Chapter 3 — Organisation Of Data

Overview: In organising data we classify it so appropriate methods of tabulation, presentation and analysis can be used. "Types of data" refers to different ways data can be grouped according to source, nature, continuity, time-dimension and measurement scale.

1. By Source

• Primary data: Collected first-hand for a specific purpose (surveys, interviews, experiments). Advantage: tailored and current. Disadvantage: costly and time-consuming.
• Secondary data: Already collected for another purpose (government publications, research papers, databases). Advantage: cheaper and quick. Disadvantage: may not fit exactly and may be outdated or biased.

2. By Nature

• Qualitative (Categorical) data: Describes qualities or categories. Examples: gender, religion, colour, type of house. Usually summarized by counts or percentages.
• Quantitative (Numerical) data: Numerical values that measure quantity. Examples: income, age, production. Can be further analysed by arithmetic operations.

3. By Continuity (for quantitative data)

• Discrete data: Countable values (integers). Example: number of students in a class, number of factories.
• Continuous data: Can take any value in an interval (measurements). Example: height, weight, distance, income measured precisely.

4. By Time-Dimension

• Time-series data: Observations of a variable over time (monthly GDP, annual rainfall). Used to study trends, seasonal effects and cycles.
• Cross-section data: Observations at a single point in time across units (income of households in 2024). Used to compare units at same time.
• Pooled (Panel) data: Combination of time-series and cross-section (same households observed across several years). Useful for dynamic analysis.

5. By Measurement Scale

• Nominal: Names or labels without order (e.g., blood group). Only mode and percentages are meaningful.
• Ordinal: Categories with order but unequal intervals (e.g., education level: primary, secondary, tertiary). Median and percentiles meaningful.
• Interval: Ordered with equal intervals but no true zero (e.g., Celsius temperature). Differences are meaningful, ratios are not.
• Ratio: Like interval but with a true zero (e.g., income, weight). All arithmetic operations and ratios are meaningful.

Use in economics: Correct identification of data type determines choice of tables, graphs and statistical measures (mean, median, mode, dispersion measures, correlation, regression). For example, categorical data are shown with bar charts/pie charts; continuous quantitative data with histograms, frequency polygons and box plots; time-series with line graphs; cross-section with scatter plots.

Overview

Data collection is the systematic process of gathering information for analysis. In economics (Class 11), we classify data by its source (where it comes from) and by the method used to collect it. Good data collection ensures accuracy, reliability and relevance for decision-making and statistical analysis.

Sources of Data

There are two broad categories:

• Primary data: Data collected first-hand for a specific purpose. It is original and fresh.
• Secondary data: Data already collected by someone else and available in published or unpublished form.

Primary sources (examples and use)

• Surveys and questionnaires — ask respondents directly.
• Interviews — structured, semi-structured, or unstructured.
• Observation — direct (researcher observes behaviour) or indirect (use of cameras, records).
• Experiments — controlled manipulation to observe effects.
• Case studies — in-depth study of a single unit (e.g., a firm or household).

Secondary sources (examples and use)

• Official publications (Census of India, NSSO/NSO reports, government statistics).
• Administrative records (tax records, hospital registers, school enrolments).
• Published sources (books, journals, newspapers, company annual reports).
• Online databases, research repositories, and archival material.

Methods of Data Collection

Choice of method depends on objectives, budget, time, required accuracy and the nature of respondents.

Census versus Sampling

• Census: Collect data from every unit in the population (e.g., decennial population census). Accurate but expensive/time-consuming.
• Sampling: Collect data from a subset (sample) of the population and generalise results. More practical and cheaper; requires careful design to avoid sampling bias.

Common sampling designs

• Probability sampling (each unit has known chance): simple random sampling, systematic sampling, stratified sampling, cluster sampling.
• Non-probability sampling (selection not random): convenience sampling, judgmental (purposive) sampling, quota sampling, snowball sampling.

Primary data collection methods — brief descriptions

• Questionnaire: Written set of questions—good for large samples. Questions can be closed or open-ended.
• Interview: Oral questioning—face-to-face, telephone, or video. Can probe deeper but costlier.
• Observation: Recording behaviour or events without asking. Useful when responses may be biased.
• Schedule: Similar to questionnaire but filled by the investigator after interviewing respondent—used in large-scale studies (e.g., household surveys).
• Experiment: Manipulating variables to test hypotheses—common in behavioural economics/lab studies.
• Case study: Intensive examination of a single entity for in-depth understanding.

Quality, Errors and Ethics

• Validity: Data measures what it is intended to measure.
• Reliability: Data collection would yield similar results under consistent conditions.
• Errors: sampling errors (due to sample selection) and non-sampling errors (measurement error, non-response, data processing mistakes).
• Ethics: informed consent, confidentiality, no harm, honest reporting.

Steps in Data Collection (practical workflow)

1. Define objective and population.
2. Choose data source (primary/secondary) and method.
3. Design instrument (questionnaire/schedule) and sample (if sampling).
4. Pilot test instrument and revise.
5. Collect data, supervise and record metadata (who, when, how).
6. Process, clean and validate the data.

When to use which source/method?

