Data Collection Methods: Primary vs. Secondary, Tools & Techniques

TAKEAWAYS:

• Primary vs. Secondary Data: Know the sources, costs, and reliability trade-offs (e.g., eSewa’s customer feedback surveys vs. NTC’s historical call data).
• Data Collection Tools: Surveys, interviews, observations, and experiments—each has pros/cons (e.g., Daraz’s online questionnaires vs. Pathao’s rider observations).
• Secondary Data Sources: Internal (company reports) and external (government stats, NEPSE reports) with their limitations.
• Ethics & Validity: Ensure data is unbiased, representative, and legally obtained (e.g., Ncell’s customer data privacy policies).
• Real-World Application: Link theories to cases like Khalti’s fraud detection (using transaction data) or Nabil Bank’s loan approvals (credit history analysis).
• Exam Focus: Differentiate methods, list sources, and justify choices (e.g., "Why use primary data for a new product launch?").

---

Primary vs. Secondary Data: The Core Distinction

Data collection methods are categorized into primary (original, firsthand) and secondary (existing, pre-collected). The choice depends on cost, time, reliability, and research goals.

1. Primary Data: Fresh, Tailored, but Costly

Primary data is collected specifically for the research problem at hand. It is original and customized to the study’s needs.

Sources of Primary Data:

• Surveys/Questionnaires: Structured or unstructured (e.g., eSewa’s customer satisfaction surveys after a transaction).
• Interviews: Face-to-face, phone, or online (e.g., Daraz conducting interviews with sellers about logistics challenges).
• Observations: Direct or indirect (e.g., NTC observing network congestion in Kathmandu).
• Experiments: Controlled tests (e.g., a bank testing new ATM interface usability with customers).
• Focus Groups: Group discussions (e.g., Pathao gathering rider feedback on delivery delays).

Advantages: ✅ Highly relevant to the research question. ✅ Controlled (researcher designs the data collection). ✅ Up-to-date (no lag in information).

Disadvantages: ❌ Time-consuming and expensive. ❌ Prone to bias (e.g., survey design flaws). ❌ Limited sample size (hard to generalize).

Worked Example: eSewa’s Customer Feedback System eSewa collects primary data via post-transaction surveys to measure user satisfaction. Their 5-point Likert scale (e.g., "How satisfied were you with the payment process?") ensures quantifiable feedback. Why primary? Secondary data (e.g., call center logs) wouldn’t capture real-time user emotions or new pain points like "slow OTP delivery."

---

---

2. Secondary Data: Quick but Questionable

Secondary data is already available from internal or external sources. It’s faster and cheaper but may lack relevance or accuracy.

Sources of Secondary Data:

Advantages: ✅ Cost-effective (no new data collection). ✅ Time-saving (instant access). ✅ Broader scope (e.g., CBS’s national income data).

Disadvantages: ❌ May not fit the research question. ❌ Outdated (e.g., a 2018 report on Kathmandu traffic won’t reflect 2024 congestion). ❌ Bias or errors (e.g., a competitor’s misleading sales data).

Worked Example: NEPSE’s Stock Analysis A researcher studying NEPSE’s market trends uses secondary data from NEPSE’s website and CBS reports on GDP growth. Why secondary? Primary data would require interviewing every investor—impossible. However, the researcher must cross-validate with primary surveys to confirm trends.

---

---

3. Data Collection Tools: How to Gather Data

Each tool has unique strengths for different research goals.

Key Decision Factors:

• Quantitative vs. Qualitative: Use surveys for numbers (e.g., "How many times did you use Daraz this month?") and interviews for opinions (e.g., "Why did you abandon your cart?").
• Budget: Secondary data (e.g., CBS reports) is free; primary tools like experiments cost more.
• Time: Need fast results? Use secondary data. Need precision? Use primary.

---

---

4. Real-World Applications: Where These Methods Shine

Case 1: Daraz’s Inventory Management

Problem: Daraz needs to predict demand for monsoon season products. Solution:

• Primary Data: Conducts surveys with customers ("What monsoon essentials do you buy?") and observes past purchase patterns.
• Secondary Data: Uses CBS population data and competitor (Amazon India) sales trends. Outcome: Reduces overstocking by 20% using a mix of both data types.

Case 2: Ncell’s Network Optimization

Problem: Ncell faces call drops in busy areas like Thamel. Solution:

• Primary Data: Observes network traffic via engineers’ tools and interviews customers about drop locations.
• Secondary Data: Uses NTC’s infrastructure reports and weather data (rain affects signals). Outcome: Identifies Thamel’s narrow streets as a bottleneck and installs repeaters.

Case 3: Khalti’s Fraud Detection

Problem: Khalti wants to reduce fake transactions. Solution:

• Primary Data: Experiments with AI to flag unusual patterns (e.g., sudden large transfers).
• Secondary Data: Uses NRA’s fraud databases and user complaint logs. Outcome: Catches 30% more frauds by combining real-time monitoring (primary) with historical fraud data (secondary).

---

5. Ethical and Practical Considerations

• Privacy: Never collect personal data without consent (e.g., Ncell’s telecom privacy policy).
• Bias: Avoid leading questions in surveys (e.g., "Don’t you love Daraz’s fast delivery?").
• Legality: Some data is restricted (e.g., bank customer records under Nepal Rastra Bank’s guidelines).
• Validity: Ensure data is reliable (e.g., cross-check NEPSE stock data with primary trader interviews).

---

Exam Tip: How to Score Full Marks

1. Differentiate Clearly:
   • Primary vs. Secondary: Always compare source, cost, and relevance. Example Answer:

     "Primary data is collected firsthand (e.g., eSewa’s customer surveys), while secondary data is pre-existing (e.g., CBS’s population reports). Primary is costly but precise; secondary is cheap but may lack timeliness."

2. List Sources with Examples:

   • For secondary data, name at least 3 sources with Nepali/global examples. Example Answer:

     *"Secondary data sources include:

     • Internal: Nabil Bank’s loan records
     • External: CBS’s national income data
     • Digital: NEPSE’s stock price history."*

3. Justify Your Choice:

   • If asked "Why use primary data for X?", explain why existing data is insufficient. Example Answer:

     "For studying Pathao’s rider stress levels, primary data (interviews/observations) is essential because secondary sources like call logs don’t capture emotional stress—only call volume."

4. Avoid Common Mistakes:

   • ❌ "Surveys are always better than interviews." → Wrong! Interviews give deeper insights for qualitative research.
   • ❌ "Secondary data is always outdated." → Partial truth; some sources (e.g., NEPSE daily reports) are real-time.

5. Use Diagrams in Exams:

   • Draw a simple flowchart (like the one above) to explain data collection steps for 2–3 marks.

---

Quick Revision Table

---

Final Challenge: Apply What You Learned

Scenario: You’re researching "Why do students in Pokhara prefer online shopping over physical stores?"

1. What primary data would you collect? (List 2 tools).
2. What secondary data would help? (List 2 sources).
3. Why not rely solely on secondary data?

Answer Key:

1. Primary:
   • Surveys: "How often do you shop online vs. in-store?"
   • Interviews: "What factors influence your choice?"
2. Secondary:
   • CBS’s internet penetration data in Pokhara.
   • Daraz’s sales reports for Pokhara region.
3. Reason: Secondary data (e.g., CBS reports) explains market trends but not student-specific reasons like "convenience" or "discounts." Primary data captures personal motivations.