Module 1: Introduction β€” Notes

Statistics for Management (QAM601)

βœ… Notes below are pulled directly from the live course (video lectures, reading pages, interactive widgets), lesson by lesson β€” the SLM PDF is only a supplementary cross-check.

Lesson 1 β€” Learning Objectives

By the end of this module, learners should be able to: 1. Discuss statistical limitations, uses, types and statistical analysis. 2. Recognise tabular and graphical descriptive techniques using MS Excel β€” types of charts and the importance of graphical representation of data. 3. Explain numerical measures using MS Excel β€” arithmetic mean overview and use, and different types of measures. 4. Describe a relevant industry example and case study.

Lesson 2 β€” 1.1 Introduction (SLM p. 1)

  • Statistical analysis = the act of gathering vast amounts of data and then using statistics and other data-analysis methods to find trends, patterns and insights.
  • A career in statistical analysis can be a great fit for someone skilled with data and statistics; statistical analysts are in high demand due to the rise of big data, machine learning and technology.

Lesson 3 β€” 1.1.1 Statistical Analysis (SLM pp. 1–2; video)

  • Data = a collection of any number of connected observations. E.g. the total number of phones placed by multiple workers in a single day, or the total installed daily by one worker across multiple days.
  • A single observation is a data point; a collection of data is a data set.
  • Statistics covers every aspect of business, not just state-specific information; it is the scientific methods used to collect, arrange, condense and evaluate data, so as to make sound judgements and valid findings.
  • A forecast's precision and accuracy significantly affect a company's performance; statistics is a vital tool for market research and manufacturing control.

Business uses of statistical methods - Time and motion studies - Consumer behaviour studies - Credit ratings - Performance evaluations and pay - Inventory management - Accounting, quality control - Distribution channel planning, and many more

It is crucial for managers to understand statistical ideas and be proficient with statistical tools. A company's growth and the market uncertainty from less competition have significantly increased the need for statistical expertise. Previously, when businesses were smaller and simpler, decisions were made by a single person (manager/owner) based on assumptions and experience rather than a scientific foundation β€” e.g. deciding where to source raw materials, how much to produce, where to market it.

Lesson 4 β€” 1.1.2 Statistics: Limitations and Uses (SLM pp. 2–3; video)

Due to their versatility, statistical approaches are popular across many sectors, but statistics has limitations and is not a panacea β€” it cannot be used in every circumstance or answer every question.

Principal restrictions of statistics 1. Only covers issues that can be quantitatively stated and subjected to mathematical/numerical analysis. Qualitative data β€” employee integrity, consumer loyalty, emotional attachment, motivation, etc. β€” is not appropriate for these methods. 2. Statistics' sole focus is data collection; each individual item is given no weight. 3. Statistical findings are not mathematically exact β€” they are approximations; a random mistake is always possible. 4. Statistics should be used only after a thorough grasp of the procedure and conceptual underpinnings; improper use can mislead. 5. Statistical laws are not exact and are liable to be misused. 6. The greatest limitation: statistical data can be properly used only by a professional with thorough knowledge of statistical methods and proper training. 7. If data are not uniform/homogenous, the study of the problem is not possible β€” homogeneity of data is essential. 8. Statistical methods are not the only way to study a problem β€” a problem can be studied in many ways.

Lesson 5 β€” 1.1.3 Descriptive vs. Inferential Statistics (SLM pp. 3–4; video)

The study of statistics has two main branches: descriptive statistics and inferential statistics.

Descriptive statistics (3 hotspots) 1. Used to sum up and graph the data for a chosen category; helps understand a particular collection of observations. 2. A sample is defined by descriptive statistics β€” no confusion in concise numbers, since we simply identify the individuals/things being calculated. 3. Gives information that describes the data in some manner. Example: a pet shop sells cats, dogs, birds and fish; if 100 pets were sold and 35 were dogs, one description of the data is that 35% were dogs.

Inferential statistics (3 hotspots) 1. Techniques that let us use samples to generalise the populations the samples were drawn from; it is crucial the sample accurately represents the population. 2. The method used is called sampling; since inferential statistics draws conclusions from a sample and generalises to a population, the sample's accuracy affects the whole process. 3. At a broad level we must: (1) define the population being studied, (2) draw a representative sample from that population, (3) use analyses that incorporate the sampling error.

πŸ’‘ Tip

Descriptive = describes what the data shows; Inferential = predicts/generalises beyond the sample to the population.

Lesson 6 β€” 1.1.4 Statistics: Importance and Scope (SLM p. 4; video)

Importance/scope of statistics (5 accordions) 1. Condensation β€” statistics reduces many numbers to a little useful information (growth rate, average sales, the BSE index). A simple record of transactions cannot give an accurate assessment of profitability, but data on profit margins, ROI, EPS etc. can be quickly recalled and used for decisions. 2. Forecast β€” statistics helps forecasting by examining patterns; forecasts from intuition/instinct can hurt a business. Example: before investing, a petrochemical plant's refining capacity requires forecasting demand for the product mix, crude-oil supply, crude price and substitute goods 10–20 years ahead. 3. Testing of hypotheses β€” hypotheses are statements about population parameters based on past information, needing verification against recent developments. Drawing inductive conclusions from sample data carries risk, but sampling reduces the cost of decision-making; statistics gives a quantitative foundation for testing population-based ideas. 4. Relationship between facts β€” statistical approaches examine cause-and-effect relationships between two or more facts, e.g. between money supply, price level, and supply and demand. 5. Expectation β€” statistics is the foundation for formulating appropriate policies; the predicted value of current decisions' outcomes determines things like how much raw material to import, how much capacity to install, how many people to hire.

Lesson 7 β€” 1.1.5 Population and Sample (SLM p. 5; video, 2 tabs)

Sample tab - A sample consists of one or more observations drawn from the population; the group of people who actually took part in the research β€” those questioned (qualitative study) or who completed the survey (quantitative study). Potential participants who did not personally take part are not counted as part of the sample. - A sample data set is a part/subset of a population. - The size of a sample is always less than the population it is drawn from β€” uses the count n βˆ’ 1 in formulas.

