Paper 3 · Statistics

Statistical Description Of Data – Chapter Notes

Complete CA Foundation Business Statistics notes on statistical description of data, collection and presentation of data, frequency distributions, histograms, frequency polygons and ogives.

Crux First

  • Statistics can mean numerical facts in the plural sense and the science of collecting, organising, analysing and interpreting data in the singular sense.
  • Primary data are collected first-hand for the present enquiry; secondary data already exist and are reused.
  • Data may be presented textually, in tables or through diagrams.
  • A frequency distribution shows how often each value or class occurs.
  • Class limits are stated limits; class boundaries are actual continuous limits.
  • Histogram uses adjacent rectangles; for unequal class widths, compare classes through frequency density.
  • Frequency polygon uses class mid-points. Ogives use class boundaries and cumulative frequencies.
  • The intersection of less-than and more-than ogives gives the median graphically.

1. Meaning and Scope of Statistics

Statistics simply means using numbers to understand what is happening.

Imagine a factory produces 10,000 parts every day. It records how many parts were made, how many were rejected, which machine created more defects and which shift performed better. These are just numbers. When we arrange and compare them to find a pattern, we are using Statistics.

Central LogicRaw data tell us the numbers. Statistics helps us understand what those numbers mean.

Statistics as Numerical Facts

In the plural sense, statistics means numerical facts. Examples are sales figures, population numbers, wages, accident counts, marks and rejection percentages.

A figure must be complete. Saying “production is 500” is not enough. We must know whether it means 500 pieces per day, 500 tonnes per month or 500 orders in a year.

Statistics as a Method

In the singular sense, statistics means the method used to collect, organise, present, analyse and understand data.

It is not only about counting or finding averages. It covers the full process from collecting the right data to drawing a useful conclusion.

CollectClassifyPresentAnalyseInterpret
MCQ TrapThe plural meaning refers to data. The singular meaning refers to the method or discipline.
Memory LinkPlural = Figures. Singular = Science.

2. Applications of Statistics

Statistics is used wherever decisions are made with the help of numbers. It helps people compare, plan, predict and control.

Economics

Statistics is used in demand analysis, index numbers, forecasting, time-series analysis, national income estimation and economic planning.

Business Management

Managers use statistical methods for forecasting, quality control, market research, inventory planning and decision-making under uncertainty.

Commerce and Industry

Past sales, production, wages, costs, competitor data and market trends are analysed to improve planning and profitability.

Government and Public Services

Census, health statistics, unemployment, agriculture, defence and welfare planning depend heavily on reliable data.

Exam LogicICAI may describe a practical situation and ask which field of statistics is being applied. Focus on the purpose: forecasting, comparison, planning, quality control or inference.

3. Limitations of Statistics

Statistics is useful, but it can mislead us if the data are wrong or the sample is biased. It should support judgement, not replace it.

  1. Statistics studies aggregates. A single isolated observation usually has little statistical meaning.
  2. Statistics mainly deals with quantitative information. Qualitative characteristics must first be coded or numerically described.
  3. Conclusions depend on conditions. Forecasts may fail when the underlying conditions change.
  4. Sampling must be representative. A biased or unrepresentative sample produces misleading conclusions.
  5. Statistics does not replace judgement. It supports decisions; it cannot compensate for faulty definitions, poor data or wrong interpretation.
Common Wrong Statement“Statistics studies every individual item separately.” This is incorrect. Its main concern is the aggregate.

4. Data, Variables and Attributes

Before solving a question, first identify what is being studied. If it can be measured in numbers, it is a variable. If it is only a category or description, it is an attribute.

Data

Data are facts or information relating to a characteristic under study. For statistical analysis, even qualitative information may be converted into numerical categories or codes.

ConceptMeaningExamplesMCQ Clue
VariableA measurable characteristicHeight, weight, profit, salaryCan take numerical values
Discrete VariableTakes isolated, countable valuesNumber of accidents, defects, childrenUsually counted
Continuous VariableCan take any value within an intervalHeight, weight, temperatureUsually measured
AttributeA qualitative characteristicGender, nationality, colourCategory, not measurement
Memory RuleDiscrete = count. Continuous = measure. Attribute = describe.

5. Primary Data and Secondary Data

The easiest rule is: Collected by me for my present work = Primary Data. Already collected by someone else = Secondary Data. The same data can be primary for one person and secondary for another.

BasisPrimary DataSecondary Data
MeaningCollected first-hand for the present purposeAlready collected earlier for another purpose
OriginalityOriginal to the investigatorNot original to the present user
Cost and TimeUsually higherUsually lower
SuitabilityDesigned for the present enquiryMust be checked for relevance and reliability
ExampleA factory conducts its own employee surveyThe factory uses government labour data
Critical LogicThe same data can be primary for one person and secondary for another. Classification depends on who collected it and for what purpose.

