Measures of Central Tendency

The full picture

Imagine your teacher records the marks of 40 students in a unit test. Listing all 40 numbers tells you nothing at a glance. What you need is a single value that stands in for the whole group — a 'central' value around which all others cluster. This is what a measure of central tendency does. The three measures you study in this chapter are the arithmetic mean, the median, and the mode. Each captures 'typical' in a slightly different way, and knowing when to use which one is as important as knowing how to calculate each.

The arithmetic mean is what most people call 'the average'. You add all the values and divide by how many there are. If Riya's scores in five unit tests are 68, 72, 75, 65, and 80, her mean score is (68 + 72 + 75 + 65 + 80) ÷ 5 = 360 ÷ 5 = 72. Simple and straightforward. But the mean has one important weakness: it is pulled by extreme values called outliers. If Riya scores 10 in one test because she was ill, her mean drops sharply even though it was an unusual event, not a true reflection of her ability. The Kerala SCERT syllabus also covers the weighted arithmetic mean, used when items carry unequal importance — for example, a school with three different class sizes cannot treat each class's average equally when computing the school's overall average.

The median is the middle value of a dataset arranged in order from smallest to largest. Sort Riya's five scores: 65, 68, 72, 75, 80. The middle value (3rd out of 5) is 72 — that is the median. When you have an even number of values, take the average of the two middle ones. The key strength of the median is that it ignores how extreme the lowest or highest value is. Whether the lowest score is 10 or 65, the median stays the same as long as the ordering of the middle values does not change. This makes the median the preferred measure whenever your data has outliers or is not spread symmetrically. In Kerala's landholding surveys, for example, the median farm size gives a fairer picture of what a 'typical' farmer owns than the mean, which is pulled upward by a few large plantations.

The mode is the value that appears most often in a dataset. In the scores 65, 68, 70, 70, 72, 75, the mode is 70 because it occurs twice — more than any other value. The mode is unique: it is the only central tendency measure you can use with non-numeric (categorical) data. If you survey students at a Thrissur college about their preferred snack and 35 choose banana chips, 28 choose tapioca chips, and 17 choose biscuits, the mode is banana chips — no arithmetic needed. A dataset can have no mode (when all values appear equally often), one mode (unimodal), or two modes (bimodal), or even more. Because the mode only uses frequency and ignores the actual value of most data points, it is the least mathematically powerful of the three measures.

Choosing the right measure is a skill in itself. Use the mean when your data is fairly symmetric and has no extreme outliers — it makes full use of every data point and is essential for further statistical work like calculating standard deviation. Use the median when your data is skewed (pulled heavily to one side) or contains outliers — income data, house prices, and land holdings almost always call for the median. Use the mode when you want to know the most popular or most common item, especially with categories. For your board exam, practise problems with grouped frequency distributions: you will estimate the mean using class midpoints, find the median using the formula involving the median class, and identify the modal class as the class with the highest frequency.

An Indian example

Consider the monthly incomes in a small neighbourhood in Kozhikode: five families earn ₹18,000, ₹22,000, ₹25,000, ₹20,000, and ₹19,000 — and one family owns a supermarket and earns ₹4,50,000. The mean income is (18,000 + 22,000 + 25,000 + 20,000 + 19,000 + 4,50,000) ÷ 6 = 5,54,000 ÷ 6 = ₹92,333. That sounds like the neighbourhood is doing well — but five of the six families earn less than ₹25,000. The median tells the real story: sort the incomes as ₹18,000, ₹19,000, ₹20,000, ₹22,000, ₹25,000, ₹4,50,000 — the average of the 3rd and 4th values = (₹20,000 + ₹22,000) ÷ 2 = ₹21,000. The median of ₹21,000 accurately reflects what a typical household in this neighbourhood earns. No income value repeats in this dataset, so there is no mode. This is exactly why the government uses median household income — not mean — when designing welfare programmes for Below Poverty Line families.

Common misconceptions to watch for

• Wrong belief: 'The mean is always the best and most accurate measure of central tendency.' Correction: The mean is the most common measure but not always the best. When a dataset has outliers — like one very high salary among many modest ones — the mean gets pulled toward that extreme and misrepresents the typical value. In such cases, the median is more reliable because it is not affected by how large or small the extreme values are.
• Wrong belief: 'Every dataset must have exactly one mode.' Correction: A dataset can have no mode at all (when every value appears the same number of times), exactly one mode (unimodal), two modes (bimodal), or three or more modes. There is no rule that forces one mode to exist. If shoe sizes 6 and 7 both appear five times each and no other size appears as often, the dataset is bimodal — both 6 and 7 are modes.
• Wrong belief: 'Arranging data in ascending order will make the mean and median equal.' Correction: Sorting changes nothing about the mean — the mean is always total ÷ count, regardless of order. Sorting only helps you locate the median's position. The mean and median are equal only when the data is perfectly symmetric (e.g., 10, 20, 30, 40, 50 — both equal 30). As soon as one extreme value is added, they diverge.