Population tab - A population includes all elements of a set of data β€” the broader group we want to generalise results to. - A sample is just a part of the population; sample size depends on the exact population. - A population data set contains all members of a specified group (the entire list of possible data values) β€” uses the count n in formulas. - Example: "ALL people living in India." - Worked example: Mr. (Mrs.) Tom analyses final exam scores of students in his own math class for the past year. Since he has no reason to generalise to all management students in the school and has all the data relevant to his investigation, this is a population data set, not a sample.

πŸ’‘ Tip

Population = N (entire group); Sample = n (subset, always smaller). Population uses count n; sample uses count n βˆ’ 1 (degrees of freedom adjustment).

Lesson 8 β€” 1.2 Tabular and Graphical Descriptive Techniques Using MS Excel (SLM pp. 5–6; no video)

  • Data can be summarised in tabular (columnar/grid) form β€” cross-tabulation and frequency tables are the two most common tabular formats β€” or graphical form (charts and graphs).

The Most Popular Types of Data Visualisations (7-hotspot widget, all viewed): 1. Bar Plots β€” compare categories using rectangular bars whose length is proportional to the value. 2. Scatter Plots β€” plot pairs of numerical data, one variable per axis, to show relationship/correlation. 3. Histograms β€” show the distribution of a numerical variable by grouping data into bins. 4. Box Plots β€” summarise a distribution's median, quartiles, and outliers. 5. Pie Charts β€” show proportions of a whole as slices of a circle. 6. Pareto Charts β€” combine a bar chart (descending frequency) with a cumulative-percentage line. 7. Line Charts β€” show trends over a continuous interval, typically time.

Case study: Descriptive Analysis of MBA Students Data - Hypothetical Data Science job-interview scenario: analyse Mba_Students_Data.csv, a dataset of 200 MBA student records across 16 variables, using Python/R to produce tabular and graphical summaries.

First Five Entries from MBA Student Data Set:

Sl.No MBA Specialisation Avg Grades (3 Sem) Gender Age PG Degree Grad Degree Grad Pass Year Grad % 10+2 Pass Year 10+2 % 10+2 Stream 10th Pass Year 10th % Work Exp. In Maths Main Work Done in Last Company
1 Marketing 8.2 M 26 No B.Tech 2011 61.65 2008 85.67 Science 2008 80 44 IT & Consultancy
2 Marketing 8.5 F 21 No B.M.S 2016 76.75 2013 87.54 Commerce 2011 88.4 NA NA
3 Marketing 8.9 M 21 No B.Sc 2015 77.25 2012 75.07 Science 2010 84.2 0 NA
4 Marketing 8.2 F 23 No B.com 2014 84.21 2011 82.77 Commerce 2009 85.09 NA NA
5 Marketing 7.2 F 21 No B.M.S 2015 71 2012 71.45 Commerce 2010 78.36 6 Marketing and Sales
πŸ“˜ Note

No quiz block in this lesson.

Lesson 9 β€” 1.2.1 Importance of Graphical Representation of Data (SLM p. 6; video)

  • Data must be processed and analysed after being collected: recording, labelling, classifying and tabulating so it's consistent with the report.
  • Data can be viewed either in tabulation form or via charts; effective use of collected data depends on how it is organised, presented and summarised.
  • Graphs and diagrams are among the most visually appealing and persuasive ways to display statistical results.

Points justifying the widespread use of graphs and diagrams (4-tab widget): 1. Diagrams and graphs attract the eye. 2. They have more memorising effect. 3. They facilitate the easy comparison of data from one period to another. 4. Diagrams and graphs give a bird's eye view of the entire data; therefore, they convey meaning very quickly.

Lesson 10 β€” 1.2.2 Bar Chart (SLM pp. 6–10; video)

  • In a bar diagram, only the length of the bar is considered (the width is not) β€” a bar is essentially a thick line, so it's a one-dimensional diagram.

Simple Bar Diagram - Symbolises a single variable. - Comparative analysis is simple because bars share the same width and differ only in length (height). - Widely used in practice; can be horizontal or vertical. - Example use: sales, production, population numbers, etc. across different years.

Example 1 β€” Birth rate per thousand:

Country India Germany U.K. New Zealand Sweden China
Birth Rate 33 16 20 30 15 40

Course narration: China has the greatest birth rate, followed by India, while Sweden and Germany have (approximately) the lowest rates.

Sub-Divided Bar Diagram - Each bar is further divided into multiple components, each representing the magnitude of a given value; each component's share of the bar is proportional to its overall share. - Illustration 1 β€” number of students by faculty:

Year/Faculty Science Humanities Commerce
2014-2015 240 560 220
2015-2016 280 610 280

Multiple-Bar Diagram - Shows components as independent, adjacent bars indicating two or more sets of linked/related data. - The true value of each component is shown by the bar's height; different hues/tones distinguish the components. - Illustration 1 β€” number of students in two colleges by faculty:

College Arts Science Commerce Total
A 1200 800 600 2600
B 700 500 600 1800

Percentage Bar Diagram - The length of the entire bar is kept equal to 100; the segments of each bar may change to represent percentages of an aggregate. - Illustration 1 β€” population share by year:

Year Men Women Children
1995 45% 35% 20%
1996 44% 34% 22%
1997 48% 36% 16%

Quiz (mid-lesson, off-topic IQR question)

⚠️ Warning

One question tests IQR of [10, 12, 15, 18, 20, 22, 25] (n=7). Standard method (exclusive median, lower/upper half of 3 values each) gives Q1=12, Q3=22, IQR=10 β€” but the platform's answer key marks 7 as correct with only 2 attempts allowed. Flagged as problematic; the platform's expected method for this specific IQR calc could not be independently confirmed to match standard formulas.

