The arithmetic mean is familiar, but equal weighting is not always a fair summary. If five small orders and one large contract are averaged by order count, each order gets the same influence even though the contract accounts for most revenue. A weighted average matches the question by giving more influence to larger quantities, longer exposure or defined importance. Always explain what each weight represents so the result can be reproduced.
Formula and definitions
Formula: Simple mean = sum of values ÷ count; weighted mean = sum(value × weight) ÷ sum(weights)
Use a simple average for equal observations
A simple average makes sense when each item should count equally, such as the average score across students when each student has equal importance. Check whether the observations are on the same scale and drawn from a comparable group. Averages can still hide outliers, spread and a skewed distribution, so pair a mean with a range or median where that information would change the decision.
Choose weights with a clear meaning
Weights may represent quantity, population size, revenue, exposure time or an explicit rubric. They should relate to why an observation deserves more influence, not to the outcome someone hopes to obtain. Normalize units and ensure weights are non-negative when the intended interpretation requires it. If weights sum to 100%, the formula can be especially easy to explain, but dividing by the total weight works even when the raw weights are quantities.
Understand the direction of distortion
A weighted average can be higher or lower than a simple average depending on which observations receive larger weights. In price analysis, higher-quantity purchases deserve more influence on average price paid. In survey work, population or sampling weights may correct known representation differences if the method is valid. Applying arbitrary weights after seeing the answer changes the metric rather than correcting it.
Document missing data and zero weights
Decide how missing observations are handled and whether zero weight means the observation is deliberately excluded. Do not treat missing as zero unless zero is the actual measured value. In a weighted average, the denominator is total included weight, so a zero total weight makes the calculation undefined. Record the excluded set and reason so another analyst can reconstruct the result.
Worked example
A buyer purchases 10 units at $8 and 90 units at $10. The simple average of the two prices is $9, giving both price quotes equal influence. The weighted average purchase price is (10 × $8 + 90 × $10) ÷ 100 = $9.80, which reflects the quantities actually purchased. If the two quotes described equal-sized stores rather than order quantities, the simple average might answer a different question more appropriately.
| Item | Calculation or value | Meaning |
|---|---|---|
| Small order | 10 units × $8 | $80 |
| Large order | 90 units × $10 | $900 |
| Simple quote average | ($8 + $10) ÷ 2 | $9.00 |
| Weighted price paid | $980 ÷ 100 units | $9.80 |
Common mistakes
Do not weight by a variable unrelated to the question. Do not use a weighted average to hide a small but important group. If data represent different populations or time windows, reconcile the definitions before summarizing them.
Next steps
Write the target question first, choose what constitutes one observation, then define the weight and denominator. Calculate both methods as a diagnostic when the weighting choice changes the outcome materially. Preserve the source values and weights instead of saving only the final average.
Frequently asked questions
Do weights need to add to 100%?
No, provided the formula divides by the total weight; percentages just make some models easier to read.
Can a weighted average equal a simple average?
Yes, when all observations have equal weights or the particular values balance to the same result.
Related tools and guides: General calculator · Percentage calculator · Compare return measures