π Lecture 23 β Statistical Data Representation
π Overview: This lecture introduces the methods for organizing and presenting statistical data in understandable formats. It covers the basis for classification, types of classification, methods of presentation, and various graphic representations including pictographs, sector graphs, column/bar graphs, and line graphs.
ποΈ Topics Covered
The lecture begins with a review and overview of Module 5 on statistical data representation. It then covers the basis for classification of data (qualitative, quantitative, geographical, chronological), types of classification (one-way, two-way, three-way), methods of presentation (text, semi-tabular, tabular, graphic), and types of graphs (pictograph, sector graph, column/bar graph, line graph) with examples.
π Lecture Summary
STATISTICAL DATA
Information is collected by government departments, market researchers, opinion pollsters and others. This information then has to be organized and presented in a way that is easy to understand.
BASIS FOR CLASSIFICATION
There are several bases for classifying data:
- Qualitative: Attributes such as sex, religion
- Quantitative Characteristics: Heights, weights, incomes, etc.
- Geographical: Regions such as provinces, divisions, etc.
- Chronological or Temporal: By time of occurrence, also known as time series
TYPES OF CLASSIFICATION
Different types of classifications exist based on how many characteristics are considered at once:
- One-way: Classifying by one characteristic, e.g., population
- Two-way: Classifying by two characteristics at a time
- Three-way: Classifying by three characteristics at a time
METHODS OF PRESENTATION
Data can be presented using different methods:
- Text: Narrating data in words, e.g., "The majority of population of Punjab is located in rural areas."
- Semi tabular: Presenting data in rows
- Tabular: Using tables with rows and columns
- Graphic: Using charts and graphs
TYPES OF GRAPHS
Various graphic representations are available:
- Picture graph (Pictograph)
- Column Graphs
- Line Graphs
- Circle Graphs (Sector Graphs)
- Conversion Graphs
- Travel Graphs
- Histograms
- Frequency Polygon
- Cumulative Polygon or Ogive
PICTURE GRAPH or PICTOGRAPH
In a picture graph or pictograph, each value is represented by a proportional number of pictures. In the example below, one car represents 10 cars.
Example: To represent 50 cars, you would draw 5 car icons, where each icon stands for 10 cars.
SECTOR GRAPHS
Sector graphs use the division of a circle into different sectors. The full circle is 360 degrees. For each percentage, degrees are calculated and sectors plotted.
π Definition β Sector Graph: A circular graph where the circle is divided into sectors proportional to the percentage of each category.
π Formula: Degrees for a sector = (Percentage / 100) Γ 360Β° β Plain-English meaning: To find how many degrees a category takes in the circle, multiply its percentage by 3.6.
π Example: If a category represents 25% of data, its sector angle = (25/100) Γ 360Β° = 90Β°
COLUMN AND BAR GRAPHS
The lecture presents an example showing the Proportion of households by size in the form of a Column and Bar graph. In a bar graph, rectangular bars represent data, with the length or height of each bar proportional to the value it represents.
LINE GRAPHS
Line graphs are the most commonly used graphs. A line graph plots data as points and then joins them with a line.
Example: A line graph showing the number of road accidents over the course of a year, with months on the x-axis and number of accidents on the y-axis, with data points connected by straight lines.
π‘ Why this matters: Line graphs are ideal for showing trends and changes over time, making them essential for analyzing time series data in business and economics.
β Key Takeaways
The most important concepts from this lecture are the different bases for classifying data (qualitative, quantitative, geographical, chronological) and the types of classification (one-way, two-way, three-way). Students must understand the methods of data presentation including textual, semi-tabular, tabular, and graphic forms. For graphs, focus on pictographs where pictures represent quantities, sector graphs that divide a circle into 360 degrees based on percentages, column/bar graphs for comparing categories, and line graphs for showing trends over time. Remember that sector graph angles are calculated by multiplying the percentage by 3.6 (360/100), and that the choice of graph depends on the type of data being presented.
π§ Quick Revision Questions
- What are the four main bases for classification of data mentioned in the lecture?
- How do you calculate the angle of a sector in a sector graph if a category represents 40% of the total?
- What is the difference between a one-way classification and a two-way classification?
- In a pictograph where one icon represents 10 units, how many icons would you need to represent 75 units?
- Which type of graph is most appropriate for showing changes in data over time?
π Lecture 24 β Statistical Representation Measures of Central Tendency Part 1
π Overview: This lecture introduces key concepts in descriptive statistics, focusing on how to represent data visually using line graphs and how to summarize data using measures of central tendency. It explains the mean, median, and mode, and discusses the importance of organizing data through stem-and-leaf displays and frequency distributions for effective analysis.
ποΈ Topics Covered
The lecture begins with a review of line graphs and their use in comparing trends across categories. It then defines measures of central tendencyβmean, median, and modeβwith examples and a comparison of their advantages and disadvantages. The lecture also covers methods for organizing numerical data, including ordered arrays, stem-and-leaf displays, and tabulating numerical data into frequency distributions, with step-by-step instructions for constructing classes, computing frequencies, and cumulative frequencies.