Lesson 11 β€” 1.2.3 Pie Chart (SLM p. 10; video)

  • A pie chart (circle chart) is a circular statistical graphic divided into slices to illustrate a numerical proportion.
  • The arc length of each slice is proportional to the quantity it represents. Categories of data are shown as wedges in the circle, proportional in size to the percent of individuals in each category.
  • Very widely used in business and mass media, generally for percentage/proportional data β€” the percentage for each category is usually labelled next to its slice.
  • Pie charts are best for displaying data with around six categories or fewer.

Lesson 12 β€” 1.2.4 Histogram (SLM pp. 10–11; video)

  • A histogram is a graphical data display using bars of different heights β€” similar to a bar chart, but with ranges (bins/intervals) as its categories. The height of each bar shows how many observations fall within each range.

A histogram can be used when: - The data is numerical. - The shape of the data's distribution needs to be viewed, especially to check whether a process's output is distributed approximately normally. - Analysing whether a process can meet customer requirements. - Analysing a supplier's process output, or checking whether a process has changed between time periods. - Determining whether the outputs of two or more processes differ, and communicating the data's distribution quickly and easily.

Lesson 13 β€” 1.2.5 Frequency Polygon (SLM pp. 11–12; video)

  • A frequency polygon plots frequencies against the mid-points of class intervals, then joins the resulting points with line segments.
  • Compared to a histogram: in a frequency polygon, points replace the bars (rectangles).
  • Frequency polygons are better than histograms when several distributions need to be compared on the same graph.

Illustration 1 β€” Age distribution (histogram + frequency polygon):

Age in Years 10–20 20–30 30–40 40–50 50–60 60–70 70–80
Number of Persons 3 16 22 35 24 15 2

Scale along y-axis: 1 cm = 5 units. The frequency polygon line passes through the midpoint of each bar's top, extending down to the x-axis at the midpoints just before the first and just after the last interval.

Lesson 14 β€” 1.2.6 Ogives (SLM pp. 12–13; video)

  • Cumulative frequencies are obtained by adding frequencies progressively. An ogive (pronounced "o-jive") is the curve that results from plotting cumulative frequencies.
  • To build an ogive: (1) compute cumulative frequencies by adding progressive totals of the frequencies for each class; (2) plot cumulative frequencies on the y-axis and classes on the x-axis.
Type Method Result
Less than ogive Data sorted ascending; frequencies added starting at the top. Cumulative frequencies plotted against upper class limits. Positive (rising) curve
Greater than ogive Frequencies added from the bottom (or deducted from the total, from the top). Cumulative frequencies plotted against lower class limits. Negative (falling) curve
Uses Ogives find the median, quartiles, quartile deviation, coefficient of skewness, etc., and are useful for comparing two distributions. β€”

Illustration 1 β€” frequency distribution:

CI 0–20 20–40 40–60 60–80 80–100 100–120 120–140 140–160
f 5 12 18 25 15 12 8 5

Cumulative table (lcf = less-than cumulative frequency, mcf = more-than/greater-than cumulative frequency):

Size lcf mcf
20 5 100
40 17 95
60 35 83
80 60 65
100 75 40
120 87 25
140 95 13
160 100 5

The less-than and more-than ogive curves intersect at approximately (73, 47) β€” the x-coordinate of this intersection gives the median graphically.

Lesson 15 β€” 1.2.7 Pareto Chart (SLM p. 14; video)

  • A Pareto chart graphs the frequency of defects and their cumulative effect. It helps identify which defects should be prioritised to achieve the greatest overall change.
  • Pareto principle (80/20 rule, law of the vital few, principle of factor sparsity): for many events, roughly 80% of the effects come from 20% of the causes.
  • Example: "Pareto Chart of Late Arrivals by Reported Cause" β€” bars (Traffic, Child Care, Public Transport, Weather, Overslept, Emergency) ranked by frequency, with a cumulative-percentage line overlaid.

When to use a Pareto Chart: - When analysing data about the frequency of problems or causes in a process. - When there are many problems/causes and you want to focus on the most significant ones. - When analysing broad causes by looking at their specific components. - When communicating with others about the data.

Lesson 16 β€” 1.2.8 Stem-and-Leaf Display (SLM pp. 14–15; video)

  • A stem-and-leaf display (plot) is a tool for visualising the shape of a distribution that presents quantitative data graphically, similar to a histogram β€” useful in exploratory data analysis.
  • The "leaf" is typically the last digit of each data value; the "stem" is the first digit(s). Leaf values are listed to the right (or left) of the stem values (which are listed down the table), grouping individual scores under their stem.
  • It's OK to repeat a leaf value (e.g. two 2.5s both appear as leaf "5" under stem "2").

Example β€” long jump results: 2.3, 2.5, 2.5, 2.7, 2.8, 3.2, 3.6, 3.6, 4.5, 5.0

Stem Leaf
2 3 5 5 7 8
3 2 6 6
4 5
5 0

Stem "2" Leaf "3" means 2.3. Each leaf here is a decimal digit; 5.0 has a leaf of "0".

Lesson 17 β€” 1.2.9 Cross Tabulations (SLM p. 15; video)

  • Cross-tabulation is a method for quantitatively analysing the relationship between multiple variables. Cross-tab tables are also called contingency tables, and this analysis is also known as crosstab, chi-square, or data tabulation. It shows how correlations vary from one group to another.
  • A cross-tabulation (crosstab) is a two- (or more) dimensional table recording the frequency of respondents having the specific characteristics described in each table cell.

Importance of Cross Tabulation (2-tab widget): - Clean and Useable Data β€” cross tabulation makes it simple to interpret data; the clarity it offers helps deliver clean data usable to improve decisions across an organisation. - Easy to Understand β€” no advanced statistical degree is needed to interpret cross tabulation; results are easy to read and explain, making it useful in any type of presentation.

