Selecting the Right Central Tendency Measure for Your Data

Every dataset tells a story – but only if you ask it the right question. One of the most foundational decisions in data analysis is choosing how to summarize that data. The mean, median, and mode are the three primary measures of central tendency, and while they each describe what is “typical” or “average” in a dataset, they do so in very different ways. Research published on PubMed highlights that a lack of understanding about when to use each metric is surprisingly common – even among researchers. Knowing which measure to reach for, and why, can be the difference between an accurate finding and a misleading one.

What are measures of central tendency? 🔗

A measure of central tendency is a single value that attempts to describe an entire dataset by identifying its central position. According to Laerd Statistics, all three measures are valid, but under different conditions some become far more appropriate than others. The three measures work as follows:

• Mean: The sum of all values divided by the number of values. It factors in every data point.
• Median: The middle value when all data points are arranged in order. If there is an even number of values, it is the average of the two middle ones.
• Mode: The value that appears most frequently in a dataset. A dataset can have one mode, multiple modes, or no mode at all.

The key to choosing correctly lies in understanding your data’s level of measurement, its distribution shape, and the presence of any extreme values (outliers).

When to use the mean 🔗

The mean is the most widely used and generally preferred measure of central tendency. Lumen Learning’s Introduction to Statistics describes it as the go-to measure for most situations, but cautions that certain data conditions undermine its reliability.

The mean works best when:

• Your data is normally distributed (symmetrical, bell-shaped curve)
• There are no significant outliers
• You are working with continuous or interval/ratio data
• You need to use the value in further statistical calculations (e.g., regression, standard deviation)

Why the mean fails with outliers 🔗

The mean’s biggest vulnerability is its sensitivity to extreme values. Consider a factory where eight workers earn between $12,000 and $18,000 per year, but two senior managers earn $90,000 and $95,000. Laerd Statistics uses exactly this kind of scenario to demonstrate how the mean salary climbs to $30,700 – a figure that does not accurately reflect what a typical worker earns. The two high salaries are distorting the result. In cases like this, the mean becomes misleading rather than informative.

This is not a rare edge case. OpenStax Introductory Business Statistics confirms that the mean is affected by outliers that do not influence the median at all – a critical distinction when analyzing real-world data like incomes, house prices, or test scores.

When to use the median 🔗

The median is the preferred measure whenever the mean is likely to be pulled away from the true center of your data. Penn State’s STAT 200 course materials explain that for distributions that have outliers or are skewed, the median is often the preferred measure of central tendency because it is more resistant to those extreme values.

Use the median when:

• The data has a skewed distribution (positively or negatively skewed)
• There are outliers present that would distort the mean
• You are working with ordinal data (ranked categories where intervals are unequal)
• There are missing or undetermined values at the extremes of your dataset

Skewed distributions and the median’s advantage 🔗

Skewness is the primary reason we need more than one measure of central tendency. Statistics LibreTexts explains that in a positively skewed distribution – where a few very high values stretch the right tail – the mean gets pulled toward those high scores, placing it well above where most of the data actually sits. The median, positioned between the mean and mode, gives a much more accurate picture of the typical value.

Income data is the classic real-world illustration. The University of Utah’s Sociology 3112 course points to Major League Soccer salaries as a striking example: the mean annual salary was approximately $138,000, but the median was only around $53,000 – a gap created almost entirely by a handful of extremely high-earning players. The median told the real story.

As a general rule from Statistics LibreTexts: when data is skewed to the right, the mean is greater than the median; when data is skewed to the left, the mean is smaller than the median. The further apart these two values are, the more the median should be preferred.

When to use the mode 🔗

The mode stands apart from the mean and median because it is the only measure of central tendency that can be used with nominal (categorical) data – data with no numerical order. A paper in PMC’s medical statistics journal notes that the mode is the only measure of central tendency applicable to data measured on a nominal scale.

The mode is most useful when:

• Working with nominal or categorical data (e.g., religion, preferred brand, mode of transport)
• You need to identify the most common outcome in survey or market research data
• The dataset has a bimodal or multimodal distribution with two or more peaks
• You need a quick, general snapshot without calculation – no arithmetic is required to find the mode

Limitations of the mode 🔗

The mode has notable limitations. A dataset may have no mode at all if every value is unique, or it may have several modes, making it difficult to identify a single “typical” value. Rio Salado College’s statistics resource notes that a dataset will always have exactly one mean and one median, but the mode offers no such guarantee. For continuous numerical data, the mode is rarely the most informative choice.

