Mean, Median and Mode

A student stares at a list of numbers and needs to pick the right average. This covers the three measures of central tendency, mean, median, and mode, and tells you exactly when to use each one, based on OpenStax Introductory Statistics 2e, sections 2.5-2.6.

The mean is the sum of all values divided by the number of values, symbol x̄ for a sample or μ for a population. The median (symbol M) is the value separating the upper half from the lower half of an ordered data set. The mode has no standard symbol; it is the most frequent value in the data set.

You use the mean for symmetric data with no extreme values, the median for skewed data or data with extreme values, and the mode for categorical data or when you need the most typical case. This shows you the calculation for each, then the decision rule that keeps you from reporting the wrong one.

Add every value in your data set, then divide by the count. For a sample, that is x̄ = sum / n. For a population, μ = sum / N.

Example: data set 2, 4, 6, 8, 10. Sum = 30. n = 5. Mean = 30 / 5 = 6.

The mean uses every value, so a single extreme value pulls it away from the centre. In data set 2, 4, 6, 8, 100, the sum is 120, n = 5, mean = 24. That 24 does not represent any of the first four values well. That is the failure case: reporting the mean for a set with one extreme value misleads everyone reading your summary.

Sort your data from smallest to largest. Find the middle position. The median is resistant to extreme values because it ignores the size of extreme values.

Odd Number of Values

Data: 2, 4, 6, 8, 10. Sorted already. n = 5, middle position is the 3rd value. Median = 6. The set with extreme values 2, 4, 6, 8, 100 also has a median of 6. The median resists the pull of the 100.

Even Number of Values

Data: 2, 4, 6, 8. n = 4. The two middle values are the 2nd and 3rd: 4 and 6. Median = (4 + 6) / 2 = 5.

The even-n case is where students misapply the rule when they also calculate quartiles. The median of an even set is the average of the two middle numbers. Do not pick one or the other. That average is the only correct value.

Count how many times each value appears. The mode is the value that appears most often.

Data: 1, 2, 2, 3, 4. Value 2 appears twice, all others once. Mode = 2.

No Mode

Data: 1, 2, 3, 4, 5. Every value appears exactly once. There is no mode. Some calculators return 'no mode' or list every value. Do not report the mode for continuous data with no repeats; it is meaningless.

Multiple Modes

Data: 1, 1, 2, 2, 3. Values 1 and 2 each appear twice. This is a bimodal data set. Three or more modes is multimodal. Most basic calculators return only one mode. Use MODE.MULT in Excel or inspect the frequency table yourself.

The choice between the mean and the median is the most common decision in descriptive statistics. The rule is straightforward: use the median for skewed distributions and data with extreme values. Use the mean for symmetric distributions with no extreme values.

Skewed right data (positive skew) has a long tail on the right. Examples: income data, house prices. In a right-skewed distribution, the mean is greater than the median. The mean is pulled toward the high end, so it overstates the typical value. The median is the correct centre.

Skewed left data (negative skew) has a long tail on the left. In a left-skewed distribution, the mean is less than the median. The mean is pulled toward the low end, so it understates the typical value. Again, the median is correct.

Symmetric data has mean approximately equal to the median. In that case, the mean is fine. Use the standard deviation with the mean for spread. Use the interquartile range (IQR) with the median for spread.

Data set: 3, 5, 5, 7, 8, 10, 12, 15, 20, 100. This set has one extreme value (100) and a slight right skew.

Mean: Sum = 3 + 5 + 5 + 7 + 8 + 10 + 12 + 15 + 20 + 100 = 185. n = 10. Mean = 185 / 10 = 18.5. Does 18.5 represent the data? No. Nine of the ten values are below 20, and the mean is 18.5 because the 100 pulled it up.

Median: Sorted data: 3, 5, 5, 7, 8, 10, 12, 15, 20, 100. n = 10, even. Two middle values: 5th (8) and 6th (10). Median = (8 + 10) / 2 = 9. The median of 9 is much closer to the bulk of the data. That is the better centre for this set.

Mode: Value 5 appears twice. All others appear once. Mode = 5.

If you report the mean of 18.5 for this data set, you misrepresent the typical value. Report the median (9) as the centre, and use the IQR for the spread. The mode (5) tells you the most common single value, which is useful if the data set represents categories or counts where the typical case matters. For this numeric set with a clear extreme value, the median is the correct choice.

Use this three-way comparison to decide which measure of central tendency to report.

Data is symmetric, no extreme values: Use the mean. It is the most efficient estimator of the centre. Accompany it with the standard deviation.

Data is skewed or has extreme values: Use the median. It is resistant to extreme values. Accompany it with the interquartile range (IQR).

Data is categorical, or you need the most typical single value: Use the mode. It works for nominal data (colours, brands) where the mean and median have no meaning. For continuous data with no repeated values, skip the mode.

Data has multiple modes: Report all modes and explain that the data set is bimodal or multimodal. Do not average the modes. That is a common error with no mathematical basis.

Skewness measures asymmetry. A positive skewness value (above 1) means the tail is on the right, so the mean is larger than the median. A negative skewness value (below -1) means the tail is on the left, so the mean is smaller than the median. Symmetric data has skewness near zero, and the mean and median are approximately equal.

Extreme values, data points flagged by the 1.5*IQR rule, pull the mean but do not move the median. If your data set contains extreme values, the median is the correct centre. Do not remove extreme values automatically; they may be legitimate extreme values that need separate investigation. The 1.5*IQR rule flags them for review, not for deletion.

The mean vs median decision is the single most common failure in reporting descriptive statistics. A student reports the mean for a skewed data set because that is what the calculator shows first. Press the button, get the mean, publish the mean. That is wrong. The median is the better centre for skewed distributions, and the IQR is the better spread measure alongside it.

Measures of central tendency suit high-school AP Statistics students who need to calculate by hand and check against a calculator. They suit college Stats 101 students who must choose between the mean and median for a given distribution. They suit analysts and researchers who need a quick, reliable summary of a small data set. They suit self-taught data enthusiasts who need a clear, worked example for each measure.

Skip these measures if you need inferential statistics, hypothesis tests, confidence intervals, p-values, regression, or probability theory. Those belong to a different branch of statistics. Skip them if you need multivariate analysis, time-series decomposition, or Bayesian methods. Go to a full statistical package like R, Python, or SPSS for those tasks.

The single thing that most often goes wrong here: a student calculates the mean, sees it on the calculator screen, and reports it without checking the shape of the data. If the data is skewed or has extreme values, that mean is the wrong centre. Sort the data, look for the tail, and report the median instead.