Exploring the Relationship Between Mean, Mode, and Median

When researchers summarise data, three values usually do the heavy lifting: the mean, the median, and the mode. Each one points to the “centre” of a dataset, but they rarely point to the exact same spot. Understanding how they connect (and when they part ways) is the difference between a tidy statistical summary and a misleading one.

Quick recap: what each measure actually tells us 🔗

Before we look at how these three values relate, a short refresher helps. The mean is the arithmetic average, calculated by adding every value and dividing by the number of observations. The median is the middle value when the data is arranged in order. The mode is the value that appears most often. According to a widely cited review in the Annals of Cardiac Anaesthesia, the mean is the most commonly used measure because it incorporates every data point, but it is highly sensitive to extreme values and therefore unsuitable for skewed distributions.

That sensitivity is the reason the three measures behave differently. The mean listens to every observation, the median only cares about position, and the mode only notices frequency. Feed them the same data and they often agree. Feed them lopsided data, and they start to disagree in predictable ways.

The symmetric case: when all three agree 🔗

In a perfectly symmetrical distribution, something elegant happens. Draw a vertical line down the middle of the histogram and the two halves mirror each other. In such a distribution, the mean, median, and mode all coincide at the same central value. The classic bell-shaped normal distribution is the poster child for this behaviour.

This alignment is more than a geometric coincidence. It tells the researcher that the data clusters evenly around one value, with no unusual pull in either direction. Human height within a single age group, measurement errors in a well-calibrated instrument, and scores on a well-designed standardised test often approximate this pattern.

A small caveat about symmetry 🔗

Not every symmetrical distribution has a single peak. If a distribution is symmetric but bimodal (two peaks of equal height), the two modes will sit on either side of the centre, while the mean and median still coincide at the middle. So while symmetry guarantees that mean equals median, the mode can behave differently when the data has multiple peaks.

Skewed distributions: where the three measures part ways 🔗

Real-world data rarely lines up so politely. Income, property prices, exam scores, response times, and rainfall are almost always skewed. Skewness simply means the distribution has one tail longer than the other. This stretched tail pulls the mean toward it, leaves the median only mildly affected, and barely disturbs the mode.

Positively (right) skewed distributions 🔗

A distribution is positively skewed when the tail extends toward higher values. Think of household income data, where a small number of very high earners stretch the right side of the curve. In this situation, the ordering is Mean > Median > Mode. The mean is the largest because those extreme high values drag it upward; the mode stays near the densest cluster of typical values on the left; and the median sits somewhere in between.

This is why economists almost always report median income rather than mean income when describing typical households. A handful of billionaires would push the mean far above what most citizens actually earn, while the median better reflects the experience of a person in the middle of the income ladder.

Negatively (left) skewed distributions 🔗

A negatively skewed distribution has its tail stretching toward the lower values. The relationship reverses: the mean is less than the median, which is less than the mode. Consider the age at retirement in a large company, where most employees retire around 60 but a few leave much earlier. The long left tail drags the mean down, while the mode stays near the typical retirement age.

A useful visual rule: the mean always chases the tail. Wherever the distribution is stretched, the mean drifts in that direction. The median resists this pull, and the mode resists it even more strongly.

Karl Pearson’s empirical formula 🔗

For distributions that are only moderately skewed, the British statistician Karl Pearson observed a striking pattern. The gap between the mean and the mode tends to be about three times the gap between the mean and the median. This observation gave us one of the most famous shortcuts in descriptive statistics:

Mode ≈ 3 × Median − 2 × Mean

Or equivalently, Mean − Mode ≈ 3 (Mean − Median). This is known as the empirical relationship because, as Pearson’s formula is not derived from a universal mathematical proof. It is based on observed patterns across many moderately skewed real-world distributions.

A quick worked example 🔗

Suppose an HR report states that the median monthly salary is ₹45,000 and the mean monthly salary is ₹52,000. Applying the empirical relation:

Mode = 3 × 45,000 − 2 × 52,000 = 1,35,000 − 1,04,000 = ₹31,000

So the most common (modal) salary is roughly ₹31,000. The fact that the mean is well above both the median and the mode tells us the salary distribution is positively skewed, probably because a few senior executives pull the average upward.

Limits of the empirical formula 🔗

The formula works well for moderate skewness, but it is not a law of nature. As a Journal of Statistics Education article points out, the textbook rule linking skewness to the order of mean, median, and mode can fail, especially in discrete distributions where the areas to the left and right of the median are not equal. There are even well-known continuous distributions (like certain Weibull densities) where the skew is positive but the mean sits to the left of the median. The empirical formula is a handy approximation, not a guarantee.

Why this relationship matters in research 🔗

Understanding how the mean, median, and mode relate is not an academic exercise. The three measures tell a researcher three different stories, and the gap between them is itself a piece of evidence about the data’s shape.

Detecting skewness without a graph 🔗

If you know all three values, you can often infer the shape of the distribution without plotting it. A large gap between the mean and median signals skewness and warns you that the mean alone may be misleading. The direction of the gap tells you which way the data is skewed.

Choosing the right measure to report 🔗

The relationship also guides a practical decision: which measure should represent the “typical” value in a report? The median is usually preferred when the distribution is skewed, because it is not strongly influenced by extreme values. The mean is best for symmetric, outlier-free continuous data. The mode comes into its own for categorical or nominal data, such as the most common religion in a census, the most frequent complaint at a grievance cell, or the most popular scheme among beneficiaries.

Public administrators and policy researchers face this choice constantly. Reporting mean land-holding size in rural districts, mean healthcare expenditure, or mean travel time to work can easily misrepresent the typical experience when the underlying data is skewed. The median is often the more honest summary.

Reading between the numbers 🔗

When a report presents only the mean, a careful reader should ask whether the median would tell a different story. A 2022 review in orthopaedic research argued that there is often a lack of understanding about when to use each metric and how various factors can impact these values. Presenting the mean alongside the median, and noting the mode where relevant, gives readers a much richer picture of the data.

A practical checklist for researchers 🔗

When working with a new dataset, it helps to follow a simple sequence: first check whether the data is approximately symmetric or clearly skewed; second compare the mean and median (a big difference suggests skew); third consider whether the mode adds useful information, especially for categorical variables; and finally choose the measure (or combination of measures) that best reflects the research question. The empirical formula can be used as a sanity check when only two of the three values are known.

Ultimately, the mean, median, and mode are not rivals. They are three lenses focused on the same data, and the relationship among them is a piece of information in its own right. A researcher who understands that relationship can describe a distribution more accurately, avoid common pitfalls, and communicate findings with far greater honesty.

What do you think? When you come across statistics about salaries, exam scores, or government scheme outcomes in news reports, do you pause to consider whether the mean being quoted actually represents a typical person? And in your own research work, have you ever found the gap between the mean and the median telling a more revealing story than either number could alone?

References

1. https://pmc.ncbi.nlm.nih.gov/articles/PMC3127352/
2. https://openstax.org/books/introductory-business-statistics-2e/pages/2-6-skewness-and-the-mean-median-and-mode
3. https://www.cuemath.com/data/relation-between-mean-median-and-mode/
4. https://byjus.com/maths/relation-between-mean-median-and-mode/
5. https://www.vedantu.com/maths/relation-between-mean-median-and-mode
6. https://www.tandfonline.com/doi/full/10.1080/10691898.2005.11910556
7. https://statistics.laerd.com/statistical-guides/measures-central-tendency-mean-mode-median.php
8. https://www.researchgate.net/publication/365036906_Choosing_the_appropriate_measure_of_central_tendency_mean_median_or_mode