Data Collection Methods: Techniques, Tools & Ethical Use

Core Concepts

1. Definition & Purpose

Data collection methods are systematic techniques used to gather information for research. They can be classified into primary (original data collected by the researcher) and secondary (existing data sourced from other studies or records).

mindmap
  root((Data Collection Methods))
    Primary Data
      Surveys
      Interviews
      Observations
      Experiments
      Focus Groups
    Secondary Data
      Internal Sources (Company Records)
      External Sources (Government Reports, Academic Papers)

Why it matters?

• Primary data is customized but time-consuming and costly.
• Secondary data is quick and cheap but may lack relevance or accuracy.

Primary Data Collection Methods

1. Surveys

Definition: Structured questionnaires administered to a sample population to collect quantitative or qualitative data.

Tools:

• Questionnaires (paper/pencil, online via Google Forms, SurveyMonkey).
• Sampling frames (lists of potential respondents, e.g., voter lists for political surveys).

Example: eSewa’s Customer Satisfaction Survey

• Method: Online questionnaire sent via SMS/email.
• Question Type: Likert scale (e.g., "How satisfied are you with eSewa’s payment speed? 1-5").
• Analysis: Used to improve transaction success rates.

Advantages/Disadvantages:

2. Interviews

Definition: Direct, interactive conversations (face-to-face, phone, or video) to gather in-depth responses.

Types:

• Structured: Fixed questions (e.g., bank loan officer interviews).
• Unstructured: Open-ended (e.g., focus groups for new product ideas).
• Semi-structured: Mix of both (common in academic research).

Example: Nabil Bank’s Loan Applicant Interview

• Method: Semi-structured interview with a loan officer.
• Question: "What challenges do you face in repaying loans?"
• Purpose: Assess creditworthiness beyond financial data.

Advantages/Disadvantages:

3. Observations

Definition: Systematic recording of behavior or events without direct interaction.

Types:

• Participant Observation: Researcher joins the group (e.g., studying employee behavior in a call center).
• Non-Participant Observation: Researcher remains detached (e.g., watching traffic flow at a Kathmandu intersection).

Example: Pathao Driver Behavior Study

• Method: Non-participant observation via GPS tracking + driver logs.
• Finding: 30% of delays caused by inefficient route planning.
• Solution: AI-powered route optimization.

Advantages/Disadvantages:

4. Experiments

Definition: Manipulating variables to test cause-and-effect relationships.

Example: Daraz’s A/B Testing for Discount Banners

• Hypothesis: Red banners increase click-through rates (CTR) more than blue.
• Method: Split traffic between two groups (Group A: red banner; Group B: blue).
• Result: Red banners increased CTR by 18% → adopted as standard.

Key Components:

flowchart TD
  A["Independent Variable<br/>(e.g., banner color)"] --> B["Dependent Variable<br/>(e.g., CTR)"]
  B --> C["Control Variables<br/>(e.g., same product, same time)"]
  B --> D["Random Assignment<br/>(e.g., users split randomly)"]

Advantages/Disadvantages:

5. Focus Groups

Definition: Group discussions (6–10 participants) moderated to explore opinions on a topic.

Example: Himalayan Java’s New Tea Blend Launch

• Method: Focus group with regular customers in Pokhara.
• Question: "What flavors would make you switch from your current tea?"
• Outcome: Led to the introduction of cardamom-infused green tea.

Advantages/Disadvantages:

Secondary Data Collection Methods

1. Internal Sources

Definition: Data already available within the organization (e.g., sales records, customer databases).

Example: NTC’s Network Performance Analysis

• Source: Internal call drop logs + customer complaint records.
• Finding: 40% of drops occur in Bhaktapur → targeted tower upgrades.

2. External Sources

Definition: Data from outside the organization (government reports, academic journals, industry databases).

Example: NEPSE’s Stock Market Research

• Source: Secondary data from NEPSE’s annual reports + global economic trends.
• Use: Predicts stock price movements for investors.

Comparison Table:

In the Real World

1. eSewa’s Data Collection for Fraud Detection

   • Method: Combines secondary data (transaction history) with primary data (real-time biometric verification).
   • How it works: Uses machine learning to flag unusual patterns (e.g., multiple failed logins from different locations).
   • Impact: Reduced fraud cases by 25% in 2023.

2. Daraz’s Supplier Feedback System

   • Method: Mixed-mode surveys (online for urban suppliers, phone interviews for rural ones).
   • Insight: Identified that 60% of delays were due to last-mile logistics → partnered with Pathao for faster deliveries.

