Can a data set have no mode? Yes, and here is when
By Jude Wallis · Published
Yes. If every value appears exactly once, the data set has no mode. That is a naming convention rather than a deep fact: the counts are all tied, and statisticians agree to say there is no mode instead of calling every value a mode, because a summary that names every observation summarizes nothing.
AP Statistics: Unit 1 (topics 1.3 Tabular Representation and Summary Statistics for One Categorical Variable, 1.6 Descriptions for One Quantitative Variable Distributions, 1.7 Summary Statistics for One Quantitative Variable). In the Fall 2026 AP Statistics course, describing the shape of a quantitative distribution, where unimodal and bimodal live, is Unit 1 topic 1.6, and summary statistics for a quantitative variable are topic 1.7. The most common category of a single categorical variable belongs to topic 1.3, which is where a mode does its real work on this exam.
The rule: no mode, one mode, or several
The mode is the value that appears most often. Counting the values gives you one of three outcomes.
- No mode. No value is more frequent than any other, because every count is the same. Almost always that means every value appears exactly once, as in 4, 7, 9, 12, 16, 21.
- One mode. A single value has the highest count, as in 2, 4, 4, 4, 7, 9, where 4 appears three times.
- Several modes. Two or more values tie for the highest count, as in 5, 5, 7, 9, 9, 12, where both 5 and 9 appear twice.
The mode is the only common measure of center that can fail to exist or fail to be unique. The mean always exists for quantitative data and is always a single number; sort the data and the median is always defined too. The mode depends on repeats, and repeats are not guaranteed.
If you need the counting procedure itself, including how to read a mode off a frequency table or a histogram, how to find the mode walks through it. This page is about the edge cases.
Why "no mode" is a convention, not a discovery
Look at 4, 7, 9, 12, 16, 21 again. Every value has a count of 1, so technically every value ties for the largest count. The tie-breaking rule for modes says all tied values are modes, and applying it literally would make all six values modes.
Statisticians do not do that, and the reason is practical. A measure of center is supposed to compress a data set into something shorter than the data set. Reporting all six values as modes returns the whole list, so it communicates nothing. Saying "no mode" is the honest short version: this data set has no most-common value.
You will occasionally meet a textbook that says every value is a mode in this case, or one that says the mode is undefined. All three phrasings describe the same counts. Nothing about the data changes, only the label, which is what makes this a convention. On an AP-style question, write "there is no mode" and you are matching standard practice.
The rule that follows from this is about ties, not about the number 1. No mode means every count is tied, whatever that count is. In a list of measurements the tie is almost always at one apiece, but four candidates holding exactly 20 votes each are tied just as completely, and naming all four as modes returns the whole tally instead of summarizing it, so the answer there is also no mode.
One related convention is worth separating out. A tie for the top between some of the values, with others lower, makes two or more modes, not "no mode." Answering "no mode" when 5 and 9 both appear twice is a genuine error, not a stylistic choice.
Measured data almost never has a mode
The finer your measurement, the less likely any two values match exactly. Eight puzzle-solving times recorded to the nearest tenth of a second, 12.4, 13.1, 13.7, 14.0, 14.2, 14.9, 15.3, 16.1, are all distinct, so there is no mode. Record them to the nearest hundredth and the chance of a tie drops further still.
That is not a defect in the data. It tells you the variable is effectively continuous, and for continuous variables the useful idea is not a single most-common value but a most-common region.
So you switch tools. Group the values into bins and report the modal class, the interval covered by the tallest bar of the histogram. With bins one second wide, each including its left endpoint and excluding its right one, those eight times fall as 1, 2, 3, 1, 1 across 12 to 13, 13 to 14, 14 to 15, 15 to 16, and 16 to 17 seconds, so the modal class is 14 to 15 seconds. Notice you can only ever name the interval, never a value inside it, because the histogram has already thrown the individual values away.
The catch is that the modal class depends on the bins you chose. Different bin widths can hand you a different tallest bar, so a dotplot, which keeps every value, is the safer graph when the data set is small. Dotplot vs histogram vs stemplot compares them.
Multimodal: two, three, or more modes
The opposite edge case is a data set with several modes. In 3, 3, 6, 6, 10, 10, 15, three values each appear twice and 15 appears once, so 3, 6, and 10 are all modes.
Here is the distinction that trips people up. Having several modes is a fact about counts. Being a bimodal distribution is a fact about shape: two clear peaks in the graph with a dip between them. The two do not have to agree.
- 1, 1, 2, 2, 3 has two modes (1 and 2) but forms one hump, so its shape is unimodal.
- A histogram of 200 exam scores can show two obvious peaks, one near 55 and one near 85, while no single score repeats often enough to matter as a count.
So describe shape from the graph and report modes from the tally, and never let one answer the other's question. What is a bimodal distribution covers the shape side, including why two peaks usually mean two groups got mixed together.
What to report when there is no mode
"No mode" is a complete answer to the question that was asked, but it is a thin description of the data. Add whatever actually describes it.
