1. Home
  2. Guides
  3. Mean, median and weighted average: choose the right summary
Everyday math

Mean, median and weighted average: choose the right summary

Learn how outliers and unequal weights affect an average, with examples you can check by hand.

Illustration of ordered data blocks and a balance scale representing averages and weighted values.

The arithmetic mean gives every value equal weight

The arithmetic mean is the sum of the values divided by the number of values. It is useful when every observation should contribute equally. Count all intended observations, including genuine zeros, and distinguish missing data from a measured zero. Leaving out zero values, or entering a missing value as zero, can change the result substantially.

Calculator Simply’s average calculator reports an equal-weight mean, along with the count, sum and range. It does not infer whether one number represents more observations than another. Before pasting a list, check whether the entries are individual measurements or already summarized groups. Averaging group averages without their group sizes can give the wrong overall mean.

Worked exampleThe mean of 10, 20 and 30 is (10 + 20 + 30) ÷ 3 = 20. Scores of 80 and 90 with one test not yet taken average 85; entering the missing test as 0 gives 170 ÷ 3 ≈ 56.67.

The median describes the middle of an ordered list

Sort the values to find the median. With an odd number of observations, use the middle value. With an even number, take the mean of the two central values. The median is less affected by an extreme observation than the arithmetic mean, although neither summary describes the entire distribution.

The NIST/SEMATECH statistics handbook notes that for skewed data the mean is pulled in the direction of the skew, and that for data with extreme tail values the median usually gives a better estimate of the center. For skewed data such as a small sample of prices, the median and mean can tell different stories. Report which one you used rather than calling either simply “the average.”

Worked exampleFor 10, 12 and 100, the mean is about 40.67 while the median is 12. For the even-sized list 3, 7, 9 and 15, the median is (7 + 9) ÷ 2 = 8.

A worked example with one unusual value

Suppose five monthly phone bills were $40, $42, $45, $47 and $180, the last including a one-time device charge. The mean is $354 ÷ 5 = $70.80, higher than four of the five bills. The median is $45, which is closer to a typical month.

Neither number is wrong; they answer different questions. If you are budgeting for total yearly spending, the mean reflects all the money spent, including unusual months. If you want to describe a normal month, the median is more representative. A useful habit is to report both alongside the range, which here is $140.

The mode: the most common value

The mode is the value that appears most often. A list can have more than one mode, or none if every value is different. The NIST handbook describes the mode as mainly useful in a qualitative way, for example to spot peaks in a histogram.

The mode is the only one of the three that works for categories, such as the most common shoe size sold or the most frequent survey answer. For measurements with many distinct decimal values, it often says little, because few values repeat exactly.

Worked exampleFor 2, 3, 3, 5, 7, 7, 7 and 9, the mode is 7, the median is (5 + 7) ÷ 2 = 6, and the mean is 43 ÷ 8 = 5.375.

Use a weighted mean when contributions are unequal

A weighted mean multiplies each value by its weight, adds those products and divides by the sum of the weights. The weights may represent credits, units purchased or the number of observations in a group. They must represent the contribution you actually intend, not arbitrary numbers chosen to obtain a preferred answer.

A GPA calculation is a familiar credit-weighted example: university registrars such as Notre Dame’s multiply each course’s grade points by its credit hours and divide by the total credit hours. For purchases, quantities can weight the unit prices. Do not treat a weighted result as an equal-weight mean, and do not use a specialized GPA tool for an unrelated scale without checking its input restrictions; Calculator Simply’s GPA tool accepts grade points from 0 to 4.

Worked exampleBuying 2 units at $10 and 8 units at $20 gives a weighted unit price of (2 × 10 + 8 × 20) ÷ 10 = $18, not $15.

Weights given as percentages

Course grades are often weighted by category: homework, tests and a final exam each count for a set percentage. Multiply each category average by its weight as a decimal and add the results. If the weights add to 100%, there is no need to divide afterward.

If the weights do not add to 100%, for example because a category has not been graded yet, divide by the sum of the weights you used. Otherwise the result is artificially low. Check the syllabus for how your instructor handles categories that are still empty.

Worked exampleHomework 20% at 95, tests 50% at 78 and final 30% at 85: 0.2 × 95 + 0.5 × 78 + 0.3 × 85 = 19 + 39 + 25.5 = 83.5. Before the final, the weighted average of the first two is (19 + 39) ÷ 0.7 ≈ 82.86.

Combining averages from groups of different sizes

When two groups have different sizes, the combined mean is not the simple average of their means. Multiply each group’s mean by its count to recover its total, add the totals and divide by the combined count. This is the same weighted-mean method with group sizes as the weights.

Worked exampleClass A has 10 students averaging 80; class B has 30 averaging 70. The combined mean is (10 × 80 + 30 × 70) ÷ 40 = 2,900 ÷ 40 = 72.5, not 75.

Be especially careful with rates and percentages

The average of two percentages is not always the overall percentage. If the underlying denominators differ, add the underlying counts before calculating the combined rate. The same warning applies to group success rates, survey percentages and discounts applied to different starting amounts.

State the population, period, weighting and exclusions alongside any average you share. A useful summary answers a defined question, such as the typical observation or the rate across all events. The calculator can perform arithmetic, but choosing which values and weights belong in that question remains your responsibility.

Worked example1 success out of 2 and 9 out of 10 combine to 10 out of 12, or about 83.33%. Averaging 50% and 90% would incorrectly give 70%.

Sources and further reading

Try the calculators