Illustrative MCQ

Professor A records the heights of his students. Professor B later uses the same record to calculate average height. The data are:

(a) Primary for both   (b) Secondary for both   (c) Primary for A and secondary for B   (d) Secondary for A and primary for B

Answer: (c). A collected the data first-hand; B reused existing data.

6. Methods of Collecting Primary Data

There is no single best method of collecting data. The correct method depends on the situation.

Selection RuleNeed detailed answers? Use an interview. Need wide coverage? Use a questionnaire. Need an exact physical fact such as height or weight? Use observation or measurement.
MethodHow it WorksMain StrengthMain Limitation
Personal InterviewInvestigator directly meets respondentsDetailed and relatively accurateCostly and difficult over a wide area
Indirect InterviewInformation is obtained from persons connected with the eventUseful where direct contact is impossibleDepends on the informant’s knowledge and neutrality
Telephone InterviewQuestions are asked over the phoneQuick and economicalNon-response and limited depth
Mailed QuestionnaireQuestionnaire is sent to respondents for self-completionWide geographical coverageHigh non-response and misunderstanding
ObservationInvestigator directly observes or measuresUseful for objective, visible factsTime-consuming and limited in scope
Enumerator MethodTrained enumerators ask questions and fill the scheduleQuestions can be explainedExpensive and vulnerable to enumerator bias

Sources of Secondary Data

  • International organisations such as the World Bank, IMF, WHO and ILO.
  • Government publications, statistical abstracts and ministry reports.
  • Research institutes, universities and quasi-government bodies.
  • Private reports, trade associations and unpublished research.
Selection Logic“Quick and cheap over a wide area” points towards telephone or mailed questionnaire. “Direct and accurate but costly” points towards personal interview or observation.

7. Scrutiny of Data

Collected data may contain mistakes. Before using it, we must check whether the figures are complete, sensible and internally consistent. This is called scrutiny of data.

Before analysis, data must be checked for accuracy, consistency, completeness and reasonableness.

  • Clerical errors: mistakes in copying, writing or totalling.
  • Internal inconsistency: related figures fail to satisfy a known relationship.
  • Enumerator bias: returns show a suspicious pattern or lack of genuine enquiry.
  • Missing or impossible values: observations fall outside any reasonable range.
Population Density = Population ÷ Area
Internal Check ExampleIf population, area and density are all given, verify whether density equals population divided by area.

8. Classification of Data

Raw data are difficult to understand because the figures are scattered. Classification means putting similar observations into groups so that the data become easy to read and compare.

Classification means arranging observations into groups or classes according to common characteristics.

Objectives

  • Condenses a large mass of data.
  • Makes comparison possible.
  • Reveals similarities, differences and relationships.
  • Prepares data for statistical analysis.
TypeBasisExample
Chronological / TemporalTimeMonthly production from January to December
Geographical / SpatialPlace or regionState-wise sales
Qualitative / OrdinalAttribute or categoryGender, literacy, smoking habit
Quantitative / CardinalNumerical variableIncome, marks, height
MCQ TrapTime-series and geographical series are generally non-frequency classifications. Qualitative and quantitative classifications can form frequency distributions.

9. Modes of Presentation of Data

Data can be shown in words, tables or diagrams. Use text for a small amount of data, a table when exact figures are important, and a diagram when you want to show a trend or comparison quickly.

Textual Presentation

Data are described through sentences or paragraphs. It is simple and useful for small amounts of information, but comparison is difficult and the presentation becomes dull for large data.

Tabular Presentation

Data are presented systematically in rows and columns. A good table should have a table number, clear title, row headings, column headings, units, totals, source and footnotes where required.

Part of TableMeaning
CaptionHeadings describing columns and sub-columns
Box-headThe complete upper part including captions, column numbers and units
StubThe left-hand part describing rows
BodyThe main field containing numerical entries
Footnote / SourceClarification and origin of data

Diagrammatic Presentation

Charts and diagrams communicate patterns quickly and can reveal trends that are not obvious in a table. They are attractive and easy to understand, but less precise than tabulation.

Accuracy RuleWhen exact values matter, prefer a table. When trend or comparison matters, a diagram may communicate better.

10. Line, Bar and Pie Diagrams

Choose the diagram according to the question: trend over time = line chart; comparison = bar chart; parts of a total = pie chart.

Line Diagram

Use a line diagram when a variable changes over time. Plot time on the horizontal axis and the value on the vertical axis, then join successive points.

