case(s)
      Collect Data (Interviews, Documents, Observations)
      Analyze Data
      Draw Conclusions
    Example
      "Nepal’s E-commerce Boom (Daraz vs. Hamrobazaar)"

4. Choosing the Right Research Design

Selecting a design depends on:

1. Research objective (exploratory, descriptive, causal).
2. Budget and time (cross-sectional is cheaper than longitudinal).
3. Ethical considerations (experiments may require consent).
4. Data availability (existing data vs. new collection).

DECISION TREE:

flowchart TD
    A["Start"] --> B["Is the goal to explore a new topic?"]
    B -->|"Yes"| C["Use Exploratory Design"]
    B -->|"No"| D["Is the goal to describe characteristics?"]
    D -->|"Yes"| E["Use Descriptive Design"]
    D -->|"No"| F["Is causation needed?"]
    F -->|"Yes"| G["Use Experimental Design"]
    F -->|"No"| H["Use Non-experimental Design"]

In the Real World

1. eSewa & Khalti (Digital Payments)
   • Design Used: Longitudinal & Experimental
   • How?
     • Longitudinal: Track transaction trends over years to identify growth patterns.
     • Experimental: Test discount incentives (e.g., "Pay with Khalti, get 10% off") to see if it boosts usage.

2. Daraz (E-commerce)

   • Design Used: Cross-sectional Surveys & A/B Testing (Experimental)
   • How?
     • Surveys: Ask customers about shopping preferences (e.g., "Do you prefer COD or digital payments?").
     • A/B Testing: Test two website layouts to see which increases sales.

3. Nepal Electricity Authority (NEA)

   • Design Used: Descriptive & Longitudinal
   • How?
     • Descriptive: Publish monthly electricity consumption reports.
     • Longitudinal: Track load shedding patterns over decades to plan infrastructure.

4. Ncell (Telecom)

   • Design Used: Case Study & Survey
   • How?
     • Case Study: Analyze why Ncell lost market share to NTC in rural areas.
     • Surveys: Ask customers about network reliability complaints.

5. Nepal Stock Exchange (NEPSE)

   • Design Used: Time-Series Analysis (Longitudinal)
   • How?
     • Track stock prices of Himalayan Java over 10 years to predict trends.

Exam Tip

1. Understand the difference between exploratory, descriptive, and causal designs—examiners often ask when to use each.
2. Memorize the advantages/disadvantages of cross-sectional vs. longitudinal designs (e.g., cost, time, depth).
3. Practice matching real-world scenarios to designs (e.g., "How would Daraz test a new feature?" → Experimental/A/B Test).
4. Case studies are common in exams—be ready to analyze a Nepali company’s research approach (e.g., "How did Nabil Bank study loan defaults?").
5. Diagrams save marks! Draw flowcharts for research processes or tables comparing designs in exams.

KEY FORMULA (For Hypothesis Testing in Experimental Design): $$\text{Effect Size}=\frac{{\text{Mean}}_{\text{Treatment Group}}-{\text{Mean}}_{\text{Control Group}}}{\text{Standard Deviation}}$$ (Used to measure the impact of an intervention.)