Research Design: Types, Models & Selection Criteria

What is Research Design?

Research design is the blueprint of a study—it outlines how you will answer your research question. It includes:

• Structure: The framework (e.g., surveys, experiments, case studies).
• Methods: Tools and techniques (e.g., questionnaires, observations, statistical tests).
• Logic: The reasoning behind your approach (e.g., why use surveys over interviews?).

A well-designed study ensures validity (measuring what you intend) and reliability (consistent results). Poor design leads to wasted time, money, and unusable data.

Why Does Research Design Matter?

Real-world analogy:

• NTC’s customer satisfaction survey: Uses a descriptive design to quantify complaints about service delays.
• Pathao’s driver app updates: Tests new features via experimental design (A/B testing) to see which increases ride bookings.
• Nepal Tourism Board’s heritage site study: Combines qualitative interviews (local guides) with quantitative surveys (tourist feedback) for a mixed-methods design.

Types of Research Design

Research designs are categorized based on purpose and nature of data. The three primary types are:

1. Exploratory Design

Definition: Used when the research problem is broad or unclear. Goal: generate insights, not final answers. When to use:

• New or complex topics (e.g., "Why are Nepali tourists shifting to domestic travel?").
• Lack of prior research.
• Need to define variables before deeper study.

Example: Problem: Why do fewer tourists visit Mustang in winter? Design: Exploratory (interviews with tour operators, literature review on climate trends). Outcome: Hypothesis generated → "Harsh weather reduces trekking permits by 30%."

Advantages:

• Flexible, adaptable.
• Reveals hidden patterns.
• Low cost for initial studies.

Disadvantages:

• No definitive answers.
• Risk of bias (small sample sizes).
• Time-consuming.

2. Descriptive Design

Definition: Quantifies characteristics of a population or situation. Focuses on "what is" rather than "why." When to use:

• Measuring trends (e.g., tourist arrivals, revenue).
• Profiling groups (e.g., demographics of cruise tourists).
• Testing hypotheses (e.g., "Do 60% of tourists prefer online bookings?").

Example: Problem: What factors influence tourist satisfaction at Kathmandu’s hotels? Design: Descriptive (survey of 500 guests using Likert scales). Tools: Closed-ended questions, statistical analysis (mean scores).

Advantages:

• Precise, measurable data.
• Easy to replicate.
• Useful for policy-making.

Disadvantages:

• Cannot establish causality.
• Limited to existing variables.
• Survey fatigue (low response rates).

3. Explanatory Design

Definition: Explains relationships between variables. Answers "why" or "how." When to use:

• Testing theories (e.g., "Does social media marketing increase bookings?").
• Understanding root causes (e.g., "Why did tourist arrivals drop in 2020?").
• Policy recommendations (e.g., "How does airport infrastructure affect tourism?").

Example: Problem: Why do Nepali students prefer study abroad over local universities? Design: Explanatory (mixed methods: surveys + interviews with alumni). Findings: "Lack of English-medium programs and perceived better job prospects abroad."

Advantages:

• Deep insights into causality.
• Supports evidence-based decisions.
• Combines qualitative and quantitative data.

Disadvantages:

• Complex and time-intensive.
• Requires skilled researchers.
• Ethical challenges (e.g., sensitive topics).

4. Experimental Design

Definition: Tests cause-and-effect under controlled conditions. Uses manipulation of variables. When to use:

• Marketing tests (e.g., "Does a 20% discount increase Daraz sales?").
• Policy interventions (e.g., "Does a new visa process reduce wait times?").
• Product development (e.g., "Does a new hotel app improve check-in speed?").

Key Components:

1. Independent Variable (IV): What you change (e.g., discount percentage).
2. Dependent Variable (DV): What you measure (e.g., sales volume).
3. Control Group: No treatment (e.g., normal pricing).
4. Experimental Group: Treatment applied (e.g., 20% discount).

