Median vs Mean in Surveys

When it comes to survey analysis, you're often faced with choosing between different statistical tools. You have a dataset filled with information, but different analysis methods can give you different results. One of the most common errors in survey research comes from a surprisingly basic choice: Mean vs.

Median. Understanding when to use each can make the difference between accurate insights and misleading conclusions.

Mean is simply another term for "average." It takes all numbers in the dataset, adds them together, and divides by the total number of entries. Median is the middle point in the data – the 50th percentile – regardless of the values above or below it. Here's a simple example with this dataset: 1, 1, 1, 1, 1, 1, 2, 2, 4.

The median is "1" since that's the middle number, while the mean (average) is 1.56. For much of survey analysis, the mean is useful. If your data falls under a normal distribution (bell curve), the mean helps remove statistical noise and gives you an overall average score for the group.

But mean is often overused. When collecting survey data, it's common to find extreme scores that can skew your results significantly.

Let's say you run a customer satisfaction survey with 9 respondents rating their satisfaction on a scale of 1 to 10. You get an average of 5.22. Since you typically retain customers with scores over 3, you feel good about this number.

But then you lose 6 of those 9 customers. You look at the raw data: 1, 3, 3, 3, 3, 5, 9, 10, 10. The median is 3 – meaning at least half your customers were unhappy.

The high scores of 9 and 10 pulled the average up, masking the fact that most customers were dissatisfied. You missed a critical warning sign.

Use Mean when: Your data follows a normal distribution (bell curve), there are no extreme outliers, and you want an overall average that accounts for all values. Use Median when: Your data has extreme scores or outliers, you're measuring income or salary data (where a few high earners can skew results), your distribution is skewed rather than symmetrical, or you want to understand what a "typical" respondent looks like. Median is particularly valuable in income research – a few millionaires can make the socio-economic status of your sample look higher than it really is.

The median gives you a much more representative picture. Good survey software will let you easily calculate both mean and median in your reports, so you can compare them and choose the metric that best represents your data.