A frequency distribution is a way of organizing data to show how often each value (or range of values) occurs. If you measured the resting pulse rate of 63 people, for instance, rather than staring at 63 individual numbers, you could group them into ranges and count how many people fall into each one. That count is the “frequency,” and the organized summary is the distribution. It’s one of the most fundamental tools in statistics, used everywhere from medical research to business analytics.
How a Frequency Distribution Works
At its core, a frequency distribution takes raw data and sorts it into categories, then tallies how many observations land in each category. The result is typically displayed as a table with two columns: the category (or range of values) and the count of observations in that category.
Consider a real example from clinical research. Researchers measured resting pulse rates in 63 healthy volunteers and organized the results like this:
- 60–64 beats/min: 2 people
- 65–69 beats/min: 7 people
- 70–74 beats/min: 11 people
- 75–79 beats/min: 15 people
- 80–84 beats/min: 10 people
- 85–89 beats/min: 9 people
- 90–94 beats/min: 6 people
- 95–99 beats/min: 3 people
In seconds, you can see that most people cluster in the 75–79 range, with fewer people at the extremes. That instant clarity is the whole point. Without the frequency distribution, you’d be scanning a list of 63 numbers trying to spot a pattern.
Ungrouped vs. Grouped Distributions
There are two main formats, and the right one depends on your data type.
An ungrouped frequency distribution lists every unique value and its count. This works well for categorical variables, things like eye color, favorite brand, or yes/no responses, where there are a limited number of distinct values. If you surveyed 100 people about their preferred coffee type and 42 said espresso, 35 said drip, and 23 said cold brew, that’s an ungrouped distribution.
A grouped frequency distribution bundles numerical data into ranges called class intervals. The pulse rate example above is grouped: instead of listing every possible pulse rate individually, the researchers created intervals of five beats per minute. This is the standard approach for quantitative variables, especially when values span a wide range. Each interval has a lower limit and an upper limit, and the intervals cannot overlap. An observation falls into whichever interval it’s greater than or equal to at the lower end and less than at the upper end.
Choosing the Right Number of Groups
When you build a grouped frequency distribution, you need to decide how many intervals to use. Too few and you lose detail; too many and the table becomes as hard to read as the raw data. Two common rules of thumb give you a starting point.
Sturges’ rule sets the number of intervals to roughly 1 + 3.3 × log₁₀(N), where N is your number of observations. For 63 data points, that works out to about 7 intervals. An alternative called the Rice rule sets the number of intervals to twice the cube root of N, which for 63 observations gives about 8. Neither rule is absolute. They’re guidelines to get you in the right ballpark before you adjust based on what makes the data easiest to interpret.
Relative and Cumulative Frequency
A basic frequency distribution tells you raw counts, but two extensions make it more useful.
Relative frequency converts each count into a proportion or percentage by dividing it by the total number of observations. In the pulse rate data, 15 out of 63 people fell in the 75–79 range, giving a relative frequency of about 23.8%. This lets you compare distributions of different sizes. If one study has 63 participants and another has 500, raw counts aren’t comparable, but percentages are.
Cumulative frequency adds up the counts as you move through the intervals. By the time you reach the 75–79 range in the pulse data, the cumulative frequency is 35, meaning 35 of 63 volunteers (55.6%) had a pulse of 79 or below. This is useful for answering questions like “what percentage of values fall below a certain threshold?”
Visualizing Frequency Distributions
Tables work, but a picture often communicates faster. The two main chart types map directly onto the two distribution formats.
A histogram is the standard graph for grouped, continuous data. The x-axis represents the class intervals, the y-axis shows frequency, and the bars touch each other. That contact between bars is intentional: it signals that the data flows along a continuous scale, with no gaps between ranges. You can’t rearrange the bars because the order is determined by the data’s natural progression.
A bar chart is used for categorical (ungrouped) data. The bars are separated by gaps to reinforce that each one represents a distinct, independent category. Unlike a histogram, you can sort the bars however you want: alphabetically, by frequency, or by any other logic that fits your analysis.
A third option, the ogive, graphs cumulative frequency. You plot a point at the upper boundary of each interval at the height of its cumulative frequency, then connect the points with a smooth curve. The result is an S-shaped line that climbs from zero to your total sample size. Ogives are particularly handy for reading off percentiles or median values at a glance.
What the Shape Tells You
Once you’ve built a frequency distribution and graphed it, the shape of that graph reveals a lot about your data.
A symmetric distribution looks roughly the same on both sides of the center. The classic bell curve is the most familiar example. If the left and right tails mirror each other, the data is balanced around its average.
A skewed distribution has one tail that stretches out longer than the other. Right-skewed (or positively skewed) data has a long tail extending toward higher values, which is common in measurements that have a natural floor but no ceiling, like income or time-to-failure in reliability testing. Left-skewed data tails off toward lower values. Skewness is essentially a measure of how lopsided the distribution is.
A distribution can also be described by how peaked or flat it is compared to a normal bell curve. A sharply peaked distribution with heavy tails means more extreme values than you’d typically expect, while a flatter distribution means values are more evenly spread out. Statisticians call this property kurtosis, though for practical purposes what matters is recognizing whether your data has an unusual number of outliers (sharp peak, heavy tails) or is relatively uniform.
Why Frequency Distributions Matter
Frequency distributions are often the very first step in data analysis because they turn a disorganized pile of numbers into something you can actually reason about. Before calculating averages, running statistical tests, or building models, you need to know what your data looks like. Is it clustered around a single value or spread out? Is it symmetric or lopsided? Are there outliers?
In medical research, frequency distributions summarize patient demographics, lab results, and clinical outcomes. In business, they reveal how customer spending, delivery times, or product ratings are distributed. In education, they show the spread of test scores across a class. The underlying logic is always the same: count how often each value or range of values shows up, organize those counts, and look for patterns. Everything else in statistics builds on that foundation.