π Lecture 40 β Measures of Central Tendency
π Overview: This lecture introduces the fundamental concepts of central tendency in statistics, focusing on various types of averages used to describe where data clusters. Understanding these measures is essential for summarizing data sets and making meaningful comparisons, whether in academics, business, or research.
ποΈ Topics Covered
The lecture covers six measures of central tendency: Arithmetic Mean, Geometric Mean, Weighted Mean, Harmonic Mean, Median, and Mode. It explains their definitions, formulas, applications, and how to calculate them for both raw and grouped data, with practical examples demonstrating when each measure is most appropriate.
π Lecture Summary
Focus of Statistics
Statistics is concerned with three main aspects of data: Measures of Location (where data is clustered or centered, also called measures of central tendency), Measures of Dispersion (how data spreads from the center), and Measures of Shape (how data sways and peaks). The central tendency is measured by averages, which describe the point about which various observed values cluster. In mathematics, an average or central tendency of a data set refers to a measure of the "middle" or "expected" value of the data set. The measures studied include: Arithmetic Mean, Geometric Mean, Weighted Mean, Harmonic Mean, Median, and Mode.
Arithmetic Mean
The Arithmetic Mean is the sum of the observation values divided by the number of observations. It is the most widely used measure of central tendency and is the only measure where the sum of the deviations of each value from the mean is always zero. The formula for calculating the arithmetic mean of n values is: ( \bar{x} = \frac{x_1 + x_2 + x_3 + ... + x_n}{n} ) or ( \bar{x} = \frac{\sum_{i=1}^{n} x_i}{n} ).
For grouped data, if ( z_1, z_2, z_3, ..., z_k ) are mid-values and ( f_1, f_2, f_3, ..., f_k ) are corresponding frequencies (where k stands for number of classes), then the mean is ( \bar{z} = \frac{\sum f_i z_i}{\sum f_i} ).
π Example: The math exam scores of 9 students are: 59, 66, 65, 74, 71, 67, 71, 62, 77. The mean score is: ( \bar{x} = \frac{59+66+65+74+71+67+71+62+77}{9} = \frac{612}{9} = 68 ).
Geometric Mean
The Geometric Mean is often used when calculating investment returns over multiple periods or to find a compound growth rate. It is computed by taking the nth root of the product of n values. When dealing with numbers that need to be multiplied, the geometric mean gives a more realistic picture than the arithmetic mean.
π Definition β Geometric Mean: For n values, ( GM = \sqrt[n]{a_1 a_2 ... a_n} = \left( \prod_{i=1}^{n} a_i \right)^{\frac{1}{n}} ).
For grouped data, if "n" non-zero positive variable values ( x_1, x_2, ..., x_n ) occur ( f_1, f_2, ..., f_n ) times respectively, then ( GM = \left[ x_1^{f_1} x_2^{f_2} ... x_n^{f_n} \right]^{\frac{1}{N}} = \left( \prod_{i=1}^{n} x_i^{f_i} \right)^{\frac{1}{N}} ) where ( N = \sum f_i ).
π Example (a): Company A has grown over 3 years by 10, 12, and 14 million dollars. Average annual growth = arithmetic mean = (10+12+14)/3 = 12 million dollars.
π Example (b): Company B's profit has grown over 3 years by 2.5%, 3%, and 3.5%. Average growth rate = geometric mean = ( [(1.025)(1.030)(1.035)]^{\frac{1}{3}} - 1 = 2.9992% ). π‘ Why this matters: When dealing with percentages and growth rates, the geometric mean is more accurate than the arithmetic mean because it accounts for compounding.
Weighted Mean
The Weighted Mean is a special case of the mean that allows different weights on different observations. Formally, for a non-empty set of data ( x_1, x_2, ..., x_n ) with non-negative weights ( w_1, w_2, ..., w_n ), the weighted mean is: ( \bar{x}_w = \frac{w_1 x_1 + w_2 x_2 + ... + w_n x_n}{w_1 + w_2 + ... + w_n} ). Note: The arithmetic mean is a weighted mean where all weights equal 1.
π Example: Two school classes: Morning class (20 students, average 80) and Afternoon class (30 students, average 90). The straight average of 80 and 90 is 85, but this does not account for class sizes. The weighted mean = ( \frac{20(80) + 30(90)}{20+30} = \frac{1600+2700}{50} = \frac{4300}{50} = 86 ). Alternatively, averaging all 50 individual scores gives the same result of 86.
Harmonic Mean
The Harmonic Mean is often used by investors to find the average cost of shares purchased over time. In situations involving rates and ratios, the harmonic mean provides the truest average.
