Research Designs & Methods: Types, Ethics & Applications

Core Concepts: What is Research Design?

Research design is the blueprint of your study—it determines how you collect, analyze, and interpret data to answer your research question. A well-structured design ensures validity, reliability, and ethical compliance.

1. Types of Research Designs

Research designs are classified based on purpose, time, and control. Below is a Mermaid classification of the five primary designs covered in TU/PU syllabi:

mindmap
  root((Research Designs))
    Exploratory
      "Purpose: Gain insight, define problems"
      "Methods: Pilot studies, literature reviews, case studies"
      "Example: Studying public perception of eSewa failures"
    Descriptive
      "Purpose: Describe characteristics, trends"
      "Methods: Surveys, observations, cross-sectional studies"
      "Example: NTC’s customer satisfaction survey"
    Explanatory
      "Purpose: Explain *why* something happens"
      "Methods: Experiments, longitudinal studies, causal analysis"
      "Example: Pathao’s driver retention study (why drivers quit)"
    Evaluative
      "Purpose: Assess program effectiveness"
      "Methods: Pre-test/post-test, control group studies"
      "Example: NEPSE’s stock market impact analysis"
    Predictive
      "Purpose: Forecast future trends"
      "Methods: Time-series analysis, regression models"
      "Example: Ncell’s customer churn prediction model"

## In the Real World

Research designs shape how companies, governments, and NGOs solve problems. Here’s how Nepali and global firms apply them:

## 1. Exploratory Research Design

Definition: Used when little is known about a topic. Aims to define problems, generate hypotheses, or identify variables.

How It Works

1. Purpose: Discover ideas, clarify concepts, or develop research questions.
2. Methods:
   • Literature reviews (secondary data).
   • Pilot studies (small-scale tests).
   • Case studies (deep dive into one instance).
   • Focus group discussions (qualitative insights).
3. Example:
   • Problem: Why do students in Pokhara University struggle with online exams?
   • Method: Conduct focus groups with 10 students to explore challenges (e.g., internet issues, stress).
   • Outcome: Identifies three key issues → leads to a larger descriptive study.

Advantages & Disadvantages

## 2. Descriptive Research Design

Definition: Describes characteristics of a population, situation, or phenomenon without manipulating variables.

Key Features

• No cause-effect relationships (only "what" and "how much").
• Uses surveys, observations, or existing data.
• Example in Nepal:
  • NTC’s Customer Satisfaction Survey:
    • Method: 5,000 users rated service on a scale of 1–5.
    • Findings: 60% reported slow speeds in hilly regions.
    • Action: NTC expanded 4G towers in those areas.

Types of Descriptive Studies

pie
  title Descriptive Research Types
  "Cross-Sectional" : 40
  "Longitudinal" : 30
  "Case Study" : 20
  "Survey" : 10

When to Use It?

✅ Market research (e.g., Daraz’s customer preferences). ✅ Social trends (e.g., youth unemployment rates in Nepal). ✅ Program evaluation (e.g., impact of a government scholarship).

## 3. Explanatory Research Design

Definition: Explains the relationship between variables (cause-and-effect). Uses experiments or controlled studies.

How It Differs from Descriptive

Worked Example: Pathao’s Driver Retention

• Research Question: Why do Pathao drivers quit within 6 months?
• Method:
  1. Survey 500 drivers (descriptive).
  2. Experiment: Offer bonus incentives to half the drivers (IV) and track quit rates (DV).
• Result:
  • Control group (no bonus): 40% quit rate.
  • Bonus group: 20% quit rate.
• Conclusion: Financial incentives reduce turnover by 50%.

## 4. Evaluative Research Design

Definition: Assesses the effectiveness of a program, policy, or intervention.

Steps in Evaluative Research

1. Define objectives (e.g., "Reduce student dropout rates by 20%").
2. Baseline data (pre-test).
3. Implement intervention (e.g., free tutoring for weak students).
4. Post-test (compare results).
5. Analyze impact (e.g., dropout rate dropped from 30% to 10%).

Example: NEPSE’s Stock Market Education Program

• Goal: Improve investor literacy.
• Method:
  • Pre-test: Survey 1,000 investors on trading knowledge.
  • Workshop: Train 500 investors (treatment group).
  • Post-test: Re-survey after 6 months.
• Result: Treatment group scored 30% higher in trading knowledge.

## 5. Predictive Research Design

Definition: Forecasts future trends using historical data, statistical models, or AI.

Methods

• Time-series analysis (e.g., Ncell’s call volume during festivals).
• Regression models (e.g., predicting loan defaults).
• Machine learning (e.g., Khalti’s fraud detection).

Example: Ncell’s Network Congestion Prediction

• Data Used: Past 5 years of call drop rates during Dashain/Tihar.
• Model: Linear regression predicts peak hours (e.g., 6–9 PM).
• Outcome: Ncell pre-allocates servers, reducing drops by 40%.

## Sampling Techniques: How to Select Participants

Sampling ensures your research is representative, feasible, and ethical. Below is a comparison table of key methods:

Shows population divided into strata (e.g., age groups, regions). (Image: Dan Kernler, CC BY-SA 4.0, via Wikimedia Commons)

## Ethical Considerations in Research Design

Ethics ensures participant rights, validity, and social responsibility. Key principles:

1. Informed Consent: Participants must knowingly agree to take part.
   • Example: NTC’s survey includes a clear consent form before data collection.
2. Anonymity/Confidentiality: Protect identities.
   • Example: Khalti’s user data is encrypted to prevent leaks.
3. Avoiding Harm: No physical/psychological damage.
   • Example: A study on child labor must not exploit participants.
4. Honesty & Transparency: Disclose funding sources, conflicts of interest.
   • Example: TU research must declare if funded by NGOs or corporations.

## Common Biases in Research Design

Biases distort results. Recognize these in experimental and survey-based studies:

## Exam Tip: How to Score Full Marks

1. Define Clearly: Always start with a precise definition of the research design (e.g., "Descriptive research aims to measure and describe variables without manipulation...").
2. Use Real Examples: Link theories to Nepali cases (e.g., NTC surveys, Daraz delivery studies).
3. Compare & Contrast: Use tables to differentiate designs (e.g., descriptive vs. explanatory).
4. Ethics & Bias: Examiners love critical analysis—discuss how biases could affect a study (e.g., "Convenience sampling in Kathmandu may overrepresent urban views").
5. Visuals: Draw diagrams (even rough sketches) for:
   • Research designs (flowcharts).
   • Sampling methods (Venn diagrams for strata).
   • Ethical checklists (tables).

Past Exam Question Analysis:

"Describe Bern’s (1974) study of masculinity and femininity. How can you utilize this in future research?"

• Answer Structure:
  1. Briefly summarize Bern’s study (gender role socialization via observational learning).
  2. Link to research design: Used case studies + surveys (descriptive + explanatory).
  3. Application: For your thesis, use mixed methods (e.g., survey + interviews on gender roles in Nepali workplaces).

## Summary Checklist

Before submitting your answer, ensure you’ve covered: ✅ All 5 research designs (exploratory, descriptive, explanatory, evaluative, predictive). ✅ Sampling methods (random, stratified, convenience) with Nepali examples. ✅ Ethical principles (consent, anonymity, avoiding harm). ✅ At least 3 real-world applications (eSewa, Khalti, NTC, etc.). ✅ Visuals (diagrams for designs, tables for comparisons, images for real objects). ✅ Exam tips (how to structure answers for full marks).