Collection of Data

The full picture

Every economic statement you have ever heard — 'India's unemployment rate rose', 'onion prices doubled', 'Kerala has the highest literacy in India' — is built on data. Data is simply collected facts or measurements about something. Before an economist can analyse anything, someone has to gather the information first. This chapter is about that gathering step: what it means, how it is done, and why doing it badly ruins the entire chain of reasoning that follows.

Data has two broad origins. Primary data is information you collect yourself, fresh, for your specific question. Secondary data is information that was already gathered by someone else — government agencies, research institutions, newspapers — and is now available to you. Think of it this way: if you survey students in your class about how many hours they study, that is primary data. If you use the Kerala State Planning Board's report on student performance, that is secondary data. Both are useful; neither is automatically better. The right choice depends on what you need, when you need it, and how much it costs.

Primary data can be collected in three main ways. The first is a census, also called complete enumeration: you survey every single unit in the group you are studying. The Census of India, conducted every ten years, tries to record details about all 1.4 billion residents — their age, education, occupation, and housing. It is the most complete picture possible, but it costs thousands of crores of rupees and takes years to process. The second method is sampling: you choose a smaller, representative sub-group and study only them. If a district wants to know the average monthly expenditure of households, surveying 500 carefully chosen families can give a reliable answer at a fraction of the cost of visiting every household. The third method covers how you actually ask questions — through written questionnaires, personal interviews, telephone calls, or direct observation. Each approach carries its own strengths and weaknesses that you must weigh before choosing.

Secondary data is everywhere if you know where to look. The Reserve Bank of India (RBI) publishes quarterly reports on credit, money supply, and inflation. The National Statistical Office (NSO), which absorbed the former National Sample Survey Organisation (NSSO), runs regular surveys on household consumption, employment, and income. The Department of Economics and Statistics, Kerala publishes detailed state-level data on agriculture, population, and education. Private rating agencies like CRISIL and ICRA track industries and companies. When you write an economics project and cite a government table, you are using secondary data. The great advantage is speed and cost: the data already exists, so you just download or find it. The danger is that it may be outdated, collected for a different purpose than yours, or contain errors you did not notice.

Quality matters more than quantity in data collection. A small sample chosen carefully through random selection can be more reliable than a large survey contaminated by bias. Bias creeps in when the selection process systematically favours certain people or answers over others. For example, if a company surveys customer satisfaction only by calling people who purchased online, it will miss the experience of customers who buy in shops — skewing the result. Similarly, a leading question like 'Don't you find our service excellent?' pushes respondents toward a positive answer. When you evaluate any data source — whether for your exam, a project, or life — always ask: who collected this, when, how, and for what purpose? Those four questions are the foundation of critical thinking about statistics.

An Indian example

Ananya runs a small kirana store in Thrissur. She hears that sales of packaged atta have fallen across Kerala and wants to know if she should reduce her stock order. She finds two options: the Ministry of Consumer Affairs publishes average wholesale prices, but the latest figures are three months old and cover pan-India trends. Alternatively, she designs a quick questionnaire — ten questions, given to 40 regular customers over two days — asking how many kg of atta they buy per month and whether they switched to a local mill brand recently. Her 40-customer survey (primary data) costs her almost nothing and takes two days. It tells her exactly what her customers in her neighbourhood are doing right now. The government report (secondary data) is free but too old and too broad to answer her specific question. She discovers that 28 of her 40 customers started buying directly from a local flour mill that offers loose atta at ₹32/kg versus her packed brand at ₹48/kg. She adjusts her order immediately and avoids stocking ₹12,000 worth of slow-moving inventory. The lesson: secondary data is your starting point, but when your decision is specific and time-sensitive, primary data collected with a simple, focused questionnaire can save you real money.

Common misconceptions to watch for

• Wrong belief: 'A bigger sample is always more accurate.' Correction: A randomly selected sample of 400 households beats a carelessly chosen sample of 4,000, because random selection removes bias — what matters is how you select, not just how many you select.
• Wrong belief: 'Secondary data is always the better choice because it is free and already available.' Correction: Secondary data may be outdated, collected for a different purpose, or too broad for your question; when your need is specific or time-sensitive, collecting primary data is often the right — and necessary — choice.
• Wrong belief: 'A census (100% survey) is the gold standard and sampling is just a compromise.' Correction: Censuses are expensive, slow, and only practical for certain situations; well-designed sample surveys like the NSO's National Sample Survey are standard professional tools that governments and researchers rely on precisely because they are faster and cheaper while still being statistically reliable.