Example: Problem: Does offering free Wi-Fi increase restaurant revenue? Design: Experimental (A/B test in two similar restaurants). Steps:

• Control Group: Restaurant A (no free Wi-Fi).
• Experimental Group: Restaurant B (free Wi-Fi + survey on usage). Result: "Revenue increased by 15% in Group B, with 70% of diners using Wi-Fi >2 hours."

Advantages:

• Strongest evidence for causality.
• Replicable and objective.
• Useful for business decisions.

Disadvantages:

• Artificial settings (may not reflect real world).
• Ethical concerns (e.g., withholding benefits from control group).
• Expensive and time-consuming.

Research Models: Qualitative vs. Quantitative vs. Mixed

Research designs also differ by data type. Choose based on your research question.

1. Qualitative Design

Definition: Explores subjective experiences, opinions, or social phenomena. Focuses on words, not numbers. Key Features:

• Flexible: Questions evolve during research.
• Context-rich: Understands why behind behaviors.
• Small samples: 5–30 participants (deep insights).

Methods:

• Interviews: One-on-one or group (e.g., with hotel managers).
• Focus Groups: Discussions with 6–10 people (e.g., tourist guides).
• Case Studies: In-depth analysis of a single entity (e.g., a heritage site).
• Observations: Watching behaviors (e.g., how tourists interact at Swayambhunath).

Example: Problem: How do local communities perceive the impact of tourism in Pokhara? Design: Qualitative (6 focus groups with residents, shopkeepers, and monks). Findings:

• "Tourism brings income but also traffic and noise pollution."
• "Monks feel cultural erosion from foreign visitors."

Advantages:

• Reveals hidden motivations.
• Adaptable to new insights.
• Useful for exploratory research.

Disadvantages:

• Subjective (bias risk).
• Hard to generalize.
• Time-consuming analysis.

2. Quantitative Design

Definition: Uses numeric data to test hypotheses or measure trends. Focuses on objectivity and generalization. Key Features:

• Structured: Predefined questions (e.g., Likert scales).
• Large samples: 100+ respondents for reliability.
• Statistical analysis: Tests correlations, averages, etc.

Methods:

• Surveys: Closed-ended questions (e.g., "Rate your experience: 1–5").
• Experiments: Controlled tests (e.g., A/B testing ads).
• Secondary data: Existing statistics (e.g., NEPSE tourism revenue reports).

Example: Problem: What percentage of tourists use online booking platforms? Design: Quantitative (survey of 500 tourists at Tribhuvan International Airport). Questions:

1. "Do you book hotels online? (Yes/No)"
2. "Which platform do you use? (List options)" Result: "68% use online platforms; 40% prefer Agoda."

Advantages:

• Objective and replicable.
• Easy to analyze with software (SPSS, Excel).
• Supports large-scale decisions.

Disadvantages:

• Limited depth (misses why).
• Risk of low response rates.
• May not capture complex behaviors.

3. Mixed-Methods Design

Definition: Combines qualitative + quantitative to get a complete picture. When to use:

• Complex problems needing both depth and breadth.
• Triangulation (cross-checking data sources).
• Policy recommendations (e.g., "Why are tourist arrivals declining?").

Example: Problem: Why do Nepali tourists prefer domestic destinations over foreign trips? Design: Mixed-methods.

1. Quantitative: Survey 300 tourists (ranking preferences).
2. Qualitative: Interviews with 10 frequent travelers. Findings:

• Quantitative: 70% prefer domestic due to cost.
• Qualitative: "Foreign trips are stressful; domestic is safer."

Advantages:

• Balances strengths of both models.
• Richer, more nuanced insights.
• Higher validity.

Disadvantages:

• Complex to design and analyze.
• Time-consuming and costly.
• Requires mixed-methods expertise.

