Histogram shapes / Left-skewed
Left-Skewed Histogram (Negative Skew)
A left-skewed histogram peaks on the right with a long tail to the left. See a worked example, how to spot negative skew, the mean vs median rule, and how left skew differs from right skew.
What a left-skewed histogram looks like
A left-skewed histogram (also called negatively skewed) is the mirror image of a right-skewed one. The tall bars sit on the right, and a long tail trails off to the left where a few small values live.
The example above is exam scores. Most students score high, clustered near the top, while a small number of low scores stretch the tail toward the left.
The name follows the tail, not the peak. That trips people up constantly. A left-skewed graph has its peak on the right side, and we still call it left-skewed because the thin tail points left. If you ever blank on which is which, say the tail direction out loud. That word is the skew.
How to tell a histogram is left-skewed
Two checks settle it in seconds.
- Find the tail. Look for the long, low stretch of bars fading out on one side. If the fade is on the left, toward the smaller values, the shape is left-skewed.
- Compare the mean and the median. In left-skewed data the mean prints lower than the median. If the stats panel shows a mean noticeably below the median, the data leans left before you study a single bar.
Both checks should agree. When they disagree, the skew is mild and the shape is closer to symmetric than to either lean.
Mean, median, and mode
With the tail on the left, the small outliers pull the mean down, so the order flips:
mean < median < mode
The mode is under the tall right-side bars, and the mean is dragged left by the low scores. When your mean prints lower than your median, the data leans left.
That gap between the mean and the median is the fingerprint of negative skew. The wider the gap, the heavier the left tail is pulling.
Left-skewed vs right-skewed
The two shapes are exact mirrors, and everything reverses between them.
| Left-skewed (negative) | Right-skewed (positive) | |
|---|---|---|
| Tail points | Left, toward small values | Right, toward large values |
| Peak sits | On the right | On the left |
| Average order | mean < median < mode | mode < median < mean |
| Caused by | A ceiling the data bunches against | A floor at zero with no ceiling |
| Typical data | Exam scores, retirement age, ratings | Income, house prices, wait times |
Side by side, the pair is easier to learn than either one alone. The full comparison, with both charts and a memory trick, is in the guide on right-skewed vs left-skewed histograms. The mirror page for the other direction is the right-skewed histogram.
Which average should you report
When a histogram is left-skewed, the mean stops describing a typical value because the low tail keeps tugging it downward. The median ignores those few extreme values and stays near the bulk of the data, so it is the honest summary.
Report the median as the typical value, and mention the mean only if the gap between the two is itself the point you want to make. On an easy exam, for instance, the median score tells you how the class actually did, while a mean sitting several points below it tells you a handful of students struggled badly.
Where you see it
Test scores on an easy exam, age at retirement, and product ratings out of five all tend to skew left. There is a ceiling the data bunches against (a perfect score, a maximum rating) and only a few values fall well below it.
That ceiling is the whole mechanism. Values cannot pass it, but they can fall a long way beneath it, so the data can only stretch in one direction. Age at death in a developed country works the same way: most people cluster in a band near the top of the human range, and the tail runs down toward the few who die young.
Compare that with a bell-shaped histogram, where there is no ceiling pressing the data against one side, so both tails come out about even.
Drop your numbers into the histogram maker and check the stats panel. A mean below the median is the fingerprint of left skew. If the shape is hard to read, the bin count may be hiding it, and a frequency distribution table will show the same lean in numbers.
Frequently asked questions
- Is a left-skewed histogram positive or negative skew?
- Negative. Left skew and negative skew are the same thing: the tail points to the left, toward the smaller values, and the mean sits below the median.
- What does a left-skewed graph look like?
- The tall bars sit on the right side and a long, low tail stretches off to the left. The peak is on the opposite side from the tail, which is why a left-skewed graph looks like it is leaning right.
- Why is the mean less than the median in left-skewed data?
- The long left tail holds a few unusually small values. The mean uses every value, so those low numbers drag it down, while the median only cares about the middle position and barely moves.
- What is the difference between left-skewed and right-skewed?
- They are mirror images. Left-skewed data has its tail on the left, peaks on the right, and runs mean < median < mode. Right-skewed data has its tail on the right, peaks on the left, and runs mode < median < mean.
- What kind of data is usually left-skewed?
- Data with a natural ceiling that most values bunch against, like scores on an easy test, retirement ages, or five-star ratings where most reviews are 4s and 5s.
- Is a left-skewed distribution the same as a negatively skewed distribution?
- Yes. The two names describe the same shape. Statisticians usually say negatively skewed because the skewness statistic comes out below zero; everyone else says left-skewed because that is where the tail goes.