Lesson 18 β€” 1.2.10 Scatter Plot and Trend Line (SLM pp. 16–17; video) β€” final lesson of Β§1.2

  • A scatter diagram is the simplest graph for illustrating the relationship between two variables β€” a straightforward representation of a bivariate distribution. Each data pair (xα΅’, yα΅’) is plotted as a point (x = abscissa, y = ordinate).
  • A scatter diagram gives two useful pieces of information: (1) the degree to which the variables are connected, and (2) the relationship/estimating equation that characterises it, if related.
  • If dots cluster around a line, the correlation is linear; if they cluster around a curve, it's non-linear (curvilinear).
  • Positive correlation: dots align from lower-left to upper-right (perfect positive β†’ all dots on a straight line).
  • Negative correlation: dots align from upper-left to lower-right (perfect negative β†’ all dots on a straight line).
  • It is not required to use the minimum values of the variables when producing a scatter diagram β€” the zero values of X and Y may be taken at the point of origin.
πŸ“˜ Note

Correlation answers whether an association exists between two variables and to what extent. Regression answers what relationship (equation) describes that association.

Example β€” Advertisement cost (X, '000 β‚Ή) vs Sales (Y, Lakh β‚Ή) for 10 years:

Advertisement cost ('000) 40 65 60 90 85 75 35 90 34 76
Sales (Lakh) 45 56 58 82 65 70 64 85 50 85

The scatter diagram of this data shows a positive correlation β€” sales generally rise with advertisement cost.

Lesson 19 β€” 1.3 Numerical Measures Using MS Excel (SLM pp. 17–18; no video) β€” start of Β§1.3

  • Excel's built-in mathematical/statistical functions let users perform a variety of statistical calculations and analyses to gain insights and make decisions.

Typical categories of Excel statistical functions (8-item accordion, all expanded): 1. Descriptive Statistics β€” AVERAGE, COUNT, SUM, MIN, MAX, MEDIAN, MODE, STDEV, VAR (description and summarisation of data). 2. Correlation and Regression β€” CORREL, COVAR, RSQ, TREND (analyse relationships between variables, calculate correlation coefficients, run regression). 3. Hypothesis Testing β€” T.TEST, Z.TEST, F.TEST (compare sample means, proportions, or variances). 4. Probability Distribution β€” NORM.DIST (normal), BINOM.DIST (binomial), POISSON.DIST (Poisson), and others. 5. Sampling β€” RAND, RANDBETWEEN (simulations, random numbers/selections). 6. Quartiles and Percentiles β€” QUARTILE, PERCENTILE (identify dataset quartiles/percentiles). 7. Frequency Analysis β€” HISTOGRAM, FREQUENCY (visualise distribution patterns, create frequency distributions). 8. Time Series Analysis β€” FORECAST, GROWTH, TREND (time series forecasting and trend analysis).

πŸ“˜ Note

No quiz in this lesson.

Lesson 20 β€” 1.3.1 Arithmetic Mean: Overview and Use (SLM pp. 18–22; video)

  • The arithmetic mean (average) is the sum of all observations divided by the number of observations.
  • Formula: xΜ„ = (1/n)Ξ£xα΅’

Properties of Arithmetic Mean: 1. The sum of all deviations of x's values from their arithmetic mean equals zero. 2. Multiplying the arithmetic mean by the number of items gives the sum of all items. 3. When multiple groups are given, an aggregate (combined) arithmetic mean can be calculated.

Demerits of Arithmetic Mean: 1. Affected by extreme values. 2. Cannot be found graphically or by visual inspection. 3. Cannot be calculated if even one observation is missing/lost. 4. Cannot be calculated when data has open-end class intervals.

Arithmetic Mean for Ungrouped Data β€” Individual Series

1. Direct Method: XΜ„ = Ξ£x / N β€” add up all values, divide by count.

Illustration 1: Values: 125, 128, 132, 135, 140, 148, 155, 157, 159, 191 (N=10). Ξ£x = 1440. XΜ„ = 1440/10 = 144.

2. Short-Cut (Indirect) Method: X = A + Ξ£d/n, where A = assumed average, Ξ£d = sum of deviations from A, n = number of items.

Illustration 1: Marks: 43, 48, 65, 57, 31, 60, 37, 48, 78, 59 (N=10). Assumed average A = 60. Deviations d = -17, -12, 5, -3, -29, 0, -23, -12, 18, -1 β†’ Ξ£d = -74. X = 60 + (-74/10) = 52.6 marks.

Combined Arithmetic Mean

  • The arithmetic mean of two or more related groups combined is the combined mean: XΜ„combined = (n₁X̄₁ + nβ‚‚XΜ„β‚‚) / (n₁ + nβ‚‚).

Example: Branch A: 500 workers, avg salary 300. Branch B: 1,000 workers, avg salary 250. Combined mean = [500(300) + 1000(250)] / (500+1000) = (150000+250000)/1500 = 266.66.

Weighted Arithmetic Mean

  • When some observations matter more than others, each value is weighted by its importance: multiply each item by its weight, then divide by the sum of weights.
  • Formula: XΜ„w = Ξ£wx / Ξ£w

Example: Monthly rice price (β‚Ή/ton) vs. quantity purchased ('000 tons):

Month Jan Feb March April May June
Price 42.5 51.25 50 52 44.25 54
No. of tones 25 30 40 50 10 45

Ξ£w (tones) = 200, Ξ£X (price) = 294, Ξ£wx = 10027.5. Simple mean = 294/6 = 49. Weighted mean = 10027.5/200 = 50.137 (β‰ˆ β‚Ή50.30). The correct average price paid is ~β‚Ή50.30, not β‚Ή49 β€” i.e., the weighted arithmetic mean is more accurate than the simple arithmetic mean here since it accounts for the varying quantities purchased each month.