The role of data distribution and level of measurement 🔗

Two factors should guide every decision about which measure to use: the level of measurement of the variable and the shape of its distribution.

Level of measurement 🔗

Scribbr’s guide on central tendency lays out a clear framework: the mode can be used for any level of measurement but is most meaningful for nominal data; the median requires at least ordinal data; and the mean requires interval or ratio data. In other words, if you are analyzing a nominal variable such as political affiliation or favourite food, only the mode is appropriate. For ordinal variables like satisfaction ratings, the median is the better choice. For interval or ratio variables like temperature, age, or income, the mean or median can both work – with the distribution shape determining which is best.

Symmetric vs. skewed distributions 🔗

When data follows a normal (symmetric) distribution, all three measures converge at the same point. Statistics by Jim explains that in a perfectly symmetrical distribution, the mean, median, and mode are all equal and all valid – but the mean is preferred because it uses every data point in its calculation, making it the most statistically efficient. For skewed data, the median is more accurate. The Ledidi Academy summarizes this cleanly: for categorical variables, use the mode; for numerical variables in a normal distribution, use the mean; for numerical variables in a skewed distribution, use the median.

Using all three measures together 🔗

In practice, calculating all three measures can give a fuller picture of your data before you decide which one to report. Scribbr points out that the three measures have complementary strengths and limitations, and using them together can reveal important information about the shape of your distribution – for instance, a large gap between the mean and median is itself a signal of skewness. Researchers often report more than one measure alongside each other, particularly when communicating findings to a non-specialist audience. The mode can describe the most common response in a survey; the median can report the typical income in a community; and the mean can anchor further inferential statistical analysis. None of them is universally superior – each one answers a slightly different question about your data.

A quick decision guide 🔗

To summarize the core decision logic:

• Nominal data: Use the mode. It is the only option.
• Ordinal data: Use the median. It captures rank without assuming equal intervals.
• Interval/ratio data, symmetric distribution: Use the mean. It is the most powerful and precise.
• Interval/ratio data, skewed distribution or outliers present: Use the median. It is the most resistant to distortion.
• Categorical data or quick frequency check: Use the mode. No calculation needed.

This framework does not eliminate judgment – context always matters. But it gives any researcher or analyst a reliable starting point for selecting the most honest and accurate summary of their data.

What do you think? When you encounter statistics reported in the news – like average income or average house prices – do you wonder whether the mean or median was used, and whether that choice changes the story being told? And if you were designing a survey to capture people’s most common political concern, which measure of central tendency would you choose, and why?

References

1. https://pubmed.ncbi.nlm.nih.gov/36322179/
2. https://statistics.laerd.com/statistical-guides/measures-central-tendency-mean-mode-median.php
3. https://courses.lumenlearning.com/introstats1/chapter/when-to-use-each-measure-of-central-tendency/
4. https://openstax.org/books/introductory-business-statistics-2e/pages/2-6-skewness-and-the-mean-median-and-mode
5. https://online.stat.psu.edu/stat200/lesson/2/2.2/2.2.4/2.2.4.1
6. https://stats.libretexts.org/Courses/Adler_University/Graduate-Level_Statistics_in_Psychology/05:_Measures_of_Central_Tendency/5.03:_The_Mean_Median_and_Mode_in_Normal_and_Skewed_Distributions
7. https://soc.utah.edu/sociology3112/central-tendency-variability.php
8. https://stats.libretexts.org/Courses/Los_Angeles_City_College/Introductory_Statistics/02:_Descriptive_Statistics/2.07:_Skewness_and_the_Mean_Median_and_Mode
9. https://pmc.ncbi.nlm.nih.gov/articles/PMC3157145/
10. https://www.riosalado.edu/web/oer/WRKDEV100-20011_INTER_0000_v1/lessons/Mod05_MeanMedianMode.shtml
11. https://www.scribbr.com/statistics/central-tendency/
12. https://statisticsbyjim.com/basics/measures-central-tendency-mean-median-mode/
13. https://ledidi.com/academy/measures-of-central-tendency-mean-median-and-mode