A single very high number can drag an average upward while barely moving the median — and that difference is not a technicality, it is the whole reason placement figures need to be read carefully. This post explains the statistics itself, with a fictional worked example, two real published examples of the same pattern, and then shows why Shalyam asks every contributor for a median placement package rather than an average or a highest figure.
In short
- The mean (average) is the sum of every value divided by how many values there are — every single value, including one extreme outlier, pulls on it.
- The median is the middle value once every value is sorted in order — one extreme value does not move it unless it changes which value sits in the middle.
- When a dataset has a long tail of unusually high values (called right-skewed), the mean sits above the median.
- The mean is not always the wrong choice: for roughly symmetric data with no extreme outliers, the mean is the more informative statistic, because it uses every value rather than just the middle one.
- Placement packages are a textbook case of right-skewed data — which is exactly why Shalyam's own Placements field asks contributors for a
median_package, not an average or a highest figure.
What “mean” and “median” actually measure
Both are ways of describing a “typical” value in a set of numbers, but they get there differently, and that difference matters.
The mean adds up every value in the set and divides by how many values there are. Because every value is part of that sum, one unusually large (or small) value changes the result — sometimes by a lot.
The median sorts every value from smallest to largest and picks the one sitting in the middle. It does not care how far above or below the middle a value sits — only where it ranks. That is why a single extreme value, however large, usually does not move the median at all, unless it is large enough to shift which value lands in the middle position.
Why one outlier moves the average but not the median
Note: the numbers in this example are fictional and illustrative only. They do not describe any real college, any real batch of students, or any real placement data — they exist purely to show the arithmetic.
Imagine a made-up “Class of Nowhere College, 2026” — a hypothetical batch of 7 fictional graduates. Their made-up package offers, in LPA (lakhs per annum), sorted smallest to largest, are:
3.5, 4.0, 4.2, 4.5, 4.8, 5.0, 30.0
- Mean: add all seven numbers (56.0) and divide by 7 → 8.0 LPA.
- Median: sort the seven values and take the middle one (the 4th of 7) → 4.5 LPA.
Six of these seven fictional graduates were offered somewhere between 3.5 and 5.0 LPA. One received an outlier offer of 30 LPA. That single invented data point drags the mean all the way up to 8.0 LPA — higher than every one of the other six offers. The median of 4.5 LPA, by contrast, sits right where six of the seven offers actually cluster.
This shape — most values clustered together, with a small number of much larger values stretching the tail upward — is called right-skewed. In a right-skewed distribution, the mean is pulled above the median, and the size of that gap is itself a quick, informal signal of how skewed the underlying data is.
The same pattern, from two real published sources
This is not a quirk of one made-up example. The same arithmetic shows up in sources that are not about Indian college placements at all:
- MacEwan University's open introductory statistics textbook compares the data sets {3, 3, 5, 7, 7} and {3, 3, 5, 7, 799}. Both have the same median — 5 — but very different means: 5 versus 203. A single extreme value (799) moved the mean a long way and left the median exactly where it was.
- Get on Board's careers help centre gives a salary example: for salaries of 1,000, 1,100, 1,200 and 6,000, the average is 2,325 but the median is 1,150. As the article puts it, the median is “the number you would budget against.”
The honest caveat: the mean is not always the wrong tool
None of this means the median always wins. For data that is roughly symmetric — no long tail, no extreme outliers — the mean is the more informative statistic, because it uses every value in the dataset rather than discarding all the information except rank order. MacEwan's own textbook states this plainly: it is better to use the mean to describe the centre of a symmetric distribution, and better to use the median when the distribution is skewed or outliers exist. The choice depends on the shape of the data, not on one measure being universally “more correct.”
Why this matters specifically for Indian college placements
Placement data for a real graduating batch very often looks like the fictional example above: most students land offers in a fairly narrow, typical range, while a small number of students receive offers several times higher — often from one or two highly selective recruiters, or from an international offer converted into rupees. InsideIIM's own explainer on this makes both points directly: a few companies that recruit very few people can push a batch's average package very high, and B-schools that include international offers “by converting them to the Indian rupee” can push the reported average higher still.
This is also why India's own NIRF management ranking has disclosed median salary figures for participating institutions since at least its 2017 cycle, alongside the average — precisely so a reader isn't left with only the more easily inflated number.
Why Shalyam asks for a median package, not an average or a highest figure
Shalyam's Placements contribution field asks a contributor for four things about a college's placements for a given year:
| Field | What it captures |
|---|---|
| Academic year | Which year's placement cycle the figures describe |
| Report link | A link to the official report the figures come from (required once a contributor says the figure is documented) |
| Median package | The median package offered that year — not an average, not a highest figure |
| Students placed | How many students were actually placed |
That field asks for a median, specifically, because placement data is exactly the kind of right-skewed distribution this post has just walked through. There is no field on Shalyam for a “highest package” at all — a single outlier figure is not a fact worth building a page around.
What to do with this
When you read a college's placement claim, ask which number you're looking at. A “highest package” headline tells you almost nothing about what a typical graduate earned. An average can be pulled a long way by a handful of outlier offers. A median, sourced from an official report, is the figure closest to what most students in that batch actually received — which is exactly what Shalyam asks contributors to share.
Browse colleges on Shalyam and add a median placement figure, with the official report it came from, for a college you know.
Comments
Log in to like or comment on this post.