Lesson 21 β€” 1.3.2 Median: Basics and Application (SLM pp. 22–23; video)

  • The median is the value that divides a series into two half-equal parts β€” one half containing all values ≀ it, the other β‰₯ it. It's the variable's "central value." To find it, values must be sorted by size/magnitude.
  • The median is a positional average (locative average) β€” its location within the sorted series equals the number of items on either side.

Merits of Median: - Has a strict definition. - Simple to compute and comprehend. - Can be located graphically. - Unlike the arithmetic mean, unaffected by extreme values. - Can be found by microscopic inspection. - Can be determined even if extreme values are unknown, as long as the number of items is known β€” useful for qualitative studies.

Demerits of Median: - For individual observations, values must be arranged by size to locate the median β€” tedious if the number of items is large. - Multiplying the median by the number of items does NOT give the total value of all items (unlike the arithmetic mean). - Not suitable for complex algebraic/mathematical treatment. - More affected by sampling fluctuations.

Application example: Determine the median from: 25, 15, 23, 40, 27, 25, 23, 25, 20 (n=9). - Sorted ascending: 15, 20, 23, 23, 25, 25, 25, 27, 40. - Median = middle (5th) value = 25.

πŸ’‘ Tip

Median is preferred over the mean when data has outliers or is skewed, since extreme values don't distort it.

Lesson 22 β€” 1.3.3 Mode: Fundamentals and Usage (SLM pp. 24–26; video)

  • "Mode" derives from the French "la mode" (fashion) β€” the most fashionable/typical item in a series. Croxtan and Cowden call it "the most typical of a series of values," summing up group characteristics more satisfactorily than the mean or median in some cases.
  • The mode is the value that appears most frequently in a distribution β€” the item size occurring most often in a sequence.

Merits of Mode: - Usually has an actual value (unlike some computed averages). - In discrete series, easily located by inspection. - Not affected by extreme values. - Can be determined even if extreme values aren't given. - Easy to understand β€” used in everyday speech.

Demerits of Mode: - Not based on all observations of the data. - In some cases, more than one mode exists in a series. - Multiplying mode by the number of items does NOT give the total value of items. - Doesn't truly represent the group if a small number of items of the same size sit within a large group of differently-sized items. - Not suitable for further mathematical treatment.

Applications of Mode

1. Individual Series β€” mode found by inspection (the most frequent value).

Illustration: Data: 7, 12, 8, 5, 9, 6, 10, 9, 4, 9, 9. On inspection, 9 has the maximum frequency (repeated 4 times) β†’ Mode (Z) = 9.

2. Discrete Series β€” calculated via grouping and analysis tables. - Grouping table: 6 columns β€” col 1 is frequency, cols 2–3 are two-way grouped frequencies, cols 4–6 are three-way grouped frequencies. - Analysis table: 2 columns β€” tally bar and frequency. - Steps: (1) group frequencies by two; (2) leave the first frequency and group the rest in twos; (3) group frequencies in threes; (4) leave the first size's frequency and add the rest in threes; (5) leave the first two sizes' frequencies and add the rest in threes; (6) prepare the analysis table to find which size occurs the maximum number of times β€” that size is the mode.

3. Continuous Series β€” find the modal class (group with maximum frequency, via inspection; grouping/analysis tables if multiple classes tie), then calculate the actual mode value:

Formula: Mo = L + [(fm βˆ’ f1) / (2fm βˆ’ f1 βˆ’ f2)] Γ— h Where L = lower boundary of modal class, fm = frequency of modal class, f1 = frequency of class preceding modal class, f2 = frequency of class following modal class, h = class width.

Lesson 23 β€” 1.3.4 Partition Values: Quartiles and Percentiles (SLM p. 26; video)

Percentile - A percentile is the value below which a given proportion of data falls. - Example: out of 20 people, you're the 4th tallest β†’ you're shorter than 80% of people β†’ you're at the 80th percentile. If your height is 1.65m, "1.65m" is the 80th percentile height in that group.

Quartiles - Quartiles split data into quarters β€” values dividing a data table into four groups with an approximately equal number of observations. The total 100% splits into 25%, 50%, 75%, 100%. - Quartile 1 (Q1) = 25th percentile; Quartile 2 (Q2) = 50th percentile (= median); Quartile 3 (Q3) = 75th percentile.

Example: For 1, 3, 3, 4, 5, 6, 6, 7, 8, 8: 25th percentile = 3, 50th percentile = 5.5, 75th percentile = 7.

Computing percentiles/quartiles: 1. Compute the f-value of each value in the data table (based on index i and count n). 2. Q1 is found by interpolating between the f-values immediately below and above 0.25. 3. Q3 is found by interpolating between the f-values immediately below and above 0.75. 4. Any other percentile is calculated similarly, by interpolating between the appropriate values.

Lesson 24 β€” 1.3.5 Measures of Dispersion: Range and Application (SLM pp. 27–29; video)

  • A measure of variation/dispersion reveals how much numerical values deviate from the average. When variation is minor, the average alone represents the data well; for significant differences, a dispersion measure adds useful information.
  • Determining dispersion is useful to: compare present vs. past outcomes; compare two or more observational sets; recommend ways to manage data variance. Dispersion also reveals consistency/uniformity β€” the foundation of quality-control techniques.

Absolute vs. Relative Measures of Dispersion: - Absolute dispersion is expressed in the original data's units (e.g. marks). It can't be used to compare series expressed in different units. - Relative (coefficient) of dispersion is the ratio/percentage of an absolute dispersion measure to a suitable average β€” its key benefit is enabling comparison across series with different units. - A good measure of dispersion should share the same desirable characteristics as a good measure of central tendency.

Range - Definition: the difference between the largest and smallest values of the data. - Formula: Range = L βˆ’ S (L = largest value, S = smallest value). - In individual observations and discrete series, L and S are directly identified. In continuous series, two methods apply: - Method 1: L = upper boundary of the highest class; S = lower boundary of the lowest class. - Method 2: L = mid-value of the highest class; S = mid-value of the lowest class. - Coefficient of Range = (L βˆ’ S) / (L + S) β€” the relative measure, used when comparing two data sets.