How to Choose the Right Research Design

Use this decision tree to select the best approach:

flowchart TD
    A["Start: What is your research question?"] --> B{"Is the problem new or unclear?"}
    B -->|"Yes"| C["Use Exploratory Design"]
    B -->|"No"| D{"Do you need to explain relationships?"}
    D -->|"Yes"| E["Use Explanatory or Experimental"]
    D -->|"No"| F{"Do you need numbers or words?"}
    F -->|"Numbers"| G["Quantitative"]
    F -->|"Words"| H["Qualitative"]
    F -->|"Both"| I["Mixed-Methods"]

Key Questions to Ask:

1. Purpose: Explore, describe, explain, or test?
2. Data type: Numbers, words, or both?
3. Resources: Time, budget, and expertise?
4. Ethics: Can you manipulate variables (experimental) or must you observe naturally?

In the Real World

1. eSewa & Khalti (Digital Payments)

   • Design: Experimental + Quantitative
   • How: Tested if QR-based payments increase transaction speed vs. traditional methods.
   • Result: "QR payments reduced checkout time by 40% in pilot stores."

2. Daraz (E-Commerce)

   • Design: Mixed-Methods
   • How: Combined A/B testing (quantitative: sales data) with customer interviews (qualitative: feedback on new features).
   • Example: "Live chat support increased conversions by 25% (quantitative), and users loved 24/7 help (qualitative)."

3. NTC (Transportation)

   • Design: Descriptive + Qualitative
   • How: Surveyed passengers on bus delays (quantitative) and interviewed drivers on route issues (qualitative).
   • Outcome: "Peak-hour delays caused 60% dissatisfaction; drivers cited poor road maintenance."

4. Nepal Tourism Board (Policy)

   • Design: Explanatory + Mixed
   • How: Studied the impact of COVID-19 on tourism using secondary data (quantitative) and expert interviews (qualitative).
   • Finding: "70% of hotels closed; local guides lost 80% income, but digital marketing helped some recover."

5. Pathao (Ride-Hailing)

   • Design: Experimental
   • How: Tested a "surge pricing" feature in Kathmandu vs. a control city (Lalitpur).
   • Result: "Surge pricing increased driver sign-ups by 35% during peak hours."

Worked Example: Designing a Study on "Impact of Social Media on Tourism Marketing"

Problem: How does Instagram marketing affect hotel bookings in Pokhara? Steps:

1. Define Variables:

   • IV: Hotel’s Instagram activity (posts/week).
   • DV: Booking rates (measured via hotel records).

2. Choose Design:

   • Experimental (A/B test) + Quantitative (survey bookings).
   • Why not qualitative? → Need measurable data for causality.

3. Methodology:

   • Group 1 (Control): 5 hotels with no Instagram (or minimal activity).
   • Group 2 (Experimental): 5 hotels posting 3x/week (photos, reels, promotions).
   • Data Collection: Track bookings for 3 months; survey guests on discovery method.

4. Expected Outcome:

   • "Hotels in Group 2 saw a 22% increase in bookings, with 40% of guests citing Instagram as their source."

5. Real-World Application:

   • Nepal Tourism Board could use this to train hotels on social media strategies.

Common Mistakes to Avoid

1. Mismatched Design: Using qualitative for a problem needing quantitative data (e.g., interviewing 5 people to measure national tourism trends).
2. Ignoring Ethics: Experimental designs must avoid harm (e.g., denying discounts to a control group without justification).
3. Overcomplicating: Mixed-methods are powerful but unnecessary for simple descriptive studies.
4. Small Samples: Quantitative studies need >100 respondents for reliability; qualitative needs depth, not size.
5. Leading Questions: In surveys, avoid bias (e.g., "Don’t you agree tourism harms culture?" → Use neutral phrasing).

Exam Tip

How This Unit is Tested

1. Definitions & Differences:

   • Expect questions like: "Distinguish between exploratory and descriptive research designs with examples." Answer: Use a comparison table (as above) and cite NTC’s survey (descriptive) vs. a new trekking route study (exploratory).