- For quantitative data, give the shape, a center, and a spread. The center is the mean if the distribution is roughly symmetric and the median if it is skewed or has an extreme value, since the median barely responds to one, and the spread is the standard deviation or the interquartile range to match. How to describe a distribution gives the full sentence, and mean vs median covers the choice.
- For grouped or continuous data, report the modal class instead of a mode, and say which bin width you used.
- For categorical data, no mode means every category has the same count, which is itself worth saying: the distribution is flat, with no most common category.
And if the mode does exist, remember it can sit far from the center. In 1, 1, 40, 50, 60 the mode is 1, which describes the low end rather than a typical value. The mode answers "what is most common," which is not always the same question as "what is typical."
Common mistakes
- Answering "no mode" for a tie at the top. Two values tied for the highest count, with the other values lower, means two modes. No mode requires every value to be tied, which in a list of numbers usually means every value appearing exactly once.
- Answering "all of them" for a data set with no repeats. The convention is to say there is no mode.
- Calling a data set bimodal because two values tie. Bimodal describes two peaks in the graph, not a tie in the tally.
- Reading a single mode off a histogram. Report the modal class, the tallest bar's interval.
- Reporting the frequency instead of the value. If navy hoodies sold 24 times and that is the largest count, the mode is navy, not 24.
- Assuming no mode means the data are uniform. No mode is a statement about the counts. A uniform distribution is a statement about shape, and a data set with no repeats can be strongly skewed.
- Forgetting to sort or tally first. Most "no mode" answers on long lists are miscounts.
A data set with no mode
Six students report how many books they read over the summer: 4, 7, 9, 12, 16, 21. Find the mode, the median, and the mean, and summarize what each one does or does not tell you.
Sort the values (already sorted): 4, 7, 9, 12, 16, 21, with .
Tally the counts: 4 appears once, 7 once, 9 once, 12 once, 16 once, 21 once. Check the tally against the sample size: , which matches .
Every count is 1, so no value appears more often than any other. The data set has no mode.
Median: is even, so average the 3rd and 4th values: books.
Mean: the sum is , so books.
Interpret. The mode is unavailable, so it contributes nothing here. The median says half the students read 10.5 books or fewer. The mean of 11.5 sits above the median, the usual sign of a slight stretch toward the high values, and 21 is the value doing the stretching.
No mode, because all six values appear exactly once. The median is 10.5 books and the mean is 11.5 books. Report those two plus the shape; "no mode" is the correct answer to the mode question but describes nothing on its own.
No mode, so report the modal class
Eight people are timed solving a puzzle, in seconds: 12.4, 13.1, 13.7, 14.0, 14.2, 14.9, 15.3, 16.1. Show that there is no mode, then group the data into bins one second wide and report the modal class.
Check for repeats: all eight times are distinct, so every count is 1 and there is no mode.
Choose bins one second wide, running 12 to 13, 13 to 14, 14 to 15, 15 to 16, and 16 to 17 seconds, with each bin including its left endpoint and excluding its right one.
Sort the values into bins. 12 to 13 holds 12.4, so its count is 1. 13 to 14 holds 13.1 and 13.7, count 2. 14 to 15 holds 14.0, 14.2, and 14.9, count 3. 15 to 16 holds 15.3, count 1. 16 to 17 holds 16.1, count 1.
Check the counts: , which matches .
The tallest bar is the 14 to 15 second bin with 3 values, so the modal class is 14 to 15 seconds.
Compare with the other centers. The sum is 113.7, so seconds, and the median is seconds. Both land inside the modal class, which is what you expect for a roughly symmetric distribution.
State the limitation. If you had chosen bins two seconds wide, the tallest bar could cover a different interval, so name your bin width whenever you report a modal class.
There is no mode, since all eight times are distinct. With bins one second wide the modal class is 14 to 15 seconds, holding 3 of the 8 times. The mean is about 14.21 seconds and the median is 14.1 seconds, both inside that interval.
Frequently asked questions
If every value appears once, is every value a mode?
By the strict tie-breaking rule you could argue so, since all the counts are tied. The standard convention is to say the data set has no mode, because listing every observation as a mode gives back the data instead of summarizing it. Write "no mode" on an exam.
Can a data set have two modes?
Yes, and three or more. Any values tied for the highest count are all modes. A tie is not the same as no mode, and it does not by itself make the distribution bimodal, since bimodal describes two peaks in the graph rather than a tie in the tally.
Does no mode mean the data are uniform?
No. No mode usually means nothing repeats, which is common in precisely measured data. A uniform distribution is about shape, roughly equal frequencies across the range. Bin the values 4, 7, 9, 12, 16, 21 into equal-width intervals and the counts are not equal, so the data have no mode and are still not uniform.
Can categorical data have no mode?
Yes, but it takes an exact tie across every category, such as 20 votes each for four candidates. That is worth reporting in words: the distribution is flat, with no most common category. If two categories tie at the top and others are lower, you have two modes, not none.
Does the mode matter on the AP exam?
Rarely for quantitative data, where the mean and the median do the work and the mode-related vocabulary you need is unimodal and bimodal for shape. It matters more for a single categorical variable, where the most common category is exactly what a frequency table or a bar graph is showing you.