Descriptive Statistics: Definition, 4 Types & Examples

A spreadsheet full of numbers tells you almost nothing until you summarize it. Descriptive statistics are the tools that do that summarizing. They describe the main features of a dataset, its center, its spread, and its shape, without trying to predict anything beyond the numbers in front of you.

I'll walk one small dataset through every measure, put the formula next to each worked number, and finish with two ways to get all of it in seconds. By the end you'll be able to summarize any column of numbers and know which chart to show it with.

What are descriptive statistics?

Descriptive statistics are a type of statistical analysis that organizes, summarizes, and presents data in a form a person can actually read. That's the whole job. They don't generalize, infer, or predict. They take a raw dataset and hand you a snapshot of its key characteristics.

Two things make that snapshot worth the effort.

First, it lets you understand a dataset almost at a glance. Instead of scrolling 500 rows, you read one mean, one standard deviation, one chart, and you know roughly what you're dealing with.

Second, it's the groundwork for everything that comes after. Before you run anything heavier, the summary tells you which methods even make sense for the data you have.

Descriptive vs. inferential statistics

Descriptive statistics summarize the data you actually collected. Inferential statistics use a sample to draw conclusions about a larger population you didn't measure. Describe what's in front of you, infer about what isn't.

That's the short version. For the full breakdown, see our guide on descriptive vs. inferential statistics.

The 4 types of descriptive statistics

Most guides group the measures into four types: frequency, central tendency, dispersion, and position. It's a cleaner way to think about your options than memorizing a list of formulas, and it maps onto the questions you'll actually ask of a dataset.

Here's what each type measures.

Frequency: how often values occur

Frequency counts how many times each value or category shows up, usually as a raw count, a percentage, or a frequency table. In a set of product ratings from 1 to 5, the frequency tells you how many customers gave each score, and it can also flag the ones who left no rating at all.

The kind of data you have shapes which measures make sense next, so it helps to know your data levels and measurement before you go further. Counts work for everything; means only work for numbers.

Central tendency: the center of your data

Central tendency points at the typical or representative value. You've got three options:

• Mean: the average. Add up every value and divide by how many there are. The formula is x̄ = Σx / n.
• Median: the middle value once the data is sorted from low to high. If there's an even count, it's the average of the middle two.
• Mode: the value that shows up most often. Some datasets have one mode, some have several, and some have none.

The mean is the usual pick, but it gets dragged around by outliers. One billionaire walks into a room and the average net worth is useless; the median barely moves. When your data is skewed, reach for the median.

Dispersion: how spread out your data is

Dispersion, also called spread, tells you how far your values sit from each other and from the center. Two classes can both average 85 on a quiz while one has everyone clustered at 84 to 86 and the other ranges from 60 to 100. The average hides that difference, and dispersion is what reveals it.

• Range: the simplest measure. Subtract the smallest value from the largest.
• Variance: the average of the squared distances from the mean. The sample formula is s² = Σ(x − x̄)² / (n − 1).
• Standard deviation: the square root of the variance (s = √s²), which puts the number back on the same scale as your data, so it's the one you'll usually report.
• Interquartile range (IQR): the spread of the middle 50% of your data, found by subtracting the first quartile from the third.

💡 Tip: report the standard deviation alongside the mean, and the IQR alongside the median. The reporting convention from clinical research pairs them that way because each spread measure matches the center measure it's most honest about.

Position: where a value falls in the pack

Position measures tell you where a single value stands in the sorted dataset. A student in the 90th percentile scored higher than 90% of the class. Quartiles are the common case: they split the sorted data into four equal chunks, where the first quartile (Q1) marks the 25% point, the second is the median, and the third (Q3) marks the 75% point. The gap between Q1 and Q3 is the IQR you just met in dispersion.

A worked example: one quiz, every measure

Say you teach a class of 11 students, and their last quiz scores were:

85, 92, 78, 90, 95, 88, 76, 85, 80, 82, 91

Sorted from low to high, that's 76, 78, 80, 82, 85, 85, 88, 90, 91, 92, 95. Let's run it through all four types.

Frequency: every score appears once except 85, which shows up twice. That's the only repeat in the set.