2. Application Questions:

   • "Design a study to measure the impact of flight delays on tourist satisfaction at Tribhuvan Airport." Answer:
     • Type: Descriptive/explanatory.
     • Model: Quantitative (survey) + qualitative (interviews with delayed passengers).
     • Tools: Likert scale for satisfaction + open-ended questions on causes.

3. Case Study Analysis:

   • "Evaluate the research design used in a study on ‘Why do Nepali students prefer study abroad?’" Answer:
     • Identify if it’s mixed-methods (likely).
     • Critique: "Strengths: Combines survey data (quantitative) with alumni interviews (qualitative). Weakness: Small interview sample (n=10) limits generalizability."

4. Diagram-Based Questions:

   • "Draw a flowchart showing the steps of an experimental design." Answer: Use the A/B testing example (control vs. experimental groups) with labeled arrows.

5. Real-World Scenarios:

   • "How would Khalti use an experimental design to improve mobile payments?" Answer:
     • IV: Payment method (QR vs. PIN).
     • DV: Transaction success rate.
     • Control: PIN-only users; Experimental: QR users.
     • Outcome: "QR reduced errors by 15%."

Quick Revision Checklist

• Can you list the 4 types of research designs and give a Nepali example for each?
• What’s the difference between qualitative and quantitative data? Draw a table.
• How would you design a study on "Impact of traffic on tourist experiences in Kathmandu"? (Hint: Mixed-methods!)
• What are the 3 key components of an experimental design? (IV, DV, control group).
• Name 2 Nepali companies using experimental designs and explain how.

Final Model Answer for Exam Questions

Question: "Explain the types of research designs with suitable examples from the tourism sector in Nepal." Model Answer:

Research designs are categorized based on their purpose and data type. The four primary types are:

1. Exploratory Design

   • Purpose: Generate insights for new or vague problems.
   • Example: Studying "Why are eco-tourism packages underutilized in Annapurna?"
     • Methods: Literature review + interviews with 10 local guides.
     • Outcome: Hypothesis → "Lack of marketing and high costs deter tourists."

2. Descriptive Design

   • Purpose: Quantify characteristics or trends.
   • Example: "Measuring tourist satisfaction at Pokhara’s hotels."
     • Methods: Survey of 300 guests using a 5-point Likert scale.
     • Tools: Closed-ended questions (e.g., "How likely are you to return?").

3. Explanatory Design

   • Purpose: Explain why or how phenomena occur.
   • Example: "Why did tourist arrivals drop by 40% in 2020?"
     • Methods: Mixed-methods (quantitative: arrival data; qualitative: interviews with tour operators).
     • Finding: "COVID-19 restrictions + economic uncertainty caused decline."

4. Experimental Design

   • Purpose: Test cause-and-effect relationships.
   • Example: "Does offering free city tours increase hotel bookings?"
     • IV: Free tour offer (yes/no).
     • DV: Booking rates.
     • Result: "Hotels with free tours saw 18% more bookings."

Visual Summary:

Question: "Differentiate between qualitative and quantitative research models with reference to a Nepali case study." Model Answer:

Case Study: Nepal Tourism Board’s Post-COVID Recovery

• Qualitative: Conducted focus groups with 8 heritage site managers to understand challenges (e.g., "Lack of foreign tourists hurt our income").
• Quantitative: Released a survey finding "65% of hotels reopened by 2022, but 30% faced bankruptcy."

Why Mixed? The Board combined both to explain (qualitative) why recovery was slow (quantitative data on revenue drops) and measure the extent of the problem.

Key Formulas & Checklists

1. Sample Size Rule of Thumb:

   • Quantitative: ≥100 for reliability.
   • Qualitative: Depth > size (e.g., 10 in-depth interviews).

2. Experimental Design Checklist:

   • Clearly defined IV and DV.
   • Randomized assignment to control and experimental groups.
   • Ethical approval (if manipulating variables).

3. Survey Design Tips:

   • Use Likert scales (1–5) for quantitative data.
   • Avoid leading questions (e.g., "Don’t you think traffic is terrible?").
   • Pilot-test questions with 5–10 people first.