What Is Mean, Median, and Mode?
Mean, median, and mode are the three primary measures of central tendency, statistical values that describe where the center of a dataset lies. The mean is the arithmetic average, calculated by adding all values and dividing by the count. The median is the middle value when data is arranged in order. The mode is the most frequently occurring value. Each tells you something different about your data's center, and choosing the right one depends on your data's shape, scale, and purpose. In survey research, all three appear regularly: means for rating scales, medians for income data, and modes for categorical preferences.
Why Mean, Median, and Mode Matter in Research
Picking the wrong measure of central tendency can distort your findings. Reporting the mean household income in a survey sample that includes a few millionaires will overstate what a typical respondent earns. The median would be more representative. Understanding when each measure is appropriate helps you summarize data accurately and prevents stakeholders from drawing wrong conclusions from your summary statistics.
How Mean, Median, and Mode Work
Mean (Arithmetic Average)
Formula:
Mean = (x1 + x2 + x3 +... + xn) / n
Where x1 through xn are the individual data values and n is the total number of values.
Worked Example:
A product team collects satisfaction scores from 8 customers: 6, 7, 7, 8, 8, 8, 9, 10.
Mean = (6 + 7 + 7 + 8 + 8 + 8 + 9 + 10) / 8 Mean = 63 / 8 Mean = 7.875
The average satisfaction score is 7.875 out of 10.
Strengths: Uses every data point. Well-suited for further statistical calculations (confidence intervals, t-tests).
Weakness: Sensitive to extreme values (outliers). One very high or very low score can pull the mean away from where most data sits.
Median (Middle Value)
How to find it:
- Sort the data from smallest to largest
- If n is odd, the median is the middle value
- If n is even, the median is the average of the two middle values
Worked Example:
Using the same 8 scores, already sorted: 6, 7, 7, 8, 8, 8, 9, 10.
Since n = 8 (even), the median is the average of the 4th and 5th values: Median = (8 + 8) / 2 = 8.0
Now consider a skewed dataset, annual salaries for 7 employees: $35,000, $38,000, $42,000, $45,000, $48,000, $52,000, $310,000.
Mean = $81,429 (pulled up by the $310,000 outlier) Median = $45,000 (the 4th value of 7)
The median ($45,000) better represents a typical employee's salary here. The mean ($81,429) is misleading because one executive salary skews it heavily.
Strengths: Resistant to outliers. Better for skewed distributions.
Weakness: Doesn't use all data points. Harder to use in downstream statistical formulas.
Mode (Most Frequent Value)
How to find it:
Count how often each value appears. The value with the highest frequency is the mode.
Worked Example:
A restaurant surveys 50 customers about their preferred meal time: Breakfast (8), Lunch (22), Dinner (15), Late Night (5).
Mode = Lunch (22 responses)
The most popular meal time is lunch. Mode is the only measure of central tendency that works for categorical (non-numeric) data like this.
Variations:
- No mode: all values appear equally often
- Unimodal: one clear peak
- Bimodal: two values tied for highest frequency (or nearly tied)
- Multimodal: three or more peaks
Strengths: Works with categorical data. Easy to understand.
Weakness: May not exist or may not be unique. Ignores the rest of the distribution.
When to Use Which
| Situation | Best Measure | Why |
|---|---|---|
| Symmetric, continuous data (ratings, scores) | Mean | Uses all data, supports further analysis |
| Skewed data (income, response times, prices) | Median | Not distorted by extreme values |
| Categorical data (preferences, categories) | Mode | Only option for non-numeric data |
| Ordinal scales (Likert 1-5) | Median or Mode | Mean is debated for ordinal data |
| You need to calculate confidence intervals | Mean | Required for standard CI formulas |
How Skewness Affects the Three Measures
In a perfectly symmetric distribution (like a bell curve), the mean, median, and mode are all equal. When data skews right (a long tail of high values, like income), the mean gets pulled right and exceeds the median. When data skews left (a long tail of low values), the mean drops below the median.
The relationship in right-skewed data: Mode < Median < Mean. The relationship in left-skewed data: Mean < Median < Mode.
Recognizing this pattern helps you choose the right summary statistic and explains why reported averages sometimes feel "off" from typical experience.
When to Use Mean, Median, and Mode
- Rating scale analysis: use the mean for interval-scale ratings (1-10), report the median alongside it if you suspect outliers or skew
- Income, salary, or price data: use the median because these distributions are almost always right-skewed
- Categorical survey questions: use the mode to identify the most popular option (favorite product, preferred channel, top concern)
- Comparing groups: means are best when you plan to run t-tests or ANOVA; medians are best with non-parametric tests like Mann-Whitney
- Time-series tracking: report means for trend lines but check medians to make sure outliers aren't creating false trends
Common Mistakes to Avoid
- Reporting only the mean for skewed data: always check the distribution shape before defaulting to the average
- Calculating the mean of ordinal data and treating it as precise: the mean of a 1-5 Likert scale is common in practice, but the intervals between points may not be truly equal
- Ignoring bimodal distributions: if your data has two peaks, reporting a single mean or median can hide the fact that you have two distinct groups in your sample
- Assuming the mode is always useful: in continuous data with many unique values, every value might appear only once, making the mode meaningless
- Forgetting to report spread alongside center: a mean of 7.0 with a standard deviation of 0.5 is very different from a mean of 7.0 with a standard deviation of 3.0
How Quali-Fi Supports Mean, Median, and Mode
Quali-Fi's analytics dashboard calculates the mean, median, and standard deviation for every numeric question in your survey automatically. For categorical questions, the mode is highlighted along with full frequency distributions. Cross-tabulation views let you compare these measures across segments, so you can instantly see whether the median satisfaction score for Enterprise customers differs from SMB customers without exporting a single spreadsheet.
Frequently Asked Questions
Can a dataset have more than one mode?
Yes. If two values share the highest frequency, the dataset is bimodal. If three or more values tie, it's multimodal. In survey research, bimodal distributions often indicate two distinct customer segments with different preferences.
Should I use mean or median for Likert scale data?
This is debated. Purists argue Likert data is ordinal (not interval), so the median is technically correct. In practice, means of Likert scales are widely used and generally informative as long as you don't over-interpret small decimal differences. Report both if precision matters.
What happens to the mean when I remove outliers?
The mean moves toward the bulk of the data. This is called a "trimmed mean", you remove a fixed percentage of extreme values from both ends before calculating. It's a compromise between the sensitivity of the mean and the outlier resistance of the median.
Is the average always the mean?
In everyday language, "average" usually refers to the arithmetic mean. But technically, the median and mode are also types of averages (measures of central tendency). In research writing, it's clearer to specify which measure you're referring to rather than using the word "average."
Related Topics
- Standard Deviation
- Normal Distribution
- Confidence Interval
- Margin of Error
- Likert Scale
- Levels of Measurement
Need automatic mean, median, and mode calculations for every survey question? Start your free 14-day Quali-Fi trial, no credit card required.