Mean: add the scores (942) and divide by 11. x̄ = 942 / 11 = 85.64

Median: with 11 values, the middle one is the 6th in the sorted list. Median = 85

Mode: the most frequent score. Mode = 85

Range: highest minus lowest. 95 − 76 = 19

Variance and standard deviation: take each score's distance from the mean, square it, add the squares (that total comes to about 378.55), then divide by n − 1 (which is 10) for the sample variance. s² = 378.55 / 10 = 37.85 s = √37.85 ≈ 6.15

Quartiles and IQR: split the sorted list at the median. The lower half (76, 78, 80, 82, 85) has a median of 80, so Q1 = 80. The upper half (88, 90, 91, 92, 95) has a median of 91, so Q3 = 91. IQR = 91 − 80 = 11

A quick note on honesty: quartiles depend on the method. I used the one that excludes the median from each half, so your software may report Q1 and Q3 a point or two off. Different tools use different conventions here, and both are valid.

Here's every value in one place.

Read together, the table says the class landed in the mid-80s, scores were fairly tight (a standard deviation of about 6 points), and nobody was a wild outlier.

How to visualize descriptive statistics

Numbers summarize a dataset. A good chart makes the summary obvious. The trick is matching the chart to the question you're asking.

• Frequency table: best when you need exact counts or you're working with categories. It's the plainest way to show how often each value appears.
• Bar chart: best for comparing categories side by side, like average score by class section. The bars don't touch, which signals the categories are separate.
• Histogram: best for the shape of one numeric variable. The bars touch here because the data is continuous, and the result shows you whether the data is bunched up, spread out, or skewed.
• Box plot: best for seeing the median, the quartiles, the IQR, and any outliers in a single picture. If you want spread and position at a glance, this is the one.

When you're ready to build these, Julius's graph maker can turn a dataset into any of them without a formula.

How to calculate descriptive statistics

You can work the numbers by hand for 11 scores. For a few thousand rows, you'll want software. Two practical routes:

Excel's Analysis ToolPak. Excel ships with a built-in descriptive statistics tool, though you may need to switch it on first (File, then Options, then Add-ins, then enable Analysis ToolPak). Once it's on, go to the Data tab, click Data Analysis, pick Descriptive Statistics, select your input range, check the "Summary statistics" box, and run it. Excel spits out the mean, median, mode, standard deviation, sample variance, range, minimum, maximum, and count in one block. Our walkthrough on using Excel for statistical analysis covers the setup in more detail.

💡 Note: if you don't see Data Analysis on the Data tab, the ToolPak add-in isn't enabled yet. That's the step most people miss.

Julius. Upload a CSV or connect a data source, then ask in plain English for descriptive statistics on each column. You get the summary table and the charts back in seconds, which is the point once your dataset outgrows a textbook example. It's the route I reach for when hand-calculating stops being realistic.

Go from raw data to descriptive statistics in seconds

Hand-calculating a quiz with 11 scores is fine. The datasets you hit at work rarely stop at 11, and that's where summarizing by hand turns into an afternoon you didn't plan to spend.

We built Julius to collapse that work into a single question. Here's how Julius helps:

• Ask in plain language: type "give me descriptive statistics for every column" and Julius returns the mean, median, spread, and quartiles without a formula in sight.
• Connect your data: link PostgreSQL, Snowflake, or BigQuery, or just upload a CSV, so your summary reflects the real dataset instead of a stale copy.
• Built-in visualization: Julius pairs the numbers with histograms, box plots, and bar charts on the spot, so you don't hop into another tool to see the shape.
• One-click sharing: turn the whole analysis into a report you can send without reformatting anything.

Ready to skip the hand-calculations and go straight to the summary? Try Julius for free today.

Frequently asked questions

What is the difference between descriptive and inferential statistics?

Descriptive statistics summarize the data you have, while inferential statistics use a sample to draw conclusions about a larger population you didn't measure. Descriptive methods describe; inferential methods generalize. For the full comparison, see our guide on inferential statistics.

What are examples of descriptive statistics?

Common examples include the mean, median, and mode (central tendency), the range and standard deviation (dispersion), frequency counts and percentages, and percentiles (position). In the quiz example above, the mean of 85.64 and the standard deviation of about 6.15 are both descriptive statistics.

What are the 5 descriptive statistics?

The five most commonly used descriptive statistics are the mean, median, mode, range, and standard deviation. These individual measures map onto the four measure types (frequency, central tendency, dispersion, and position), so the "five" refers to the specific values people calculate most often.

What are the limitations of descriptive statistics?

Descriptive statistics summarize a dataset but can't generalize beyond it, so they tell you nothing about a wider population. A single summary number can also hide outliers or a skewed shape, and descriptive methods never show cause and effect. For that, you need inferential or other advanced analysis.