A stem and leaf plot splits every number in a dataset into two parts: a “stem” (the leading digits) and a “leaf” (the last digit). Reading one is straightforward once you understand that each row represents a group of numbers sharing the same leading digit, and the individual leaves tell you exactly which numbers fall in that group. Unlike a bar chart or histogram, the plot preserves every original value, so you can recover the full dataset just by reading it.
How Stems and Leaves Work
For most two-digit datasets, the tens digit becomes the stem and the ones digit becomes the leaf. Take this small dataset: 17, 18, 20, 25, 28, 34, 34, 37, 38, 50. The plot would look like this:
- 1 | 7 8
- 2 | 0 5 8
- 3 | 4 4 7 8
- 5 | 0
The vertical line separates stems from leaves. To reconstruct a number, combine the stem with each leaf. The row “3 | 4 4 7 8” gives you 34, 34, 37, and 38. Notice there’s no row for 4, because no values in the 40s exist in the data. Some plots include an empty row to make the gap obvious; others skip it entirely.
Why the Key Matters
Every stem and leaf plot should include a key (sometimes called a legend) that tells you how to interpret the digits. A typical key looks like “3 | 4 represents 34 sit-ups.” This seems obvious for two-digit numbers, but it becomes essential when the data involves hundreds, thousands, or decimals. If you’re looking at three-digit values like 340, 345, and 372, the stems might be 34 and 37 while the leaves are 0, 5, and 2. Without the key, you wouldn’t know whether “3 | 4” means 34 or 340. Always check the key before reading any values.
Reading the Shape of the Data
One of the biggest advantages of a stem and leaf plot is that it doubles as a rough histogram turned on its side. The length of each row shows how many values fall in that range. Longer rows mean more data points are clustered there. You can quickly spot where values concentrate, where gaps appear, and whether the distribution is symmetric (roughly even on both sides of the center), skewed (trailing off to one side), or bimodal (having two peaks).
In the example above, the longest row belongs to the 30s, so that’s where the data clusters. The single value at 50 sits apart from everything else, which might signal an outlier. A common rule for identifying outliers formally is the 1.5 × IQR method: any point more than 1.5 times the interquartile range above the third quartile or below the first quartile counts as an outlier. But visually, a leaf sitting alone far from the rest of the plot is your first clue.
Finding the Median
Because a stem and leaf plot lists every value in order from smallest to largest, it’s a natural tool for finding the median. Count the total number of leaves to get your sample size, then locate the middle value.
If the dataset has an odd number of values, the median is the single middle number. With 10 values (an even count), you take the two middle numbers and average them. In the dataset above, the 5th and 6th values are 25 and 28. Average those: (25 + 28) ÷ 2 = 26.5. That’s the median. You can also find the mode (the most frequent value) by scanning for repeated leaves in the same row. Here, 34 appears twice, making it the mode.
Back-to-Back Plots
A back-to-back stem and leaf plot compares two datasets side by side using a shared column of stems in the center. One dataset’s leaves extend to the right, the other’s extend to the left. You might see test scores for two different classes, or heights of males and females, arranged this way.
Reading the right side works exactly like a standard plot. The left side is the same idea, but the leaves grow outward to the left, so the digit closest to the stem is the ones place. For example, if the left side shows “8 5 2 | 3” with a key stating “2 | 3 represents 32,” then the values are 38, 35, and 32. You read away from the stem to reconstruct each number.
These plots make it easy to compare the center, spread, and shape of two distributions at a glance. If one side’s leaves stretch much further than the other, that group has a wider range. If one side bulges near the top and the other bulges near the bottom, the two groups have different typical values.
Handling Larger or Smaller Numbers
Stem and leaf plots aren’t limited to two-digit whole numbers. For three-digit values, the stem typically holds the hundreds and tens digits while the leaf holds the ones digit. For single-digit data or decimals, the split shifts accordingly: with values like 3.2, 4.7, and 5.1, the whole number becomes the stem and the decimal digit becomes the leaf. The key clarifies everything. “3 | 2 represents 3.2” tells you decimals are in play.
Split Stems for Crowded Data
Sometimes a single stem accumulates too many leaves, making one row unreasonably long and the rest very short. When this happens, the stem can be split into two rows: one for leaves 0 through 4 and another for leaves 5 through 9. You’ll see the same stem number appear twice. For instance, a stem of 3 might have one row for 30 through 34 and a second row for 35 through 39. This spreads the data out and gives you a better picture of the distribution within that range. The plot usually labels both rows with the same stem digit, so if you see a repeated stem, that’s what’s happening.
Quick Steps to Read Any Plot
- Check the key first. It tells you whether stems represent tens, hundreds, or whole numbers before a decimal.
- Reconstruct values. Combine each stem with its leaves, one at a time, to get the actual data points.
- Count the leaves. The total number of leaves equals the total number of data points.
- Look at row lengths. Longer rows show where data clusters. Very short rows or isolated leaves suggest gaps or potential outliers.
- Find the center. Use the ordered values to locate the median directly from the plot.
- Note the range. The smallest value is the first leaf in the top row. The largest is the last leaf in the bottom row. Subtract to get the range.
Once you’ve practiced on a few examples, reading these plots becomes almost instant. The structure does most of the work for you: the data is already sorted, every value is visible, and the overall shape jumps out at a glance.