π Lecture Summary
Statistical Representation
Line graphs are presented as the most commonly used graphical tool. They are beneficial for understanding trends in data clearly. Examples illustrate their use: a line graph showing the occurrence of causes of death due to cancer in males and females reveals that after age 40, occurrence is much greater in males. Similarly, a line graph of heart diseases shows the disease is more prominent in males.
Another example compares temperature in four cities (A, B, C, and D) . The graph shows that although the general pattern is similar, city A has the lowest temperature, followed by D, B, and C. In city C, the highest temperature is close to 30 degrees, while in city A and B it is about 25, and in city D, the highest temperature is about 28 degrees.
Measures of Central Tendency
The term central tendency refers to the middle value (sometimes a typical value) of the data. Measures of central tendency are measures of the location of the middle or the center of a distribution.
Mean
Also known as the arithmetic mean, the mean is typically what is meant by the word average. It is perhaps the most common measure of central tendency. The mean of a variable is given by (the sum of all its values)/(the number of values).
π Definition β Mean: the sum of all the results included in the sample divided by the number of observations.
π Formula: Mean = (Sum of all values) / (Number of values)
π Example: Find the mean of 58, 69, 73, 67, 76, 88, 91, and 74 (8 marks). Sum = 596. Mean = 596/8 = 74.5. Note that the mean is affected by extreme values.
Median
Another typical value is the median. To find the median of a number of values, first arrange the data in ascending or descending order, then locate the middle value. If there are an odd number of data points, the median is the middle value. If there are an even number of data points, the median is the mean of the two middle values.
π Definition β Median: the middle value of all the numbers in the sample. π Example: For the data 3, 6, 11, 14, 19, 19, 21, 24, 31 (9 values), the median is the middle value, which is 19. The median is easier to find than the mean, and unlike the mean, it is not affected by values that are unusually high or low.
Mode
The mode is the most common score in a set of scores. There may be more than one mode, or no mode at all.
π Definition β Mode: the most frequently observed value of the measurements in the sample. π Example: For the data set 2, 2, 1, 2, 0, 3, 2, 1, 1, 4, 1, 1, 1, 2, 2, 0, 3, 2, 1, the mode is 1.
A table compares the advantages and disadvantages of each measure:
- Mean: Quick and easy to calculate, but may not be representative of the whole sample.
- Median: Takes all numbers into account equally; fair to calculate; half of the sample (normally) lies above the median. Disadvantages: more tedious to calculate; can be affected by a few very large (or very small) numbers.
- Mode: Tedious to find for a large sample which is not in order.
Organizing Numerical Data
Numerical data can be organized in forms such as the Ordered Array, Stem-leaf Display, Tabulating and Graphing Numerical Data, Frequency Distributions (Tables, Histograms, Polygons) , and Cumulative Distributions (Tables, the Ogive) .
Stem and Leaf Display
A stem and leaf display (also called a stem and leaf plot) is particularly useful when the data is not too numerous. For example, 21 = 20 + 1 = (10 Γ 2) + 1. This is represented as a stem of 2 and a leaf of 1. The digit at the tenth place is taken as the stem and the digit at the units place is taken as the leaf. A stem is displayed once, and the leaf can take values from 0 to 9.
π Example: A stem and leaf display for the number of touchdown (TD) passes thrown by each of the 31 teams in the National Football League in the 2000 season (data: 32, 33, 33, 37, ...) is shown in the table. The stems are 3, 2, 1, and 0.
3|2337 represents the numbers 32, 33, 33, and 37.
2|001112223889 represents 12 data points.
One purpose of a stem and leaf display is to clarify the shape of the distribution. By looking at the stems and the shape of the plot, you can tell that most of the teams had between 10 and 29 passing TDs, with a few having more and a few having less.
Tabulating and Graphing Univariate Categorical Data
Different ways of organizing univariate categorical data include the Summary Table and Bar and Pie Charts, the Pareto Diagram. Bivariate categorical data can be organized using Contingency Tables and Side by Side Bar charts.
Tabulating Numerical Data
It is important that data is organized in a professional manner and graphical excellence is achieved in its presentation. In some cases, it is necessary to group the values of the data into classes to summarize the data properly. The process is described below.
Step 1: Sort Raw Data in Ascending Order Data: 12, 13, 17, 21, 24, 24, 26, 27, 27, 30, 32, 35, 37, 38, 41, 43, 44, 46, 53, 58
Step 2: Find Range Range = Maximum value β Minimum Value. Thus, Range = 58 - 12 = 46.
Step 3: Select Number of Classes Select the number of classes (usually between 5 and 15). In this example, let us make 5 classes.
Step 4: Compute Class Width Class width = Range / Number of classes = 46 / 5 = 9.2. Round up 9.2 to 10. You must round up, not off. If the range divided by the number of classes gives an integer value (no remainder), then you can either add one to the number of classes or add one to the class width.
Step 5: Determine Class Boundaries (limits) Pick a suitable starting point less than or equal to the minimum value. In this example, start with 10. Continue to add the class width (10) to this lower limit to get the lower limit of other classes: 10, 20, 30, 40, 50. To find the upper limit of the first class, subtract one from the lower limit of the second class. The upper limit of the first class is 20 β 1 = 19. Rest upper limits are: 29, 39, 49.
