What does “typical” mean?
Suppose you want to describe the usual wait at a service desk. You have five recorded waits: 4, 5, 6, 7 and 28 minutes. A single number would be convenient, but which one represents the experience? Most people in this small group waited no more than 7 minutes, while one waited much longer. The answer depends on what you want your summary to convey.
The mean, median and mode are measures of center: ways to describe a dataset with a compact summary. None is automatically the “correct” typical value. Each answers a different question, and each leaves out information. Before choosing one, ask what kind of data you have, how the values are distributed and what decision the summary is meant to support.
Use the mean when every value should count
The mean is the total of all values divided by the number of values. For the five waits above, the total is 50 minutes, so the mean is 10 minutes. Every observation contributes to that calculation. This makes the mean useful when the total matters, such as estimating the combined time spent waiting across many visits.
That same feature makes the mean sensitive to unusually high or low values. No one in the example waited exactly 10 minutes, and four of the five waits were shorter. The mean is not wrong; it simply answers a question that may differ from “What did a typical visitor experience?” If you report it, consider whether a few extreme values are pulling it away from where most observations lie.
The mean works especially well as a description when values cluster reasonably closely and there are no exceptional observations dominating the total. It can also be useful with uneven data when your purpose genuinely requires every unit to affect the result. State that purpose rather than treating the mean as a default.
Use the median for the middle observation
The median is the middle value after the observations are sorted. In the wait-time example, the median is 6 minutes. Half the recorded waits are at or below 6 minutes, and half are at or above it. With an even number of observations, the median is usually found by averaging the two middle values.
Because it depends on position rather than the size of every value, the median changes less than the mean when an extreme observation becomes even more extreme. If the 28-minute wait had been 58 minutes, the median would still be 6 minutes. This makes it a useful choice when you want to describe a middle experience in data with a long tail, such as purchase amounts where a few purchases are much larger than the rest.
The median does not tell you how far apart the observations are. Two groups can have the same median but very different ranges. A median of 6 minutes could describe waits tightly grouped near 6, or a mixture of very short and very long waits. Pair it with a measure of spread or a simple description of the distribution when those differences matter.
Use the mode for the most frequent value or category
The mode is the value that appears most often. For the numbers 2, 2, 3, 4 and 9, the mode is 2, the median is 3 and the mean is 4. The mode answers “What occurred most often?” rather than “Where is the middle?” A dataset can have more than one mode if values tie for the highest frequency; it may also have no useful mode when every value appears only once.
Unlike the mean or median, the mode can describe categories. If you are counting which product size customers selected, the most commonly selected size may be more actionable than trying to calculate a numerical center. But frequency alone does not show how dominant that choice was. A leading category chosen by a narrow margin tells a different story from one chosen by nearly everyone. Include counts or proportions when that distinction matters.
Choose the measure after checking the data
Start with the question. If you need a per-item average that reflects a total, consider the mean. If you need the middle observation in a dataset with extreme values, consider the median. If you need the most common value or category, consider the mode. Then inspect the observations rather than trusting the label “average.” A quick ordered list, frequency table or simple chart may reveal clusters, gaps or outliers that a single number hides.
Also check how the data was collected. Missing entries, inconsistent units and narrowly defined groups can change what any measure seems to say. For example, combining short routine visits with lengthy complex visits might produce a center that describes neither group particularly well. In that case, showing separate summaries may be clearer.
A useful report often gives both a center and context: the number of observations, a sense of spread and a brief explanation of any unusual values. Mean, median and mode are tools for making information manageable, not substitutes for looking at the information itself. The best choice is the one that answers your question without concealing a difference your reader needs to understand.