Example: Data: 10, 5, 8, 11, 12, 9. L = 12, S = 5. Range = 12 βˆ’ 5 = 7. Coefficient of range = (12βˆ’5)/(12+5) = 7/17 β‰ˆ 0.41.

Lesson 25 β€” 1.3.6 Inter Quartile Range: Concept and Use (SLM pp. 29–30; video)

  • Inter-Quartile Range (IQR) = difference between the upper quartile (Q3) and lower quartile (Q1): IQR = Q3 βˆ’ Q1.
  • Quartile Deviation (QD) = the average of the difference between Q3 and Q1: QD = (Q3 βˆ’ Q1) / 2. QD also equals the average deviation of Q1 and Q3 from the median.
  • Coefficient of QD (relative measure) = (Q3 βˆ’ Q1) / (Q3 + Q1).

Example β€” Weekly wages of labourers:

Weekly wages 100 200 400 500 600 Total
No. of weeks 5 8 21 12 6 52

Cumulative frequency: 5, 13, 34, 46, 52 (N = 52).

Solution: - Q1 position = (N+1)/4 = 53/4 = 13.25 β†’ Q1 = 13th value + 0.25(14th value βˆ’ 13th value) = 200 + 0.25(400βˆ’200) = 250. - Q3 position = 3(N+1)/4 = 39.75 β†’ Q3 = 39th value + 0.75(40th value βˆ’ 39th value) = 500 + 0.75(500βˆ’500) = 500. - QD = (500 βˆ’ 250)/2 = 125. - Coefficient of QD = (500βˆ’250)/(500+250) = 250/750 = 0.333.

Lesson 26 β€” 1.3.7 Standard Deviation and Variance (SLM pp. 30–34; video)

Variance - The average squared deviation of data points from the mean. - Sample variance σ²ₓ: sum of squared deviations from the mean divided by n βˆ’ 1. Population variance σ²: divided by N.

Sample Variance: σ²ₓ = Ξ£α΅’β‚Œβ‚βΏ(xα΅’ βˆ’ xΜ„)Β² / (n βˆ’ 1) Population Variance: σ² = Ξ£(xα΅’ βˆ’ ΞΌ)Β² / N = [Ξ£xα΅’Β²/N] βˆ’ ΞΌΒ² Alternate form: Var(x) = E(XΒ²) βˆ’ [E(X)]Β²

Standard Deviation (SD) - The root mean square deviation of values from their arithmetic mean; symbol Οƒ (sigma). SD is the positive square root of the variance β€” also called the RMS of the deviations. - SD of a sample = square root of sample variance; SD of a population (Οƒ) = square root of population variance.

Characteristics of SD: - The most significant and extensively applied variability measure. - Every observation forms its basis (uses all data points). - Further mathematical analysis is feasible with it. - Least affected by sampling fluctuations. - Extreme values affect it β€” values deviating from the mean are weighted more heavily. - Main drawback: can't compare variability across data sets given in different units.

Formula for calculating SD: Οƒ = √[(Ξ£xΒ²/n) βˆ’ (Ξ£x/n)Β²] Step-deviation (short-cut) formula for grouped data: Οƒ = √[(Ξ£fdΒ²/n) βˆ’ (Ξ£fd/n)Β²] Γ— C, where d = X βˆ’ A (A = assumed mean), C = true class interval, N = total frequency.

Example β€” find SD:

Class Interval 0–10 10–20 20–30 30–40 40–50 50–60 60–70
Frequency 6 14 10 8 1 3 8

Class marks (mα΅’): 5,15,25,35,45,55,65. Ξ£fi = 50, Ξ£fiΒ·mi = 1500, Ξ£fiΒ·diΒ² = 19250 (di = mi βˆ’ A). Mean = 1500/50 = 30. SD = √(19250/50) = 19.62.

Combined Standard Deviation

For a combined group of (n₁ + nβ‚‚) items with means x̄₁, xΜ„β‚‚ and SDs σ₁, Οƒβ‚‚: - Combined mean: XΜ„ = (n₁x̄₁ + nβ‚‚xΜ„β‚‚) / (n₁ + nβ‚‚) - Combined SD: Οƒ = √[(n₁σ₁² + nβ‚‚Οƒβ‚‚Β² + n₁d₁² + nβ‚‚dβ‚‚Β²) / (n₁ + nβ‚‚)], where d₁ = x̄₁ βˆ’ XΜ„, dβ‚‚ = xΜ„β‚‚ βˆ’ XΜ„ (or their absolute values). - Extends to 3 samples: Οƒ = √[(n₁σ₁² + nβ‚‚Οƒβ‚‚Β² + n₃σ₃² + n₁d₁² + nβ‚‚dβ‚‚Β² + n₃d₃²) / (n₁ + nβ‚‚ + n₃)].

Lesson 27 β€” 1.3.8 Relative Dispersion: Coefficient of Variation (SLM pp. 34–35; video)

  • The Coefficient of Variation (CV) is the ratio of standard deviation to mean, multiplied by 100 (expressed as a percentage): CV = (SD / Mean) Γ— 100.
  • CV is a measure of relative dispersion/variability. A smaller CV indicates greater stability and lesser variability.
  • Since it's a normalised (unit-free) measure, CV allows comparison of variability between datasets with different units or widely different means β€” something absolute measures like SD can't do.

Example:

Company A Company B
Mean annual sales $50,000 $200,000
SD of annual sales $5,000 $40,000
  • CV_A = (5,000/50,000) Γ— 100 = 10%
  • CV_B = (40,000/200,000) Γ— 100 = 20%

Interpretation: Despite Company B having far higher mean sales, its sales are more variable relative to its mean (CV 20%) than Company A's (CV 10%) β€” Company A is relatively more stable/consistent. CV standardises the comparison of variability between the two companies despite their different scales.