Step 6: Compute Class Midpoints Class Midpoint = (Lower limit + Upper limit) / 2. First midpoint is (10+19)/2 = 14.5. Other midpoints are: 24.5, 34.5, 44.5, 54.5.
Step 7: Compute Class Intervals First class: Lower limit is 10, Upper limit is 19. We can write the first class interval as 10 to 19, 10 β 19, or β10 but under 20β. Other class intervals are 20β29, 30β39, 40β49, and 50β59.
Important points to remember: There should be between 5 and 15 classes. The classes must be mutually exclusive, all inclusive, and continuous. The classes must be equal in width (except for possibly the first or last class).
Frequency Distribution: Count Observations & Assign to Class Intervals Looking through the data, the frequencies are: 10β19: 3, 20β29: 6, 30β39: 5, 40β49: 4, 50β59: 2. Total frequency = 20.
Relative Frequency of a class = Frequency of the class interval / Total Frequency. For the first class, relative frequency is 3/20 = 0.15. Percent Relative Frequency is obtained by multiplying the relative frequency by 100 (e.g., 15%).
Cumulative Frequency: If we add the frequency of the second interval to the frequency of the first interval, we get the cumulative frequency for the second interval. The cumulative frequencies are: 10β19: 3 20β29: 3 + 6 = 9 30β39: 3 + 6 + 5 = 14 40β49: 3 + 6 + 5 + 4 = 18 50β59: 3 + 6 + 5 + 4 + 2 = 20
Percent Cumulative Relative Frequency is calculated like cumulative frequency, but using percent relative frequency for each class interval. The percent cumulative relative frequency for the last class interval is 100%.
β Key Takeaways
A student must understand that measures of central tendencyβmean, median, and modeβdescribe the center or typical value of a dataset, each with its own strengths and weaknesses; the mean is sensitive to extreme values, while the median is not. The ability to construct and interpret stem-and-leaf displays is crucial for visualizing small datasets and understanding their shape. For larger numerical datasets, the skill of grouping data into a frequency distribution by calculating range, determining class width, and setting class boundaries is essential. Finally, understanding absolute, relative, and cumulative frequencies (including their percentage forms) is fundamental for summarizing data in tables and for later graphical representations like histograms and ogives.
π§ Quick Revision Questions
- What is the primary disadvantage of using the mean as a measure of central tendency, and how does the median address this weakness?
- For a dataset with an even number of values, such as 10, 15, 20, and 25, how is the median calculated?
- In the process of creating a frequency distribution, what is the purpose of rounding up the class width, and what should be done if the range is exactly divisible by the number of classes?
- What key information about a dataset is revealed by constructing a stem-and-leaf display that might not be immediately obvious from a simple list of numbers?
- If the total frequency for a dataset is 50 and the frequency for a class interval 30β39 is 10, what is the percent relative frequency for that class?
π Lecture 25 β Statistical Representation Measures of Central Tendency Part 2
π Overview: This lecture continues the study of statistical representation and measures of central tendency. It introduces the histogram as a graphical tool for numerical data and provides a comprehensive list of central tendency measures. The lecture also explains how to calculate key measures like mean, median, and mode using Excel functions, and demonstrates the calculation of the arithmetic mean for grouped data.
ποΈ Topics Covered
The lecture begins by reviewing the histogram as a bar graph of a frequency distribution. It then presents a comprehensive list of measures of central tendency, including arithmetic mean, median, mode, geometric mean, and others. The core of the lecture focuses on using Excel functions such as AVERAGE, AVERAGEA, MEDIAN, MODE, COUNT, and FREQUENCY. Finally, it provides a step-by-step example and Excel calculation for the arithmetic mean of grouped data.
π Lecture Summary
Graphing Numerical Data: The Histogram
A histogram is a bar graph of a frequency distribution. The widths of the bars are proportional to the classes into which the variable has been divided, and the heights of the bars are proportional to the class frequencies.
Measures of Central Tendency
Measures of central tendency can be summarized as:
- Arithmetic Mean a. For discrete data (Sample Mean, Population Mean) b. For grouped data
- Geometric Mean
- Harmonic Mean
- Weighted Mean
- Truncated Mean or Trimmed Mean
- Winsorized Mean
- Median (for grouped data, for discrete data)
- Mode (for grouped data, for discrete data)
- Midrange
- Midhinge
The main measures are Arithmetic Mean, Median, and Mode. These measures are used in different situations to understand data behavior for decision making, such as determining average, median, or mode salary in an organization before salary increases.
π‘ Why this matters: Without such summaries, it is not possible to compare large volumes of data.
AVERAGE
Returns the average (arithmetic mean) of the arguments.
π Definition β AVERAGE: The arithmetic mean of a set of numbers.
π Formula: =AVERAGE(number1,number2,...) where number1, number2 are 1 to 30 numeric arguments.
Remarks: Arguments must be numbers or references containing numbers. If an array or reference contains text, logical values, or empty cells, those values are ignored; cells with the value zero are included.
π Example: Data entered in cells A4 to A8 with formula =AVERAGE(A4:A8) returns 11.
AVERAGEA
Calculates the average (arithmetic mean) of the values in the list of arguments. Unlike AVERAGE, text and logical values such as TRUE and FALSE are included in the calculation.
π Definition β AVERAGEA: The arithmetic mean that includes text and logical values in the calculation.
