1. What Is Data?

Data are raw symbols, measurements, or observations that capture attributes of people, objects, events or concepts. When processed, analysed and contextualised, data yield information, and ultimately, knowledge.

2. Typologies of Data

2.1 By Source

• Primary Data – gathered firsthand by the current investigator to answer the contemporary problem.
• Secondary Data – pre-existing data originally compiled for another purpose.

2.2 By Measurement Scale

• Nominal (categories without order), e.g., blood type.
• Ordinal (ranked categories), e.g., satisfaction levels.
• Interval (equal intervals, no true zero), e.g., temperature (°C).
• Ratio (interval plus absolute zero), e.g., income, weight.

2.3 By Nature

• Quantitative – numeric, allows statistical operations.
• Qualitative – textual, visual, or audio descriptions capturing richness of context.

3. Meaning, Features & Distinction

3.1 Primary Data

Generated directly via fieldwork or experimentation. Researchers control definitions, timing, sampling, and measurement instruments, ensuring high relevance.

3.2 Secondary Data

Obtained from published or internal records: census tables, hospital registers, transaction logs, remote-sensing images, social-media APIs, etc. Valuable for historical analyses, benchmarking and exploratory work.

3.3 Comparison at a Glance

4. Selecting a Data-Collection Method – Decision Factors

The optimal method balances research objectives (depth vs breadth, causal vs descriptive), required precision, budget, timeline, respondent accessibility, ethical constraints, and the analytical strategy (e.g., multivariate statistics require quantifiable variables). A prudent rule is “use secondary data first”; if inadequate, design primary collection. Mixed-method triangulation often yields the most credible insights.

5. Primary Data-Collection Methods (Detailed)

a. Surveys

• Modes:
  • Self-administered (web, mobile, mail). Lowest cost, broad reach.
  • Interviewer-administered (CATI—Computer-Assisted Telephone Interviewing, CAPI—Computer-Assisted Personal Interviewing).
• Merits: Standardisation, scalability, robust for inference if sampling is probabilistic.
• Demerits: Non-response, social-desirability bias, limited probing depth.

b. Interviews

• Structured: uniform questions → comparability.
• Semi-structured: core guide + flexible probes.
• Unstructured / in-depth: narrative exploration.
• Focus: attitudes, motivations, sensitive topics.
• Demerits: interviewer bias, transcription burden, smaller samples.

c. Observation

• Participant vs Non-participant: immersion level affects objectivity.
• Naturalistic vs Controlled: trade-off realism vs rigor.
• Mechanical: CCTV, eye-tracking, IoT sensors—objective, high-frequency streams.

d. Experiments

• Laboratory: full control → high internal validity.
• Field / A-B Tests: natural environment → higher external validity.
• Online experimentation: rapid iteration at scale (e.g., website layout tests).
• Ethics: randomisation and informed consent crucial.

e. Focus Groups

• 6–10 homogeneous participants, 60–90 min guided discussion.
• Useful for idea generation, language testing, uncovering group norms.
• Moderator skill vital to avoid dominance or groupthink.

f. Emerging Approaches

• Mobile Ethnography – diary apps, geo-tagged photos, real-time context.
• Passive Digital Tracking – clickstreams, wearables (heart-rate, steps).
• Citizen Science / Crowdsourcing – distributed data collection (e.g., biodiversity counts).

Efficiency Spotlight
Why Online Surveys Often Win
Fast deployment, auto-coding, real-time dashboards, and negligible marginal cost make online surveys the most cost-efficient for large literate populations, provided internet penetration and response incentives are adequate.

6. Secondary Data Sources & Evaluation

6.1 Typical Sources

• Government census, vital statistics, economic indicators.
• International organisations (ILO, FAO, IMF).
• Academic journals, theses, open-access repositories (Zenodo, ICSSR).
• Commercial databases (Nielsen retail audit, Euromonitor).
• Administrative data (tax filings, hospital records).
• Digital traces (social-media posts, satellite imagery, Google Trends).

6.2 Quality-Appraisal Checklist (CRAFT)

• Coverage – Does the dataset fully cover the target population/timeframe?
• Reliability – Was the collection process documented, standardised?
• Accuracy – Error rates, validation checks, sampling design?
• Freshness – Publication lag, update frequency?
• Transparency – Access to metadata, coding manuals?

7. Comparative Merits & Demerits of Key Primary Methods

8. Ethics & Data-Quality Safeguards

• Informed consent – explain purpose, risks, rights.
• Confidentiality – anonymise identifiers; follow data-protection laws (e.g., GDPR).
• Minimisation – collect only data necessary for objectives.
• Data-quality controls – pilot testing, logic checks, duplicate detection.

9. Triangulation & Mixed-Method Strategies

Combining multiple sources (e.g., survey + focus groups + administrative data) enhances validity through corroboration, contextual depth and error reduction. Sequence can be:

• Exploratory Sequential: Qualitative → design survey instrument.
• Explanatory Sequential: Quantitative → qualitative follow-up for “why”.
• Concurrent: Collect both strands simultaneously; merge in analysis.

10. Key Takeaways

• Primary data offer precision and current relevance but demand higher resources; secondary data provide speed and breadth but require critical evaluation.
• Method choice is a multi-criteria optimisation of accuracy, cost, time, ethics, and analytic needs.
• Online surveys are generally the most cost-efficient for large literate populations, while passive sensors are unrivalled for fine-grained behavioural measures.
• Triangulation mitigates individual method weaknesses and strengthens credibility.