How can mean and median describe the same dataset differently?
Quick answer
For the dataset 2, 3, 3, 4, 18, the mean is 6 because the values sum to 30 and 30 ÷ 5 = 6. The median is 3 because it is the middle value in the ordered list. The large value pulls the mean upward, while the median marks the dataset’s middle position. Each summarizes the same data in a different way.
The question, the details, the next step.
How can the mean and median give different impressions of the same data? I want to understand the difference with actual numbers.
Consider this small invented dataset: 2, 3, 3, 4, 18. It is already in order, and the value 18 is much larger than the others.
The mean is the sum of all the values divided by how many values there are: (2 + 3 + 3 + 4 + 18) ÷ 5 = 30 ÷ 5 = 6. The median is the middle value in the ordered data, so here it is 3. [S1]
So the mean is 6, while the median is 3. Neither calculation is a mistake: one summarizes the data through the sum of all five values, and the other identifies the central position in the ordered list.
Why does the value 18 change the mean so much more than the median?
The mean includes every value in the sum. Because 18 is much larger than the other four numbers, it raises that sum—and therefore raises the mean. The median depends on the position of the middle value after sorting. In this list, the third value is still 3, so the median remains 3. Extreme values can affect the mean more than the median. [S1]
A useful comparison is to remove 18 just for a check: 2, 3, 3, 4 has mean 3 and median 3. Adding 18 increases the mean to 6, while the median of the five-value list is 3. That comparison illustrates how one unusually large value can change these summaries by different amounts.
How should I explain what each number says without claiming one is always the right answer?
Describe the calculation and its meaning, then note the uneven spread. For this example, you could write: “The mean is 6, the total of the five values divided by five. The median is 3, the middle value when the data is ordered. Because one value, 18, is much larger than the others, the mean is higher than the median.”
That wording says what each summary captures without declaring a universal winner. The mean gives a summary based on all values in the sum; the median gives the middle position. Which is more useful depends on what you want your summary to convey. These two numbers alone do not explain why the values differ or what produced the large value.
Can you give me a repeatable method I can use with another small dataset?
Use this sequence:
- Write down the values and count them.
- Add all values and divide the sum by the count to find the mean.
- Put the values in order and identify the middle value for the median. [S1]
- Compare the results and inspect whether a value far from the rest is influencing the mean.
- Explain both results in relation to the actual data.
For the example, that gives a mean of 6 and a median of 3. The calculations describe this list; they do not, by themselves, establish why 18 is unusually large or which summary is best for a particular decision.
A practical checklist
Record every value and count how many there are.
Add the values, then divide by the count for the mean.
Sort the values before identifying the median.
Compare the summaries and check whether an extreme value affects them differently.
State what each number represents without calling either one universally best.
A few more useful answers
Does the median always stay the same when an extreme value is added?
No. It depends on the number of values and where the added value falls after sorting. In this example, adding 18 to the original four values leaves the median at 3, but other datasets can have different results.
Why can the mean be higher than the median?
The mean uses the sum of all values, so a particularly large value can raise it. The median is the middle value in the ordered list. In the example, the mean is 6 and the median is 3.
Which summary should I report?
Choose based on what you want the summary to express, and consider reporting both when their difference helps explain the data. The calculations alone do not decide which is most useful for every purpose.
Sources
References used for this guide.
- [S1] 2.5 Measures of the Center of the Data — Introductory Statistics 2e openstax.org · Checked