π Formula: =AVERAGEA(value1,value2,...) where value1, value2 are 1 to 30 cells, ranges, or values.
Remarks: Array or reference arguments that contain text evaluate as 0. Arguments that contain TRUE evaluate as 1; FALSE evaluates as 0.
π Example: With data in A2:A6 and A7 as empty cell:
=AVERAGEA(A2:A6)returns 5.6 (includes the text "Not Available" as 0)=AVERAGEA(A2:A5,A7)returns 7 (includes the empty cell as 0, but ignores it in count)
MEDIAN
Returns the median of the given numbers. The median is the number in the middle of a set of numbers; half the numbers have values greater than the median, and half have values less.
π Definition β MEDIAN: The middle value in an ordered set of numbers.
π Formula: =MEDIAN(number1,number2,...) where number1, number2 are 1 to 30 numbers.
Remarks: If there is an even number of numbers, MEDIAN calculates the average of the two numbers in the middle. Text, logical values, or empty cells are ignored.
π Example: With numbers in cells A14 to A19 (1,2,3,4,5,6):
=MEDIAN(1,2,3,4,5)returns 3=MEDIAN(A14:A19)returns 3.5 (average of 3 and 4, the two middle values)
MODE
Returns the most frequently occurring, or repetitive, value in an array or range of data. Like MEDIAN, MODE is a location measure.
π Definition β MODE: The most frequently occurring value in a data set.
π Formula: =MODE(number1,number2,...) where number1, number2 are 1 to 30 arguments.
Remarks: If the data set contains no duplicate data points, MODE returns the #N/A error. No single measure of central tendency provides a complete picture of the data.
π Example: With data in cells A27 to A32 (values include 4 appearing most frequently), formula =MODE(A27:A32) returns 4.
COUNT Function
Counts the number of cells that contain numbers and also numbers within the list of arguments.
π Definition β COUNT: Counts the number of cells containing numbers in a range or array.
π Formula: =COUNT(value1,value2,...) where value1, value2 are 1 to 30 arguments.
Remarks: Arguments that are numbers, dates, or text representations of numbers are counted. Empty cells, logical values, text, or error values are ignored.
π Example: With data in A2:A8 (Sales, 12/8/2008, empty, 19, 22.24, TRUE, #DIV/0!):
=COUNT(A2:A8)returns 3 (counts numbers: 19, 22.24, and the date as a number)=COUNT(A5:A8)returns 2 (counts numbers in last 4 rows)=COUNT(A2:A8,2)returns 4 (adds the value 2 to the count)
FREQUENCY
Calculates how often values occur within a range of values, and then returns a vertical array of numbers.
π Definition β FREQUENCY: Counts how many values fall within specified intervals (bins).
π Formula: =FREQUENCY(data_array,bins_array) where data_array is the set of values and bins_array is the intervals.
Remarks: FREQUENCY returns an array and must be entered as an array formula. The number of elements in the returned array is one more than the number of elements in bins_array.
π Example: With scores in A2:A10 and bins in B2:B5 (70, 79, 89):
=FREQUENCY(A2:A10,B2:B5)returns: 1 score β€ 70, 2 scores in 71-79, 4 scores in 80-89, 2 scores β₯ 90
Arithmetic Mean Grouped Data
This section demonstrates calculating the arithmetic mean of grouped data where marks (classes) and frequency are given. Class marks are the class midpoints calculated as the average of lower and higher limits.
π Definition β Arithmetic Mean (Grouped Data): The weighted average using class marks as the representative value for each class.
π Formula: Mean = Ξ£(fX) / n
Where: f = frequency, X = class mark (class midpoint), n = total frequency
Steps:
- Calculate class marks: X = (lower limit + upper limit) / 2
- Multiply each frequency f by its class mark X to get fX
- Sum all fX values: Ξ£(fX)
- Sum all frequencies: n
- Mean = Ξ£(fX) / n
π Example: Given marks distribution:
| Marks | Frequency | Class Marks | fX |
|---|---|---|---|
| 20-24 | 1 | 22 | 22 |
| 25-29 | 4 | 27 | 108 |
| 30-34 | 8 | 32 | 256 |
| 35-39 | 11 | 37 | 407 |
| 40-44 | 15 | 42 | 630 |
| 45-49 | 9 | 47 | 423 |
| 50-54 | 2 | 52 | 104 |
| TOTAL | 50 | 1950 |
n = 50, Ξ£(fX) = 1950
Mean = 1950/50 = 39 Marks
Excel Calculation: Lower limits in A54:A60, higher limits in B54:B60, frequency in D54:D60. Class marks in F54:F60 using formula =(A54+B54)/2. fX in H54:H60 using formula =D54*F54. Total frequency in D61 using =SUM(D54:D60). Sum of fX in H61 using =SUM(H54:H60). Mean in H62 using =ROUND(H61/D61,0).
β Key Takeaways
The histogram is a fundamental graphical tool for visualizing frequency distributions. While there are many measures of central tendencyβincluding geometric mean, harmonic mean, weighted mean, trimmed mean, winsorized mean, median, mode, midrange, and midhingeβthe three primary measures are arithmetic mean, median, and mode, each serving different analytical purposes. Excel provides dedicated functions for these calculations, with AVERAGE, MEDIAN, and MODE being the most essential. When dealing with grouped data, the arithmetic mean is calculated by summing the products of frequencies and class marks, then dividing by the total frequency. Understanding which measure to use in different situations is critical for accurate data analysis and decision-making.