Lesson 28 β€” 1.3.9 Relevant Industry Example (SLM pp. 35–36; video)

Example β€” Two plants of a factory, C and D:

Plant C Plant D
No. of workers 5,000 6,000
Average monthly wages $2,500 $2,500
Standard deviation 9 10

To determine: which plant has more wage variability per worker, using CV (higher CV = more variable).

  • CV_C = (9/2500) Γ— 100 = 0.36%
  • CV_D = (10/2500) Γ— 100 = 0.4%
  • Since CV_D (0.4) > CV_C (0.36), Plant D's individual salary variability is higher, even though both plants have the same average wage.
πŸ“˜ Note

Industry applications: in retail, standard deviation helps analyse demand volatility; in manufacturing, low variability in measurements indicates high quality and consistency (process control).

Lesson 29 β€” 1.3.10 Case Study (SLM pp. 36–37; video) β€” final content lesson of Module 1

Case Study: Interpretation of Interquartile Range - IQR is used to understand dispersion/spread of a dataset β€” especially useful for skewed distributions or datasets with outliers.

Scenario: A retail electronics business surveys 1,000 customers on satisfaction (1–10 scale) with a new smartphone.

Analysis: - Data sorted ascending; Q1 (25th percentile) and Q3 (75th percentile) computed: - Q1 = 6, Q3 = 8 - IQR = Q3 βˆ’ Q1 = 8 βˆ’ 6 = 2 - Interpretation: The middle 50% of customer satisfaction ratings falls within a 2-point range, from 6 to 8.

Uses of IQR (3-item accordion, all expanded): 1. Finding Outliers β€” a data point is an outlier if it's above Q3 + 1.5Γ—IQR or below Q1 βˆ’ 1.5Γ—IQR. 2. Comparing Variability β€” comparing IQRs across datasets reveals information about their relative spread/variability. 3. Robust Measure β€” IQR is a robust measure of spread because it's unaffected by extreme values/outliers.

Conclusion: IQR is a robust statistical tool for understanding data distribution β€” helps analysts assess variability and spot potential outliers across domains.

Lesson 30 β€” Module 1 Summary

  • Data is a collection of any number of connected observations (e.g. total phones placed by multiple workers in a day, or phones installed by one worker across multiple days).
  • Statistics has two main branches: descriptive statistics and inferential statistics.
  • Pareto principle (80/20 rule): roughly 80% of effects come from 20% of causes.
  • A stem-and-leaf display visualises the shape of a distribution graphically, similar to a histogram β€” useful for exploratory data analysis.
  • Cross tabulation quantitatively analyses the relationship between multiple variables; also called contingency tables or cross tabs.

Lesson 31 β€” Module 1 Glossary

  • Sample: One or more observations drawn from the population β€” the group of people who took part in the research.
  • Population: All elements from a set of data β€” the broader group results are generalised to.
  • Frequency Polygon: Frequencies plotted against the mid-points of class intervals, with the points joined by line segments.
  • Range: The difference between the largest and smallest values of data.
  • Histogram: Shows numerical data dispersion via a sequence of bars (bins), each bar's height representing the frequency/count of observations in that bin.

Lesson 32 β€” Check Your Understanding (graded self-check, 20 questions)

  • Scored 95% (19/20) β€” passing threshold 0%, marked Passed.
  • 19 questions answered correctly on first attempt (no retry option offered on this quiz β€” straight to the next question after each submit).
  • Q1 was marked wrong by the platform's answer key: it says "Statistics can only analyse quantitative data" is the true statement, which is factually incorrect β€” statistics analyses both quantitative and qualitative data (the option I selected: "Statistics can analyse both quantitative and qualitative data"). Flagged as a platform answer-key error; did not affect the Passed outcome.
πŸ“˜ Note

Questions 2–20 covered: condensation of data, tabular summarisation (cross-tabulation & frequency tables), percentage bar diagrams (bar length = 100%), ogives (cumulative frequency vs. data value), descriptive statistics' goal, limitations of statistical analysis, inferential vs. descriptive statistics examples, samples vs. populations, homogeneity of data, key business applications of statistics, frequency distribution tables, qualitative vs. quantitative data types, sample statistics vs. population parameters, and correlation analysis for studying relationships between variables β€” all answered correctly and consistent with earlier lesson notes in this module.

Lesson 33 β€” Exercise (self-study prompts)

Attempt these questions in your notebook to strengthen concepts and writing skills:

  1. Explain which branch of statistics focuses on summarising and describing the characteristics of a dataset without making inferences about a larger population? (β†’ Descriptive statistics β€” see Lesson 6.)
  2. Explain the concepts of population and sample in statistics and discuss their significance in research and decision-making processes. (β†’ Lesson 7.)
  3. Discuss the importance of graphical representation of data in statistics and data analysis. (β†’ Lesson 9.)
  4. Compare and contrast histograms and frequency polygons as graphical representations of frequency distributions in statistics. (β†’ Lessons 12–13.)
  5. Discuss the significance and applications of Pareto Charts in quality management and process improvement. (β†’ Lesson 15.)

Lesson 34 β€” Learning Activities (self-study prompts)

Attempt these questions in your notebook to strengthen concepts and writing skills:

  1. Have a group discussion about the similarities and differences between cross-tabulation and frequency tables. In what scenarios would you prefer to use cross-tabulation over frequency tables, and vice versa? (β†’ Lessons 8, 17.)
  2. Design a simple bar diagram to compare the sales figures of different products for a given period, such as a month or a quarter. (β†’ Lesson 10.)

Lesson 35 β€” Further Readings and Bibliography (reference material)

Core reference texts for Statistics for Management:

  1. Levin, R. I., Rubin, D. S., Rastogi, S., & Siddiqui, M. H. (2016). Statistics for Management. Pearson Education, 7th Edition. - Primary course textbook (same authors as module material).