  • Multiple line chart: two or more related series measured in the same unit.
  • Multiple-axis chart: related series measured in different units.
  • Logarithmic or ratio chart: useful when fluctuations cover a very wide range and relative changes matter.

Bar Diagram

  • Horizontal bars: commonly used for qualitative or geographical data.
  • Vertical bars: commonly used for time-series or quantitative comparison.
  • Multiple bars: compare two or more related series.
  • Component bars: show parts of a total.
  • Percentage bars: compare proportional composition where each bar represents 100%.

Pie Chart

A pie chart shows how a total is divided among components. Each sector’s angle is proportional to its share.

Central Angle = (Component Value ÷ Total Value) × 360°

Mini Example

A company spends ₹80 lakh on materials out of total expenditure of ₹400 lakh.

(80 ÷ 400) × 360° = 72°
The material sector will have an angle of 72°.
MCQ TrapA pie chart is best for parts of one whole. It is not the best choice for showing a trend over time.

11. Frequency Distribution

Suppose you have marks of 500 students. Reading all 500 marks one by one is difficult. So we group similar marks together and count how many students fall in each group. This table is called a frequency distribution.

A frequency distribution is a table showing the number of observations falling against each value or class interval.

Types

  • Discrete or ungrouped frequency distribution: frequency is recorded against individual values.
  • Grouped frequency distribution: frequency is recorded against class intervals.

Steps in Construction

  1. Identify the smallest and largest observations.
  2. Calculate the range.
  3. Select class length and number of classes.
  4. Create mutually exclusive and exhaustive classes.
  5. Use tally marks.
  6. Count tallies and verify that total frequency equals the number of observations.
Range = Largest Observation − Smallest Observation
Approximate Number of Classes × Class Length ≈ Range
Rounding LogicIf the computed number of classes is not an integer, generally move to the next integer so that the maximum value is included.

12. Important Terms in a Frequency Distribution

Class limits, class boundaries, mid-points and class width look similar, but they mean different things. Many MCQs are based on this confusion.

Class Limits

The stated minimum and maximum values of a class are the lower class limit and upper class limit.

Class Boundaries

Class boundaries are the actual continuous limits separating adjacent classes. In exclusive classes such as 10–20, 20–30, boundaries coincide with limits. In inclusive classes such as 10–19, 20–29, adjustment is required.

D = Lower Limit of Next Class − Upper Limit of Present Class
Lower Class Boundary = Lower Class Limit − D/2
Upper Class Boundary = Upper Class Limit + D/2

Example

For classes 44–48 and 49–53:

D = 49 − 48 = 1
LCB = 44 − 0.5 = 43.5
UCB = 48 + 0.5 = 48.5

Mid-point or Class Mark

Mid-point = (Lower Limit + Upper Limit) ÷ 2
Mid-point = (Lower Boundary + Upper Boundary) ÷ 2

Class Width

Class Width = Upper Class Boundary − Lower Class Boundary
MCQ TrapFor graphs such as histograms and ogives, class boundaries—not merely stated class limits—are often required.

13. Cumulative, Relative and Percentage Frequency

Ordinary frequency tells us how many observations are in one class. Cumulative frequency tells us how many are below or above a point. Relative and percentage frequencies show each class as a share of the total.

Less-than Cumulative Frequency: Count Upwards

Add frequencies progressively from the first class downward. The values rise from zero towards total frequency.

More-than Cumulative Frequency: Count Downwards

Begin with total frequency and subtract class frequencies progressively. The values fall towards zero.

ClassFrequencyLess-than CFMore-than CF
0–103312
10–20479
20–305125

Relative Frequency

Relative Frequency = Class Frequency ÷ Total Frequency

Percentage Frequency

Percentage Frequency = (Class Frequency ÷ Total Frequency) × 100

Frequency Density

Frequency Density = Class Frequency ÷ Class Width
Check TotalsRelative frequencies add to 1. Percentage frequencies add to 100%.

14. Histogram

A histogram is used for continuous grouped data. Its rectangles touch each other because the classes are continuous. The area of each rectangle should represent the frequency.

A histogram represents a continuous frequency distribution through adjacent rectangles.

  • The horizontal axis carries class boundaries.
  • The vertical axis carries frequency when class widths are equal.
  • When class widths are unequal, use frequency density.
  • There are no gaps between adjacent rectangles.
Area of Rectangle ∝ Frequency
Height for Unequal Classes = Frequency Density = Frequency ÷ Class Width
Bar DiagramHistogram
May have gaps between barsRectangles are adjacent
Used for discrete or categorical comparisonUsed for continuous grouped data
Width usually has no numerical meaningWidth represents class interval
Height represents magnitudeArea represents frequency; height may represent density
High-Value TrapFor unequal class widths, using raw frequency as height gives a misleading histogram. Use frequency density.