π§ Quick Revision Questions
- What is the difference between the AVERAGE and AVERAGEA Excel functions?
- How is the median calculated when a data set has an even number of values?
- What does the MODE function return if there are no duplicate values in a data set?
- What is a class mark, and how is it calculated for grouped data?
- Why must the FREQUENCY function be entered as an array formula in Excel?
π Lecture 26 β Statistical Representation Measures of Dispersion and Skewness Part 1
π Overview: This lecture covers key statistical tools for data representation and analysis, including frequency distributions, graphical methods, measures of central tendency (geometric and harmonic means), and positional measures. It introduces fundamental concepts of data dispersion and skewness, which are essential for understanding how data varies and is distributed in real-world business applications.
ποΈ Topics Covered
The lecture begins with a review of the FREQUENCY function and its application in creating frequency distributions. It then covers frequency polygons, cumulative frequency distributions, and ogives. Methods for tabulating and graphing univariate data are presented, including summary tables, bar charts, pie charts, and Pareto diagrams. The lecture introduces contingency tables and side-by-side charts, followed by geometric and harmonic means, quartiles, deciles, and percentiles. It concludes with discussions on symmetrical and skewed distributions, empirical relationships, trimmed and winsorized means, and types of dispersion measures.
π Lecture Summary
FREQUENCY-EXAMPLE
The FREQUENCY function in Excel calculates how often values occur within a range of values and returns a vertical array of numbers. The syntax is FREQUENCY(data_array, bins_array). To use it, you enter the data in one range, the bins (upper limits) in another range, select cells one more than the number of bins, type the formula, and press CTRL+Shift+Enter to enter it as an array formula.
π Definition β FREQUENCY Function: A function that counts how many values in a data set fall into specified intervals (bins).
π Formula: =FREQUENCY(data_array, bins_array) β Counts values less than or equal to each bin and values above the last bin.
π Example: Data in cells A3:A11 = {values}. Bins in B3:B5 = {70, 79, 89}. Formula = FREQUENCY(A3:A11, B3:B5). Result: β€70 = 1, 71-79 = 2, 80-88 = 4, β₯89 = 2. The Number of result cells is one more than the number of bins.
Application of FREQUENCY Function in Frequency Distribution
To create a frequency distribution using Excel: Write data in column A. In columns C and D, write class lower and upper limits. Use class upper limits as bins. Select cells E3 to E8 (one more than number of classes). Type =FREQUENCY(A2:A21, D3:D7). Press Ctrl+Shift+Enter to fill the frequency column.
FREQUENCY POLYGONS
A frequency polygon is a line graph obtained from a frequency distribution by joining with straight lines points whose x-coordinates are the midpoints of successive class intervals and whose y-coordinates are the corresponding class frequencies.
CUMULATIVE FREQUENCY
Relative frequency can be converted into cumulative frequency by adding the current frequency to the previous total. Percent cumulative frequency is calculated by dividing the cumulative frequency by the total number of observations and multiplying by 100.
π Example: First interval: relative frequency = 3, cumulative frequency = 3. Second interval (20-30): relative frequency = 6, cumulative frequency = 3 + 6 = 9 (meaning 9 values are β€ 30). Total cumulative frequency = 20. Percent cumulative frequency for first interval = 3/20 Γ 100 = 15%.
CUMULATIVE % POLYGON-OGIVE
An ogive is a cumulative percentage frequency polygon that starts from the first limit (not the midpoint as in relative frequency polygons). The maximum value in an ogive is always 100%. Ogives are used for determining cumulative frequencies at different values (not limits).
TABULATING AND GRAPHING UNIVARIATE DATA
Univariate data (data with one variable) can be tabulated in summary form or in graphical form using three types of charts: Bar Charts, Pie Charts, or Pareto Diagrams.
SUMMARY TABLE
A summary table is built specifically from detailed data. It contains summaries of the data and is used to speed up analysis.
π Example: Investor's Portfolio
| Investment Category | Amount (thousand Rs) | Percentage |
|---|---|---|
| Stocks | 46.5 | 42.27 |
| Bonds | 32 | 29.09 |
| Cash Deposit | 15.5 | 14.09 |
| Savings | 16 | 14.55 |
| TOTAL | 110 | 100 |
BAR CHART
A bar chart displays data using rectangular bars with heights proportional to the values. It can be prepared using the Excel Chart Wizard.
PIE CHARTS
Pie charts are very useful charts to show percentage distribution. They are made with the help of the Chart Wizard and clearly show how categories like stocks and bonds stand out.
PARETO DIAGRAMS
A Pareto diagram is a cumulative distribution where the first point is the first relative frequency. Each subsequent point adds the next category's percentage. The diagram shows both relative and cumulative frequency simultaneously.
π Example: First category (Stocks) = 42%. Add Bonds (29%) = 71%. Add Savings (15%) = 86%. Add CD (14%) = 100%.
CONTINGENCY TABLES
A contingency table is another form of data presentation that shows the comparison of multiple categories across different groups.