π Definition β Harmonic Mean: For a set of positive data values ( x_1, x_2, ..., x_n ), ( HM = \frac{n}{\frac{1}{x_1} + \frac{1}{x_2} + ... + \frac{1}{x_n}} = \frac{n}{\sum_{i=1}^{n} \frac{1}{x_i}} ).
π Example (c): 100 km trip, traveling half the time at 40 km/h and half at 60 km/h. Average speed = arithmetic mean = (40+60)/2 = 50 km/hr.
π Example (d): 100 km trip, traveling half the distance at 40 km/h and remaining half at 60 km/h. Average speed = harmonic mean = ( \frac{2}{1/40 + 1/60} = \frac{2}{0.025 + 0.0167} = \frac{2}{0.0417} = 48 ) km/hr. π‘ Why this matters: When the distances are fixed but speeds vary, the harmonic mean correctly calculates the average speed; the arithmetic mean would overestimate it.
Median
The Median is the middle value of observations such that the number of observations above it equals the number below it. If n is odd, the median is the middle value: ( M_e = x_{\frac{1}{2}(n+1)} ). If n is even, the median is the average of the middle two values: ( M_e = \frac{1}{2} \left( x_{\frac{n}{2}} + x_{\frac{n}{2} + 1} \right) ).
For grouped data: ( M_e = L_0 + \frac{h}{f_0} \left( \frac{n}{2} - F \right) ), where:
- ( L_0 ) = Lower class boundary of the median class
- h = Width of the median class
- ( f_0 ) = Frequency of the median class
- F = Cumulative frequency of the pre-median class
Steps to find median of grouped data: (1) Compute less-than-type cumulative frequencies. (2) Determine n/2. (3) Locate the median class where cumulative frequency exceeds N/2. (4) Determine lower limit of median class (( L_0 )). (5) Sum frequencies of all classes prior to median class (F). (6) Determine frequency of median class (( f_0 )). (7) Determine class width (h). (8) Apply formula.
Mode
Mode is the value of a distribution for which frequency is maximum. It is the value of a variable that occurs with the highest frequency.
π Example: Mode of list (1, 2, 2, 3, 3, 3, 4) is 3. Mode is not necessarily well-defined. List (1, 2, 2, 3, 3, 5) has two modes: 2 and 3.
For grouped data: ( M_0 = L_1 + \frac{\Delta_1}{\Delta_1 + \Delta_2} h ), where:
- ( L_1 ) = Lower boundary of modal class
- ( \Delta_1 ) = Difference between frequency of modal class and class before it
- ( \Delta_2 ) = Difference between frequency of modal class and class after it
- h = Class interval
π Example (Commuter Routes): Route V and W journey times (minutes) recorded over 10 days:
- Route V: 53, 52, 48, 51, 49, 47, 42, 48, 57, 53. Mean = 500/10 = 50. Sorted: 42, 47, 48, 48, 49, 51, 52, 53, 53, 57. Median = (49+51)/2 = 50.
- Route W: 43, 41, 39, 108, 52, 42, 38, 45, 39, 51. Mean = 498/10 = 49.8. Sorted: 38, 39, 39, 41, 42, 43, 45, 51, 52, 108. Median = (42+43)/2 = 42.5. Route W is quicker (lower median time).
π Example (Data Comparison): Consider sets A and B:
- A: 48, 52, 60, 60, 60, 68, 72
- B: 0, 10, 60, 60, 60, 110, 120 Both sets have mean = median = mode = 60. However, the sets are not the same; B is much more spread out than A. This demonstrates that measures of central tendency alone are insufficient for fully describing data; measures of dispersion are also needed.
β Key Takeaways
Students must remember that measures of central tendency (mean, median, mode, geometric mean, harmonic mean, and weighted mean) each serve different purposes and are appropriate for different data types. The arithmetic mean is best for additive data but is sensitive to outliers, while the median is robust to extreme values. The geometric mean is essential for growth rates and percentages, and the harmonic mean is correct for rates with fixed distances or quantities. The weighted mean accounts for varying importance of data points. Importantly, having identical measures of central tendency does not mean two data sets are the sameβmeasures of dispersion are also needed to fully describe data.
π§ Quick Revision Questions
- For a data set with values 10, 20, 30, 40, and an extreme outlier of 1000, which measure of central tendency would be least affected by the outlier?
- When would you use the geometric mean instead of the arithmetic mean, and give one real-world example?
- If you travel 50 km at 30 km/h and 50 km at 70 km/h, what is the correct average speedβarithmetic mean or harmonic meanβand why?
- How do you calculate the median for grouped data when the total number of observations N is even?
- In the commuter routes example, why did the median provide a better comparison than the mean for Route W?