Calculating the Median: A Robust Measure of Central Tendency

When studying a dataset, the average is often the first thing people reach for. But averages can lie-especially when a handful of extreme values quietly pull them off-centre. This is where the median steps in. As a positional measure that represents the exact middle of an ordered dataset, the median offers a more honest picture of the “typical” value, particularly when data is uneven or skewed. For researchers, policymakers, and students working with real-world information like incomes, exam scores, or household sizes, understanding the median is not just a statistical skill-it is a safeguard against misleading conclusions.

What is the median?
Why position matters more than magnitude
Calculating the median for ungrouped data
Step-by-step procedure
Calculating the median for grouped data
The median formula for grouped data
Steps to compute the median of grouped data
A worked example
Why the median is a robust measure
Dealing with skewed distributions
Resistance to outliers
When to use the median
Skewed or asymmetric data
Ordinal data
Presence of outliers or suspected errors
Limitations of the median
The median in public administration and research

What is the median?

The median is the middle value of a dataset that has been arranged in ascending or descending order. It divides the distribution into two equal halves: 50% of the observations lie below it, and 50% lie above. As a measure of central tendency, the median is simply the middle number in an ordered data set, offering a single value that summarises where the centre of the data lies.

Unlike the arithmetic mean, the median does not depend on the magnitude of every observation. It depends only on the position of values once they are sorted. This property makes it especially useful when a dataset contains unusually high or low figures that might distort the mean.

Why position matters more than magnitude

Consider a simple dataset of monthly salaries in a small firm: ₹18,000, ₹20,000, ₹22,000, ₹25,000, and ₹3,00,000. The mean works out to ₹77,000-a figure that does not reflect what most employees actually earn. The median, however, is ₹22,000, which honestly represents the typical salary. This illustrates why the median is often called a resistant or robust measure: only the middle observation was used to determine its value, which means extreme values cannot pull it off-centre.

Calculating the median for ungrouped data

Ungrouped data refers to raw observations that have not been organised into class intervals. Calculating the median here is straightforward and involves two cases depending on whether the number of observations is odd or even.

Step-by-step procedure

The process begins with arranging the data. The value of the middle-most observation obtained after arranging the data in ascending or descending order is called the median of the data.

For an odd number of observations, the median is the value at the (n+1)/2 position, where n is the total number of observations. For example, in the dataset 3, 7, 9, 12, 15, there are 5 values. The median is the (5+1)/2 = 3rd value, which is 9.

For an even number of observations, the median is the average of the two middle values. If the number of observations is even, the median is the mean value of the observations located in the n/2 and (n+2)/2 positions. For example, in the dataset 4, 6, 8, 10, 12, 14, the median is the average of the 3rd and 4th values, i.e., (8 + 10) / 2 = 9.

Calculating the median for grouped data

When data is organised into class intervals-common in survey research, census reports, and policy studies-individual values are no longer visible. Instead, we see how many observations fall within each range. Here, the median cannot be read off directly; it must be estimated using a formula.

The median formula for grouped data

The standard formula used is:

Median = l + [(n/2 − cf) / f] × h

Where:

l is the lower boundary of the median class, n is the total number of observations, cf is the cumulative frequency of the class preceding the median class, f is the frequency of the median class, and h is the size (width) of the class interval. This formula is outlined in standard references on grouped data analysis, including the method used to locate the class whose cumulative frequency is greater than or nearest to n/2, called the median class.

Steps to compute the median of grouped data

The calculation follows a logical sequence. First, construct a frequency distribution table with class intervals and their frequencies. Second, compute the cumulative frequency by adding each frequency to the sum of the previous ones. Third, find n/2, which marks the position of the middle observation. Fourth, identify the median class-the class interval where the cumulative frequency first equals or exceeds n/2. Finally, substitute the values into the formula.

A worked example

Suppose a survey collects data on the daily study hours of 50 students, organised as follows: 0-2 hours (frequency 6), 2-4 hours (frequency 14), 4-6 hours (frequency 18), 6-8 hours (frequency 9), and 8-10 hours (frequency 3). The cumulative frequencies are 6, 20, 38, 47, and 50 respectively. Since n = 50, n/2 = 25. The cumulative frequency just greater than 25 is 38, which corresponds to the class 4-6. This is the median class.

Substituting into the formula with l = 4, cf = 20, f = 18, and h = 2:

Median = 4 + [(25 − 20) / 18] × 2 = 4 + (5/18) × 2 ≈ 4.56 hours.