15. Frequency Polygon

A frequency polygon is made by plotting class mid-points against frequencies and joining the points. It is useful for comparing two or more distributions on the same graph.

A frequency polygon is drawn by plotting class mid-points against corresponding frequencies and joining successive points with straight lines.

  1. Calculate class mid-points.
  2. Plot each pair: (mid-point, frequency).
  3. Join the points.
  4. Add one imaginary class at each end with zero frequency to close the polygon.
Visual LogicHistogram uses full rectangles. Frequency polygon uses the mid-points of the top sides of those rectangles.
Best UseFrequency polygons are convenient for comparing the shapes of two or more distributions on the same axes.

16. Ogives or Cumulative Frequency Curves

Ogives are cumulative frequency curves. They help us find how many observations are below or above a value and also help locate the median and quartiles.

An ogive is obtained by plotting cumulative frequency against class boundaries.

Less-than Ogive

  • Plot upper class boundaries on the horizontal axis.
  • Plot less-than cumulative frequencies on the vertical axis.
  • The curve generally rises from left to right.

More-than Ogive

  • Plot lower class boundaries on the horizontal axis.
  • Plot more-than cumulative frequencies on the vertical axis.
  • The curve generally falls from left to right.

Graphical Quartiles

  • The intersection of less-than and more-than ogives gives the median.
  • Quartiles can also be located using cumulative frequency positions.
Q₁ Position = N/4    |    Median Position = N/2    |    Q₃ Position = 3N/4
MCQ TrapHistogram and polygon use ordinary frequency. Ogives use cumulative frequency.

17. Frequency Curves

A frequency curve shows the overall shape of the distribution. It tells us whether most observations are near the centre, near the ends or mainly on one side.

A frequency curve is a smooth curve representing the general shape of a distribution. It may be treated as a smooth limiting form of a histogram or frequency polygon.

ShapePatternTypical Interpretation
Bell-shapedLow at both ends, highest near the centreMany natural characteristics such as height or marks
U-shapedHigh at both ends, low in the middleTwo extreme groups dominate
J-shapedStarts low and rises strongly towards one endFrequency accumulates towards one extreme
MixedCombination of shapesMore complex population structure
Memory PictureBell = centre-heavy. U = extremes-heavy. J = one-side-heavy.

18. Decision Guide: Which Method Should You Use?

Question RequirementBest MethodReason
Show movement over yearsLine diagramEmphasises trend over time
Compare categoriesBar diagramDirect visual comparison
Show components of one totalPie chart or component barShows composition
Summarise repeated valuesFrequency distributionShows occurrence count
Represent continuous grouped dataHistogramArea represents frequency
Compare distribution shapesFrequency polygonMultiple polygons can share one graph
Find median or quartiles graphicallyOgiveUses cumulative frequencies
Exact detailed figures requiredTableMore precise than a diagram

19. Integrated Worked Example

The weights of 20 components, in kilograms, are:

44, 48, 52, 55, 58, 61, 63, 64, 67, 69, 51, 54, 57, 59, 62, 65, 68, 70, 72, 73

Step 1: Find the Range

Range = 73 − 44 = 29 kg

Step 2: Form Classes of Width 5

Suitable inclusive classes are 44–48, 49–53, 54–58, 59–63, 64–68 and 69–73.

ClassFrequencyClass BoundariesMid-pointLess-than CF
44–48243.5–48.5462
49–53248.5–53.5514
54–58453.5–58.5568
59–63458.5–63.56112
64–68463.5–68.56616
69–73468.5–73.57120
What Each Column Is Used ForClass boundaries are used on the horizontal axis of a histogram and ogive. Mid-points are used for a frequency polygon. Less-than cumulative frequency is used for the rising ogive.
Exam ApproachDo not calculate every column automatically. First identify what the question asks, then calculate only the required column.

20. Solved ICAI-Style MCQs

1. Statistics in the singular sense refers to:

(a) Numerical facts   (b) A scientific method   (c) A government report   (d) A frequency table

Answer: (b). Singular statistics means the discipline or method.

2. Number of defective components produced in a shift is:

(a) Continuous variable   (b) Attribute   (c) Discrete variable   (d) Secondary data

Answer: (c). Defects are counted in whole numbers.

3. Weight of a forged component is:

(a) Discrete   (b) Continuous   (c) Qualitative   (d) Chronological

Answer: (b). Weight can assume any value within an interval.