π Example: A table comparing three investors along with their combined total investment.
SIDE BY SIDE CHARTS
Side by side charts represent the same data as a contingency table using different colors to differentiate between groups.
GEOMETRIC MEAN
The geometric mean is defined as the nth root of the product of n individual values. It is used for data that grows multiplicatively.
π Definition β Geometric Mean (GM): The nth root of the product of n values.
π Formula: G = (x1 Γ x2 Γ x3 Γ ... Γ xn)^(1/n)
π Example: Find GM of 130, 140, 160. GM = (130 Γ 140 Γ 160)^(1/3) = 142.8
HARMONIC MEAN
The harmonic mean is defined as the reciprocal of the arithmetic mean of the reciprocals of the data values.
π Definition β Harmonic Mean (HM): The number of observations divided by the sum of the reciprocals of the observations.
π Formula: HM = n / (1/x1 + 1/x2 + ... + 1/xn) = n / Ξ£(1/xi)
π Example: Find HM of 10, 8, 6. HM = 3 / (1/10 + 1/8 + 1/6) = 7.66
QUARTILES
Quartiles divide data into 4 equal parts.
π Definition β Quartiles: Values that divide a data set into four equal parts.
π Formula (Ungrouped data):
- Q1 = (n+1)/4 position
- Q2 = 2(n+1)/4 position
- Q3 = 3(n+1)/4 position
π Formula (Grouped data): Qi = l + h/f[(Ξ£f/4 Γ i) β cf]
Where l = lower boundary, h = width of class interval, f = frequency, cf = cumulative frequency
DECILES
Deciles divide data into 10 equal parts.
π Definition β Deciles: Values that divide a data set into ten equal parts.
π Formula (Ungrouped data):
- D1 = (n+1)/10 position
- D2 = 2(n+1)/10 position
- D9 = 9(n+1)/10 position
π Formula (Grouped data): Di = l + h/f[(Ξ£f/10 Γ i) β cf] (for i = 1, 2, ..., 9)
PERCENTILES
Percentiles divide data into 100 equal parts.
π Definition β Percentiles: Values that divide a data set into one hundred equal parts.
π Formula (Ungrouped data):
- P1 = (n+1)/100 position
- P2 = 2(n+1)/100 position
- P99 = 99(n+1)/100 position
π Formula (Grouped data): Pi = l + h/f[(Ξ£f/100 Γ i) β cf] (for i = 1, 2, ..., 99)
Symmetrical Distribution
In a symmetrical distribution, the mean, median, and mode are all equal: mean = median = mode.
Positively Skewed Distribution
In a positively skewed distribution (tilted to the left), the mean is greater than the median, which is greater than the mode: mean > median > mode.
Negatively Skewed Distribution
In a negatively skewed distribution (tilted to the right), the mode is less than the median, which is less than the mean: mode < median < mean.
EMPIRICAL RELATIONSHIPS
For a moderately skewed and unimodal distribution, there is an empirical relationship: Mean β Mode = 3(Mean β Median).
π Example: If mode = 15 and mean = 18, find the median. Median = 1/3[mode + 2(mean)] = 1/3[15 + 2(18)] = [15 + 36]/3 = 51/3 = 17
TRIMMED MEAN
A trimmed mean (or truncated mean) is a measure of central tendency. First, sort the data. Then discard an equal number of data at both ends (most often 25 percent of the ends). The mean of the remaining data is called the trimmed mean.
π Definition β Trimmed Mean: The mean calculated after removing a specified percentage of observations from both ends of a sorted data set.
WINSORIZED MEAN
The Winsorized mean involves calculating the mean after replacing given parts of data at the high and low ends with the most extreme remaining values (most often 25 percent of the ends are replaced).
π Definition β Winsorized Mean: The mean calculated after replacing extreme values with the nearest remaining values in the data set.
π Example: Data: 9.1, 9.2, 9.3, 9.2, 9.2, 9.9 (sorted: 9.1, 9.2, 9.2, 9.2, 9.3, 9.9) Q1 position = (6+1)/4 = 1.75, Q1 β 9.2 (2nd value) Q3 position = 3(6+1)/4 = 5.25, Q3 β 9.3 (5th value) Trimmed Mean = (9.2 + 9.2 + 9.2 + 9.3)/4 = 9.225 Winsorized Mean = (9.2 + 9.2 + 9.2 + 9.2 + 9.3 + 9.3)/6 = 9.233
π‘ Why this matters: Trimmed and Winsorized means reduce the influence of outliers while keeping most of the data, providing more robust central tendency measures.
DISPERSION OF DATA
Dispersion of data is defined as the degree to which numerical data tend to spread about an average.
π Definition β Dispersion: The spread or variability of data points around a central value.