  2. Mann, P. S. (2016). Introductory Statistics, 7th Edition. Wiley India. - Complementary introductory-level resource for foundational concepts.

  3. James, G., Witten, D., Hastie, T., & Tibshirani, R. (2016). An Introduction to Statistical Learning with Applications in R. Springer. - Bridges theory to computational practice; excellent for R-based learning.

  4. Srivastava, V. K. et al. Quantitative Techniques for Managerial Decision Making. Wiley Eastern Ltd. - Focus on applied quantitative methods for business decisions.

  5. Levin, R. I., & Kirkpatrick, C. A. Quantitative Approaches to Management. McGraw Hill, Kogakusha Ltd. - Strategic quantitative methods for management-level problems.

  6. Mann, P. S. (2016). Introductory Statistics, 7th Edition. Wiley India. - (Reiterated for emphasis; same as ref 2.)

  7. Budnik, F. S., Mcleaavey, D., & Mojena, R. Principles of Operation Research. AITBS New Delhi. - Operations research foundations for advanced quantitative modeling.

  8. Sharma, J. K. Operation Researchβ€”Theory and Applications. McMillan, New Delhi. - Comprehensive OR reference for optimization and decision science.

Note: These references collectively cover descriptive statistics, probability, distributions, inference, decision-making frameworks, and computational methods β€” the full scope of Module 1's "Introduction" content.


πŸ“‹ To-Do / Gaps

  • (none yet)

✨ Quick Revision Pack (Module 1)

60-second summary

  • Descriptive statistics summarise data (tables, charts, averages, spread); inferential statistics draw conclusions about a population from a sample.
  • Charts: bar (compare categories), pie (share of a whole), histogram (continuous frequency), frequency polygon, ogive (cumulative), Pareto (80/20), stem-and-leaf, scatter plot (relationship).
  • Central tendency: mean (uses all values, hit by outliers), median (middle value, robust), mode (most frequent).
  • Dispersion: range, IQR (Q3 - Q1), variance, standard deviation, coefficient of variation (CV = Οƒ/mean x 100) for comparing variability across data sets.
  • Relationship: mean - mode = 3(mean - median) for moderately skewed data.

Diagrams

Which chart?

flowchart TB
  Q{"What do you want to show?"} -->|"Compare categories"| B["Bar chart"]
  Q -->|"Parts of a whole"| P["Pie chart"]
  Q -->|"Distribution of continuous data"| H["Histogram / frequency polygon"]
  Q -->|"Cumulative totals, median, quartiles"| O["Ogive"]
  Q -->|"Relationship between two variables"| S["Scatter plot + trend line"]
  Q -->|"Biggest causes first"| PA["Pareto chart"]

Choosing a measure of centre

Situation Best measure
Symmetric numerical data Mean
Skewed data or outliers Median
Categorical / most common value Mode
Rates and ratios Harmonic mean
Growth rates Geometric mean

Shape and the averages

flowchart LR
  A["Negative skew
mean < median < mode"] --- B["Symmetric
mean = median = mode"] --- C["Positive skew
mode < median < mean"]

Memory tricks

πŸ’‘ Tip

Mean, Median, Mode = "average, middle, most". In a right-skewed distribution the mean is pulled toward the tail (largest of the three).

πŸ’‘ Tip

Dispersion order of sensitivity: Range (uses only 2 values, most affected by outliers) β†’ IQR (middle 50%) β†’ SD (uses all values).

❗ Important

Variance = (SD)Β²; sample variance divides by n - 1, population by N. CV = SD / Mean x 100% β€” the lower CV means more consistent.

πŸ“˜ Note

Stem-and-leaf keeps the original digits; histogram groups them (bars touch because data are continuous; bar charts have gaps).

⚠️ Warning

A pie chart is poor for many categories or for comparing close values; use a bar chart.

Formula sheet

Measure Formula
Arithmetic mean xΜ„ = Ξ£x / n
Weighted mean Ξ£wx / Ξ£w
Combined mean (n1xΜ„1 + n2xΜ„2)/(n1 + n2)
Median (grouped) L + [(n/2 - cf)/f] x h
Mode (grouped) L + [(f1 - f0)/(2f1 - f0 - f2)] x h
Range / coefficient Max - Min ; (Max - Min)/(Max + Min)
IQR Q3 - Q1
Variance σ² = Ξ£(x - ΞΌ)Β² / N ; sΒ² = Ξ£(x - xΜ„)Β² / (n - 1)
Coefficient of variation CV = (Οƒ / xΜ„) x 100

Practice problems (with answers)

  1. Mean: 12, 15, 18, 20, 25 β†’ Ξ£ = 90, n = 5 β†’ 18.
  2. Median: 12, 15, 18, 20, 25, 30 β†’ (18 + 20)/2 = 19.
  3. Mode: 3, 5, 5, 7, 5, 9 β†’ 5.
  4. Weighted mean: marks 80, 90, 70 with weights 2, 3, 1 β†’ (160 + 270 + 70)/6 = 83.33.
  5. Standard deviation: data 2, 4, 6, 8 β†’ mean 5; squared deviations 9 + 1 + 1 + 9 = 20; population variance = 20/4 = 5, SD = 2.236.
  6. Which is more consistent? A: mean 50, SD 5; B: mean 80, SD 10. β†’ CV(A) = 10%, CV(B) = 12.5% β†’ A is more consistent.
  7. Empirical relation: mean = 40, median = 38 β†’ mode β‰ˆ 3(38) - 2(40) = 34.
  8. Combined mean: group 1: n = 20, mean 60; group 2: n = 30, mean 70 β†’ (1,200 + 2,100)/50 = 66.
  9. Short answer (5 marks): Why is the median preferred to the mean for incomes? β†’ A few very high incomes pull the mean up; the median is the middle value and is not affected by extremes.