4. Data collected by a researcher directly from respondents are:

(a) Primary   (b) Secondary   (c) Spatial   (d) Published

Answer: (a). They are collected first-hand for the enquiry.

5. A major weakness of mailed questionnaires is:

(a) No wide coverage   (b) Very high cost   (c) High non-response   (d) No written record

Answer: (c). Many recipients may not return the questionnaire.

6. State-wise production data are classified as:

(a) Chronological   (b) Geographical   (c) Qualitative   (d) Discrete

Answer: (b). Classification is based on region or place.

7. The left-hand part of a statistical table describing rows is called:

(a) Caption   (b) Body   (c) Stub   (d) Box-head

Answer: (c). Stub contains row descriptions.

8. The angle of a pie sector for a component equal to 25% of total is:

(a) 45°   (b) 72°   (c) 90°   (d) 120°

Answer: (c). 25% × 360° = 90°.

9. For classes 20–29 and 30–39, the upper boundary of the first class is:

(a) 29   (b) 29.5   (c) 30   (d) 30.5

Answer: (b). Gap D = 1, so add 0.5 to 29.

10. The mid-point of class 40–50 is:

(a) 40   (b) 45   (c) 50   (d) 90

Answer: (b). (40 + 50) ÷ 2 = 45.

11. When class intervals are unequal, histogram height should be based on:

(a) Class mark   (b) Relative frequency only   (c) Frequency density   (d) Cumulative frequency

Answer: (c). This keeps rectangle area proportional to frequency.

12. A frequency polygon is plotted using:

(a) Class boundaries and cumulative frequency   (b) Mid-points and frequencies   (c) Limits and percentages   (d) Raw values only

Answer: (b). Plot each class mid-point against its frequency.

13. The intersection of less-than and more-than ogives gives:

(a) Mean   (b) Mode   (c) Median   (d) Range

Answer: (c). The horizontal coordinate of intersection gives the median.

14. Relative frequencies of all classes together equal:

(a) 0   (b) 1   (c) 10   (d) 100

Answer: (b). Percentage frequencies, in contrast, total 100.

15. Which statement is incorrect?

(a) Statistics deals with aggregates   (b) Poor sampling can mislead   (c) Statistics always gives exact future predictions   (d) Qualitative data may be coded numerically

Answer: (c). Forecasts depend on assumptions and may fail when conditions change.

16. Which method is most suitable for obtaining the actual weight of each student?

(a) Mailed questionnaire   (b) Observation or measurement   (c) Indirect interview   (d) Published report

Answer: (b). Weight is an objectively measurable fact, so direct measurement is more reliable than asking respondents.

17. A table showing monthly sales for five years is classified mainly as:

(a) Geographical data   (b) Qualitative data   (c) Chronological data   (d) Attribute data

Answer: (c). The classification is based on successive time periods.

18. In a less-than cumulative frequency distribution, the final cumulative frequency equals:

(a) Highest class limit   (b) Total number of observations   (c) Class width   (d) Zero

Answer: (b). All class frequencies have been accumulated by the final boundary.

19. A histogram differs from a bar diagram mainly because:

(a) It has no vertical axis   (b) Its rectangles are adjacent and area is meaningful   (c) It cannot show frequency   (d) It is used only for qualitative data

Answer: (b). A histogram represents continuous grouped data, and rectangle area corresponds to frequency.

20. Which graph should be chosen to estimate the median directly from cumulative frequencies?

(a) Pie chart   (b) Multiple bar chart   (c) Ogive   (d) Line diagram

Answer: (c). The intersection of less-than and more-than ogives gives the median graphically.

21. Final Rapid Revision Sheet

ConceptOne-Line Recall
Plural StatisticsNumerical facts or data
Singular StatisticsScience of collecting, organising, analysing and interpreting data
Discrete VariableCountable isolated values
Continuous VariableAny value within an interval
Primary DataCollected first-hand for present purpose
Secondary DataAlready collected and reused
RangeMaximum − Minimum
Mid-point(Lower + Upper) ÷ 2
Class WidthUpper boundary − Lower boundary
Relative Frequencyf ÷ N
Percentage Frequency(f ÷ N) × 100
Frequency Densityf ÷ class width
HistogramAdjacent rectangles for continuous grouped data
Frequency PolygonMid-points plotted against frequency
OgiveClass boundaries plotted against cumulative frequency
Median by OgivesX-value of intersection of two ogives
Thirty-Second Recall

Trend → Line. Comparison → Bar. Composition → Pie. Continuous distribution → Histogram. Shape comparison → Polygon. Cumulative position → Ogive.