TYPES OF MEASURES OF DISPERSION
There are two types of measures of dispersion:
- Absolute measures
- Relative measures (coefficients)
DISPERSION OF DATA β Types of Absolute Measures
- Range
- Quartile Deviation
- Mean Deviation
- Standard Deviation or Variance
Types of Relative Measures
- Coefficient of Range
- Coefficient of Quartile Deviation
- Coefficient of Mean Deviation
- Coefficient of Variation
β Key Takeaways
The FREQUENCY function in Excel is essential for creating frequency distributions by counting values in specified intervals. Geometric and harmonic means provide alternative central tendency measures for multiplicative and reciprocal data, respectively. Quartiles, deciles, and percentiles divide data into 4, 10, and 100 parts and are calculated using specific formulas for both ungrouped and grouped data. Skewness describes distribution shape, with empirical relationships linking mean, median, and mode in moderately skewed unimodal distributions. Finally, dispersion measures like range, quartile deviation, mean deviation, and standard deviation quantify data spread, with relative coefficients allowing comparisons between different data sets.
π§ Quick Revision Questions
- What is the syntax of the FREQUENCY function in Excel, and why must one more cell than the number of bins be selected?
- How do you calculate the cumulative frequency and percent cumulative frequency from relative frequencies?
- What is the difference between a trimmed mean and a Winsorized mean?
- Write the formulas for Q1, D1, and P1 for both ungrouped and grouped data.
- In a negatively skewed distribution, how do the mean, median, and mode relate to each other?
π Lecture 27 β Statistical Representation Measures of Dispersion and Skewness Part 2
π Overview: This lecture covers the review of measures of central tendency and introduces key measures of dispersion including range, quartiles, and quartile deviation. It explains how to determine the shape of a distribution using the 5-number summary, box-and-whisker plots, and the relationship between median, midhinge, and midrange.
ποΈ Topics Covered
The lecture reviews measures of central tendency (mean, median, mode, geometric mean, midrange) and dispersion (range, interquartile range, variance, standard deviation, coefficient of variation). It covers quartiles and their calculation, midhinge, quartile deviation, shape of distributions (symmetrical, right-skewed, left-skewed), the 5-number summary, box-and-whisker plots, and summary measures of central tendency and variation.
π Lecture Summary
MEASURES OF CENTRAL TENDENCY, VARIATION AND SHAPE FOR A SAMPLE
There are many different measures of central tendency. These include Mean, Median, Mode, Midrange, Quartiles, and Midhinge. Measures of dispersion include Range, Interquartile Range, Variance, Standard Deviation, and Coefficient of Variation. Distributions can be Right-skewed, Left-skewed, or Symmetrical. Additional topics include Exploratory Data Analysis, Five-Number Summary, Box-and-Whisker Plot, Proper Descriptive Summarisation, Exploring Ethical Issues, and Coefficient of Correlation.
MEANS
The most common measure of central tendency is the mean. The Arithmetic Mean represents the overall average. The Median divides data into two equal parts. The Mode is the most common value. Geometric mean is used in compounding such as investments. Harmonic mean is the mean of inverse values. Each of these has its own utility.
π Definition β Mean (Arithmetic): The sum of all values divided by the number of values. π Formula: Mean = Ξ£x / n β the total of all observations divided by the sample size. π‘ Why this matters: When sample data is used to estimate population mean, the number is reduced by 1 to improve the estimate and avoid errors.
EXTREME VALUES
An important point to remember is that arithmetic mean is affected by extreme values (outliers). π Example: For values 1, 3, 5, 7, and 9, the mean is 5. For values 1, 3, 6, 7, and 14, the value 14 is an outlier, and the mean becomes 6 β an increase of 1 or about 20% due to the outlier. While preparing data for mean, it is important to spot and eliminate outlier values.
THE MEDIAN
The Median is derived after ordering the array in ascending order. If the number of observations is odd, it is the middle value; otherwise, it is the average of the two middle values. It is not affected by extreme values.
THE MODE
The Mode is the value that occurs most frequently. π Example: In data where 8 appears most frequently, the mode is 8. Mode is also not affected by extreme values. An important point about Mode is that there may not be a Mode at all (no value occurs frequently), or there may be more than one mode. The mode can be used for numerical or categorical data.
RANGE
Another measure of dispersion of data is the Range. It is the difference between the largest and smallest value.
π Definition β Range: The difference between the largest and smallest value in a dataset.
MIDRANGE
Midrange is the average of the smallest and largest value. In other words, it is half of a range. Midrange is affected by extreme values as it is based on smallest and largest values.
π Formula: Midrange = (Smallest value + Largest value) / 2
QUARTILES
Quartiles are not exclusively measures of central tendency. However, they are useful for dividing data into 4 equal parts, each containing 25% of the data. There are three quartiles: 25% of data falls below the first quartile (Q1), 50% below the second quartile (Q2/median), and 75% below the third quartile (Q3).
π Formula: Position of ith quartile = i(n + 1) / 4, where i = 1, 2, 3 and n is the number of data points.
π Example: Find first, second, and third quartile of the following data: 11, 22, 17, 16, 12, 21, 16, 13, 18 Step 1: Arrange data in ascending order: 11, 12, 13, 16, 16, 17, 18, 21, 22 (n = 9) Position of Q1 = (9+1)/4 = 2.5 Q1 = 2nd value + 0.5 Γ (3rd value β 2nd value) = 12 + 0.5 Γ (13 β 12) = 12 + 0.5 Γ 1 = 12.5 Position of Q2 = 2 Γ (9+1)/4 = 5 Q2 = 5th value = 16 Position of Q3 = 3 Γ (9+1)/4 = 7.5 Q3 = 7th value + 0.5 Γ (8th value β 7th value) = 18 + 0.5 Γ (21 β 18) = 18 + 0.5 Γ 3 = 19.5
MIDHINGE
Midhinge is the average of the first and third quartiles.