This tells us that roughly half the students study fewer than 4.56 hours a day, and the other half study more. For grouped data, we cannot find the exact mean, median and mode-we can only give estimates, but this estimate is typically very close to the true centre of the distribution.

Why the median is a robust measure

The word “robust” in statistics means that a measure holds its ground even when the data misbehaves. The median earns this label because it is unaffected by outliers and skewness. The median is less affected by outliers and skewed data, which is precisely why analysts prefer it for many real-world applications.

Dealing with skewed distributions

A skewed distribution is one where data clusters more heavily on one side, producing a long tail in the opposite direction. Income distributions are the classic case-most people earn modest amounts, while a few earn very high incomes. The more skewed the distribution, the greater the difference between the median and mean, and the more important it becomes to rely on the median for a faithful summary.

This is why government agencies and research organisations routinely report median income rather than average income. The Ministry of Statistics and Programme Implementation, for example, uses medians in its Household Consumption Expenditure Survey to capture the consumption patterns of typical Indian households, where a handful of very wealthy households could otherwise distort the picture.

Resistance to outliers

An outlier is a value that sits far from the rest of the data. Because the median is a positional statistic, its value hardly changes even if the smallest or largest values are replaced with far more extreme numbers. Since the median only finds the middle value of a dataset, it isn’t affected by extremely small or large values on either end of a dataset. This is a crucial advantage in fields such as public health, education research, and economics, where data errors, anomalies, or genuine extreme observations are common.

When to use the median

The median is not always the best choice. It is most valuable in specific research contexts.

Skewed or asymmetric data

Whenever a distribution is lopsided-incomes, house prices, land holdings, waiting times, or disease prevalence-the median provides a more representative centre. Income is the classic example of when to use the median instead of the mean because its distribution tends to be skewed.

Ordinal data

For data that can be ordered but not meaningfully averaged-such as satisfaction ratings on a five-point scale or stages of cancer-the median is appropriate while the mean is not. It reflects the middle rank without assuming equal intervals between categories.

Presence of outliers or suspected errors

When a dataset may contain unusual values, reporting the median alongside the mean provides a sanity check. A large gap between the two suggests skewness or outliers, prompting the researcher to investigate further.

Limitations of the median

No statistic is perfect, and the median has genuine limitations that every researcher should recognise.

First, it ignores most of the data. The median does not consider the information from extremely small or large values in a dataset. While this resistance is a strength for skewed data, it means information is discarded.

Second, the median cannot be used to compute totals. Knowing the mean salary and the number of employees lets you estimate the total payroll. Knowing the median does not. This limits its usefulness in contexts that require aggregation, such as calculating national income or tax revenue.

Third, for grouped data, the median is only an estimate based on the assumption that values are evenly distributed within the median class. This assumption may not always hold.

The median in public administration and research

Public administration is a field full of skewed data-income, landholding, household size, access to services, and budget allocations often follow asymmetric patterns. In such contexts, the median becomes a vital analytical tool. The Periodic Labour Force Survey published by the National Statistical Office relies on measures like the median to describe earnings and employment patterns that would otherwise be distorted by a few extremely high or low values.

International development agencies follow similar practice. The Human Development Index framework maintained by the United Nations Development Programme draws on median-based indicators when comparing living standards across countries, recognising that mean-based measures can mask inequality.

For students of research methodology, mastering the median is therefore not just an academic exercise. It is a practical competence for interpreting surveys, reading policy documents, and drawing defensible conclusions from messy, real-world data.

What do you think? When you read a news report that cites an “average” figure-whether it is income, exam scores, or house prices-do you pause to ask whether the median would tell a different story? And in your own research or reading, can you recall a case where using the median instead of the mean would have completely changed the conclusion?

References

PDF 📄

Table of Contents

What is the median? 🔗

Why position matters more than magnitude 🔗

Calculating the median for ungrouped data 🔗

Step-by-step procedure 🔗

Calculating the median for grouped data 🔗

The median formula for grouped data 🔗

Steps to compute the median of grouped data 🔗

A worked example 🔗

Why the median is a robust measure 🔗

Dealing with skewed distributions 🔗

Resistance to outliers 🔗

When to use the median 🔗

Skewed or asymmetric data 🔗

Ordinal data 🔗

Presence of outliers or suspected errors 🔗

Limitations of the median 🔗

The median in public administration and research 🔗