Mean, median, mode, and range describe different features of data. Exam problems often ask how they change, what can be inferred from them, or how to reconstruct missing information. The definitions provide more reliable reasoning than memorized shortcuts.
Mean is a weighted balance
For observations, Equivalently, the total is Each observation contributes equally unless explicit weights or frequencies are given.
For a frequency table, the mean is the sum of value times frequency divided by total frequency. A value appearing ten times contributes ten observations, not one. For group means, group size plays the same weighting role.
The deviations from the mean sum to zero: This explains why high and low observations balance around it and why changing one value by changes the mean by when the count is fixed.
Example: combine unequal groups
A class of 12 has mean 75 and a class of 18 has mean 85. Their total scores are 900 and 1,530. The combined mean is
Averaging the means to get 80 would be valid only if the class sizes were equal. The larger class pulls the combined mean closer to 85. That directional check helps catch a weighting error.
If one score in the combined group was recorded 15 points too high, correcting it lowers the mean by giving 80.5.
Median is positional, not a balance calculation
After sorting values, an odd-sized data set has one middle observation. An even-sized set has a median equal to the average of its two middle observations. The median can therefore be a value not actually present.
The median is resistant to changing the magnitudes of extreme observations, provided their order positions do not cross the middle. The mean is sensitive to those magnitudes. Resistance is not immunity: changing the number of observations or moving a value across the middle can change the median.
For 1, 3, 5, 7, 100, the median is 5 but the mean is 23.2. The high value makes the mean much larger than most observations. Which center is useful depends on the question; total-resource calculations may still require the mean.
Example: constrain a missing value using a median
The numbers 2, 6, 9, 13, and have median 9. If were below 9, the three smallest values would include 2, 6, and , making the middle less than 9. Thus is required.
Every value of at least 9 works: there are already two values below 9 and two at or above it besides the central 9. A median condition may determine a range rather than a unique value. Do not automatically set the unknown equal to the median.
Mode and range capture limited information
Mode identifies the highest frequency, with ties possible. Range is maximum minus minimum. Two very different data sets can share both their center and range, so these summaries do not reconstruct the full distribution.
Adding the same constant c to every value adds c to mean, median, and mode, while leaving range unchanged. Multiplying all values by a positive multiplies those summaries and range by . For negative , the order reverses and range scales by
Try it yourself
Eight values have mean 14. Removing one value leaves seven values with mean 12. Find the removed value. Would the original median be determined from this information alone?
Answer: The original total is 112 and the remaining total is 84, so the removed value is 28. The median is not determined: many arrangements of the remaining values can share total 84 but have different middle values.
What to remember
Use sums and weights for means, sorted positions for medians, counts for modes, and endpoints for range. Distinguish a uniquely determined answer from a condition that allows many data sets.