• Use primary data when specific, up-to-date, or otherwise unavailable information is needed (e.g., consumer preferences for a new product).
• Use secondary data for background, trend analysis, or when reliable official data exist (e.g., GDP series, census counts).

Note: In practice researchers combine methods (mixed methods) to improve coverage and validity—e.g., use secondary data for context and primary surveys for current measurements.

Meaning: Classification of data is the process of arranging raw data into groups or classes that have similar characteristics so that it becomes meaningful and easier to analyze. Proper classification helps in summarizing large volumes of data and in choosing appropriate methods of presentation and analysis.

Why classify? Raw data are often voluminous and unstructured. Classification simplifies data, highlights patterns, facilitates comparison and helps in drawing conclusions.

Main bases (types) of classification:

• By source: Primary data (collected first-hand through surveys, experiments, observations) and Secondary data (collected earlier by someone else — censuses, reports, books, databases).
• By nature: Qualitative (attribute/categorical data describing qualities — e.g., gender, occupation) and Quantitative (numerical data measuring quantity — e.g., income, marks).
• By measurement/ordering:
  • Nominal: categories without order (e.g., religion, blood group).
  • Ordinal: categories with a meaningful order but not equal intervals (e.g., grades: A, B, C; satisfaction levels: low, medium, high).
  • Interval/Ratio: numerical scales with meaningful differences; ratio scale has a true zero (e.g., weight, income).
• By continuity: Discrete data (take distinct separate values — counts such as number of children) and Continuous data (take any value in a range — measurements such as height, time).
• By statistical unit: Individual data (values for each unit — marks of each student) and Aggregate data (summary measures for groups — total sales of a firm, average income of a region).
• By time/space: Cross-sectional data (observations collected at the same time for different units — incomes of households in 2025) and Time-series data (observations of a variable taken over time — monthly CPI, yearly GDP).

How classification affects presentation: Choice of charts and tables depends on the type of data. Qualitative data are usually shown with bar charts or pie charts. Quantitative discrete data are shown with bar charts; quantitative continuous data are shown with histograms, frequency polygons or ogives. Time-series data are best shown with line graphs.

Important terms related to classification:

• Class interval or class: a group into which continuous data are divided (e.g., 10–19, 20–29).
• Class width: size of a class interval.
• Class mark (midpoint): the central value of a class interval used for certain calculations.
• Frequency: number of observations in a class or category.

Practical tips: Use appropriate categories (mutually exclusive and exhaustive). For continuous data choose suitable class width and number of classes so that the frequency distribution is neither too detailed nor too coarse.

Definition: Tabulation is the process of arranging raw data in rows and columns (a table) according to a plan so that the information becomes clear, systematic and easily interpretable. In economics and statistics it is the first step in organising data for further analysis.

Purpose / Importance:

• Summarises large volumes of data in compact form.
• Makes comparison and pattern recognition easier.
• Prepares data for calculation of statistical measures and for graphical presentation.

Components of a Table:

• Title: concise description of what the table contains.
• Headings: column headings that describe variables (e.g., Year, Population).
• Stubs: row labels (e.g., States, Categories).
• Body: the cells where data values are placed.
• Footnote / Source: any clarifications or data sources.

Types of Tables:

• Simple table: Lists observations (e.g., names and marks).
• Frequency distribution table: Groups data into classes with frequencies.
• One-way table: Frequency distribution of a single variable.
• Two-way (contingency) table: Cross-classifies two variables (e.g., gender × employment status).

Rules / Guidelines for Constructing Tables:

• Give a clear and complete title.
• Arrange rows and columns logically (e.g., chronological order, increasing magnitude).
• Use mutually exclusive and exhaustive classes when grouping continuous data.
• Choose an appropriate number of classes (not too many, not too few).
• Use consistent units and state them.
• Place totals (marginal totals) where useful.
• Make the table readable: align numbers, use adequate spacing and clear headings.

How Tabulation Helps Further Analysis: After tabulation you can compute frequencies, relative frequencies, percentages, cumulative frequencies and prepare graphs (histograms, pie charts, ogives) that summarise the data visually.

Limitations: A table can hide individual observations; grouping may cause loss of detail; poor design can mislead readers.

What is Frequency Distribution? A frequency distribution organises raw data into classes (intervals) and shows how often (frequency) values occur in each class. It simplifies large data sets and helps identify patterns (central tendency, dispersion, shape).

Why grouping? Raw data can be unwieldy. Grouping into classes reduces detail while preserving important structure so we can display data using histograms, frequency polygons and ogives.

Steps to construct a grouped frequency distribution

• Collect and sort the raw data (optional).
• Find the range: maximum − minimum.
• Decide number of classes (k). Use Sturges' rule as a guideline: k ≈ 1 + 3.322 log10(n), then round.
• Determine class width (h) = range / k, then round up to a convenient value.
• Choose class limits (non-overlapping continuous intervals) so they cover the entire range.
• Tally observations into classes to get frequency (f).
• Calculate cumulative frequency, relative frequency, class midpoints, etc. for analysis and graphs.