π Formula: Midhinge = (Q1 + Q3) / 2
QUARTILE DEVIATION
Quartile Deviation (Q.D.) is half the difference between the first and third quartile.
π Formula: Q.D. = (Q3 β Q1) / 2
π Example: Find Q.D. of the following data: 14, 10, 17, 5, 9, 20, 8, 24, 22, 13 n = 10 Position of Q1 = (10+1)/4 = 2.75 Q1 = 2nd value + 0.75 Γ (3rd value β 2nd value) = 8 + 0.75 Γ (9 β 8) = 8 + 0.75 = 8.75 Position of Q3 = 3 Γ (10+1)/4 = 8.25 Q3 = 8th value + 0.25 Γ (9th value β 8th value) = 20 + 0.25 Γ (22 β 20) = 20 + 0.25 Γ 2 = 20.50 Q.D. = (20.50 β 8.75) / 2 = 5.875
SHAPE OF DISTRIBUTION
The shape of a distribution can be:
- Symmetrical Distribution
- Asymmetrical Distribution: a. Right-skewed or positively-skewed distribution b. Left-skewed or negatively-skewed distribution
We can find the shape of distribution using the 5-number summary.
5-NUMBER SUMMARY
The 5-number summary consists of:
- Smallest value (Xsmallest)
- 1st Quartile (Q1)
- Median (Q2)
- 3rd Quartile (Q3)
- Largest value (Xlargest)
BOX AND WHISKER PLOT
Box and whisker plot shows the 5-number summary graphically and gives a good idea about the shape of the distribution.
Symmetrical Distribution: Data is perfectly symmetrical if:
- Distance from Q1 to Median = Distance from Median to Q3
- Distance from Xsmallest to Q1 = Distance from Q3 to Xlargest
- That is: Median = Midhinge = Midrange
Right-skewed distribution: Distance from Xlargest to Q3 greatly exceeds distance from Q1 to Xsmallest
- That is: Median < Midhinge < Midrange
Left-skewed distribution: Distance from Q1 to Xsmallest greatly exceeds distance from Xlargest to Q3
- That is: Median > Midhinge > Midrange
π Example: Suppose there are nine homes valued at Rs. 150,0000, 140,0000, 160,0000, 150,0000, 160,0000, 170,0000, 160,0000, 150,0000, and 160,0000. One small home is built with a valuation of Rs. 20,0000. Find whether the frequency polygon is negatively skewed or positively skewed. Following sample represents annual costs (in '000 Rs) for attending 10 conferences: 13.0, 14.5, 14.9, 15.2, 15.2, 15.4, 15.6, 16.2, 17, 23.1 Find 5-number summary and shape of the distribution. n = 10 Position of Q1 = (10+1)/4 = 2.75 Q1 = 14.5 + 0.75 Γ (14.9 β 14.5) = 14.5 + 0.75 Γ 0.4 = 14.8 Position of Q3 = 3 Γ (10+1)/4 = 8.25 Q3 = 16.2 + 0.25 Γ (17 β 16.2) = 16.2 + 0.25 Γ 0.8 = 16.4 Median = (15.2 + 15.4)/2 = 15.3 5-number summary: Smallest = 13.0, Q1 = 14.8, Median = 15.3, Q3 = 16.4, Largest = 23.1 Midrange = (23.1 + 13.0)/2 = 18.05 Midhinge = (14.8 + 16.4)/2 = 15.6 Since Median (15.3) < Midhinge (15.6) < Midrange (18.05), the shape is right-skewed.
SUMMARY MEASURES
The summary measures include measures of central tendency and variation. In variation, there are range, Interquartile range, standard deviation, variance, and coefficient of variation.
MEASURES OF VARIATION
In measures of variation, the sample and population standard deviation and variance are the most important measures. The coefficient of variation is the ratio of standard deviation to the mean expressed as a percentage.
INTERQUARTILE RANGE
Interquartile range is the difference between the 1st and 3rd quartile.
π Definition β Interquartile Range (IQR): The difference between Q3 and Q1, representing the middle 50% of data.
β Key Takeaways
The arithmetic mean is affected by extreme values (outliers), while median and mode are not. Quartiles divide data into four equal parts, and quartile deviation measures half the spread between Q1 and Q3. The shape of a distribution can be determined by comparing median, midhinge, and midrange: symmetrical when all three are equal, right-skewed when median < midhinge < midrange, and left-skewed when median > midhinge > midrange. The 5-number summary (smallest, Q1, median, Q3, largest) is essential for creating box-and-whisker plots and understanding data distribution. Interquartile range and quartile deviation are robust measures of dispersion not affected by outliers.
π§ Quick Revision Questions
- What is the formula for finding the position of the ith quartile in an ordered dataset?
- How does the presence of an outlier affect the arithmetic mean versus the median?
- Calculate Q1 and Q3 for the dataset: 5, 8, 12, 15, 18, 22, 25, 28, 30.
- If Median = 20, Midhinge = 22, and Midrange = 25, what is the shape of the distribution?
- What is the difference between quartile deviation and interquartile range?