3. Nabil Bank’s Loan Default Prediction

   • Method: Experiments (simulated loan scenarios) + secondary data (credit bureau reports).
   • Outcome: Reduced default rates by 15% by adjusting interest rates dynamically.

Tools & Technologies

1. Digital Platforms

• Google Forms/SurveyMonkey: For online surveys.
• Zoom/Teams: For virtual interviews and focus groups.
• Tableau/Power BI: For visualizing secondary data (e.g., NTC’s network heatmaps).

2. Sampling Tools

• Random Sampling: Ensures unbiased selection (e.g., NEPSE’s investor surveys).
• Stratified Sampling: Divides population into subgroups (e.g., Daraz’s urban vs. rural customers).

Example: Khalti’s User Experience Study

• Method: Stratified random sampling (20% urban, 30% semi-urban, 50% rural users).
• Tool: Online survey via Khalti’s app notification.

Ethical Considerations

Key Principles:

1. Informed Consent: Participants must know the purpose and risks (e.g., eSewa’s privacy policy).
2. Anonymity/Confidentiality: Protect identities (e.g., Nabil Bank’s loan applicant data).
3. Avoiding Harm: No deceptive experiments (e.g., Pathao’s driver studies use opt-in participation).
4. Data Security: Encrypt sensitive data (e.g., NTC’s customer records).

Case Study: Ncell’s Ethical Dilemma

• Issue: Wanted to track customer locations for "personalized offers" without consent.
• Solution: Implemented opt-in GPS tracking with clear disclosures → improved trust scores.

Exam Tip

How This Unit is Tested in PU Exams

1. Definitions & Classifications (20% weight):

   • Expect questions like: "Differentiate between structured and unstructured interviews with examples."
   • Tip: Use comparison tables (like above) to memorize differences.

2. Scenario-Based Applications (30% weight):

   • Example Question: "How would you collect data to study ‘customer satisfaction with Daraz’s delivery delays’? Justify your choice of method."
   • Answer Structure:
     • Method: Mixed-mode (online survey + phone interviews with delayed customers).
     • Tools: Google Forms + stratified sampling.
     • Ethics: Ensure anonymity; explain purpose upfront.

3. Advantages/Disadvantages (20% weight):

   • Example Question: "Evaluate the use of experiments in marketing research. Provide a Nepali business example."
   • Tip: Use bullet points (like the tables above) to list pros/cons concisely.

4. Data Collection Plan (30% weight):

   • Example Question: "Design a data collection plan for a study on ‘traffic congestion in Pokhara.’"
   • Answer Framework:
     1. Objective: Reduce congestion by 20% in 6 months.
     2. Method: Observation (camera data) + Surveys (driver interviews).
     3. Tools: Drones for aerial observation + Google Forms for surveys.
     4. Sampling: Stratified (commercial vs. residential areas).
     5. Ethics: Anonymize driver data; get traffic police approval.

Visual Summary for Quick Revision

flowchart LR
  A["Primary Data"] --> B["Surveys<br/>(Questionnaires)"]
  A --> C["Interviews<br/>(Structured/Unstructured)"]
  A --> D["Observations<br/>(Participant/Non-Participant)"]
  A --> E["Experiments<br/>(A/B Testing)"]
  A --> F["Focus Groups<br/>(Group Discussions)"]
  G["Secondary Data"] --> H["Internal<br/>(Sales Records)"]
  G --> I["External<br/>(Govt Reports)"]
  J["Ethical Considerations"] --> K["Informed Consent"]
  J --> L["Confidentiality"]
  J --> M["Avoid Harm"]

Worked Example: NTC’s Network Optimization Study

Problem: High call drop rates in Kathmandu Valley. Data Collection Plan:

1. Method:
   • Primary: Observation (technicians log drops via network monitors).
   • Secondary: Internal data (NTC’s call logs) + External data (IMF telecom reports).
2. Tools:
   • Observation: Signal strength meters.
   • Surveys: Customer feedback via SMS (Likert scale for call quality).
3. Sampling:
   • Stratified: High-drop zones (e.g., Thapathali, Koteshwor).
4. Analysis:
   • Cross-referenced drop locations with terrain maps (e.g., drops near tall buildings).
5. Solution:
   • Installed micro-towers in identified hotspots → 30% drop reduction.

Exam Answer Tip: Always link your method to the real-world context (like NTC’s case) to score full marks. Use bullet points for clarity:

• Method: Observation + secondary data.
• Tools: Signal meters + SMS surveys.
• Sampling: Stratified by drop zones.
• Ethics: Anonymized customer data.