Key terms

• Class limits: lower and upper bounds of an interval.
• Class boundaries: exact cut-points between classes (useful for continuous data).
• Class mark (midpoint) = (lower limit + upper limit)/2.
• Cumulative frequency: running total of frequencies up to a class.
• Relative frequency = frequency / total observations.
• Frequency density = frequency / class width (used when class widths differ).

Important notes on presentation: For continuous quantitative data, use a histogram (adjacent bars). For discrete or categorical data, use bar diagrams (bars separated). Use frequency polygon (join class midpoints) to compare distributions. Use ogive (cumulative frequency curve) to read medians and percentiles.

What it is: Presentation of data means arranging and representing numerical or categorical information in pictorial form so that patterns, trends and comparisons become easy to understand. Diagrams and graphs convert raw data (tables and frequency distributions) into visual forms.

Why it matters: Visual representations make complex data accessible, help comparison, highlight trends, and support decision-making in real life (business, economics, public policy, education).

Main types:

• Qualitative / Categorical diagrams: Bar diagram, multiple bar diagram, component bar diagram, pictogram and pie chart — used for categorical comparisons (e.g., sectoral contribution to GDP, market share).
• Quantitative / Frequency diagrams: Histogram, frequency polygon, ogive (cumulative frequency curve), and line graph — used for grouped numerical data and to show distributions or trends (e.g., marks distribution, income classes, time series).

How to choose: Use bar/pictogram for comparing categories, pie chart for showing percentage composition, line graph for time-series/trends, histogram/frequency polygon/ogive for continuous grouped data distribution, and pictorial/stacked/component bars when parts-of-whole or multiple comparisons are needed.

Basic construction principles:

• Label axes and units clearly; give a descriptive title.
• Choose an appropriate scale so the plot fills the space and is easy to read.
• For grouped data, use class boundaries/widths correctly; for unequal class widths use frequency density in histograms.
• Maintain proportionality — areas/heights/angles must reflect data values (e.g., pie-chart angles proportional to frequencies).

Advantages: Quick visual comparison, identification of central tendency and dispersion patterns, easy presentation to non-technical audiences.

Limitations: Can hide details or distort data if scales or areas are chosen improperly; not all kinds of analysis (e.g., exact values, hypothesis testing) can be done from a graph alone.

Overview
This topic collects short definitions and bite-sized concepts used in the chapter "Organisation of Data". It explains types of data, methods to organise data, basic components of frequency distributions and the common graphical forms used to present data.

Basic terms

• Data — Raw facts and figures collected for analysis (numbers, attributes, observations).
• Variable — A characteristic that can take different values (e.g., height, income).
• Attribute — Qualitative characteristic (e.g., gender, colour).
• Qualitative (categorical) — Non‑numeric data (names, categories).
• Quantitative — Numeric data; can be discrete (countable values) or continuous (measurable on a continuum).
• Univariate / Bivariate / Multivariate — Number of variables involved (one, two, many).

Sources of data

• Primary data — Collected firsthand by the investigator (surveys, experiments, observations).
• Secondary data — Collected earlier by someone else (census reports, published statistics).
• Census — Complete enumeration of all units in a population.
• Sample — Subset of the population used when census is impractical.

Sampling methods (short concept)

• Random (simple random) — Every unit has equal chance.
• Stratified — Population divided into strata; samples taken from each strata (improves representativeness).
• Systematic — Every k-th unit chosen (after random start).
• Cluster — Population divided into clusters; a few clusters chosen and all units in them surveyed.
• Quota — Interviewers collect data until quotas for subgroups are met (non‑probability).

Frequency distribution — key components

• Class interval — A range of values grouped together (e.g., 10–14).
• Class limits — Lower and upper boundary values of a class (inclusive/exclusive depending on convention).
• Class boundaries — Exact dividing lines between classes used to avoid gaps (useful for continuous data).
• Class width (size) — Difference between upper and lower limits of a class.
• Class mark (mid-point) — (Lower limit + Upper limit) / 2; used as representative value for the class.
• Frequency (f) — Number of observations in a class.
• Cumulative frequency (CF) — Running total of frequencies up to (≤) a class.
• Relative frequency — f / N (proportion of total).

Tabulation principles (short)

• Have a clear title and source.
• Use appropriate class widths; classes should be mutually exclusive and exhaustive.
• Arrange classes in ascending or descending order.
• Include totals and cumulative columns if needed.

Graphical presentation — quick notes

• Bar chart — For categorical (qualitative) data; bars separated.
• Histogram — For continuous grouped data; contiguous bars with area proportional to frequency.
• Frequency polygon — Join class‑marks at heights equal to frequencies by straight lines; useful for comparing distributions.
• Ogive (cumulative frequency curve) — Plots class upper (or lower) boundaries against cumulative frequency; useful for finding medians/percentiles graphically.
• Pie chart — Shows composition (percentages) of a whole; use for a few categories only.

Short concept tips

• Continuous data needs contiguous classes (no gaps); discrete data can have single‑value classes.
• Open-ended classes (like "60 and above") are allowed but require caution for mid‑point or mean estimation.
• Choose class width so you have between about 5 and 15 classes (practical rule) depending on sample size.