Statistics formula
Formula for a weighted average
A weighted average is a mean in which some values count for more than others. Use it when the items being combined differ in importance or size: scoring job offers against your own priorities, combining grades from modules of different lengths, or averaging pay rates across unequal hours.
Formula
weighted average = (value 1 × weight 1 + value 2 × weight 2 + ...) ÷ (weight 1 + weight 2 + ...)
- weighted average
- the combined figure, which always lies between the lowest and the highest value
- value
- each number being averaged, such as a score for one factor of an offer or an hourly rate
- weight
- how much that value counts: an importance rating, a number of hours, a credit value or a group size; never negative
- total weight
- all the weights added together, which need not come to 100
What it means
In an ordinary mean every value has one vote. A weighted average gives each value as many votes as its weight. Multiplying a value by its weight is the same as writing the value down that many times, and dividing by the total weight is the same as dividing by the number of entries in that longer list. If all the weights are equal, the result is exactly the simple mean.
Weights come from one of two places. Sometimes they are facts: hours worked at each pay rate, credits attached to each course, the number of applications in each batch. Sometimes they are judgements, as when you decide that pay matters twice as much to you as the commute. Factual weights make the average correct; judgement weights make it personal, and the answer is only as sound as the thinking behind them.
Only the proportions between weights matter. Weights of 40, 30, 20 and 10 give the same answer as 4, 3, 2 and 1 or 0.4, 0.3, 0.2 and 0.1, because the division by total weight cancels the scale. Making them add up to 100 is a convenience that lets you read each weight as a percentage share; it is not a requirement of the formula.
When scoring job offers, set the weights before you score, and score every offer on the same scale. The total then reflects your stated priorities instead of your mood on the day. Treat a narrow margin as a tie: the method organises a decision but cannot make it, and factors you left out or could not score still count. The guide to evaluating a job offer lists what to consider.
Other forms
- With weights written as shares that add up to 1
- weighted average = value 1 × share 1 + value 2 × share 2 + ...
- Turning weights into shares
- share = weight ÷ total weight
- Value needed on the last item to reach a target
- needed value = (target × total weight − sum of the other values × their weights) ÷ last weight
- Simple mean as a special case
- if every weight is 1: weighted average = sum of the values ÷ number of values
Worked example
You hold two offers and score each from 1 to 10 on four factors, weighted by how much each matters to you: pay (40), prospects for progression (30), flexibility (20) and commute (10). Offer A scores 6, 7, 9 and 9. Offer B scores 9, 8, 6 and 5.
- Total weight: 40 + 30 + 20 + 10 = 100
- Offer A products: 6 × 40 + 7 × 30 + 9 × 20 + 9 × 10 = 240 + 210 + 180 + 90 = 720
- Offer A weighted average: 720 ÷ 100 = 7.2
- Offer B products: 9 × 40 + 8 × 30 + 6 × 20 + 5 × 10 = 360 + 240 + 120 + 50 = 770
- Offer B weighted average: 770 ÷ 100 = 7.7
- Simple means for comparison: A is (6 + 7 + 9 + 9) ÷ 4 = 7.75 and B is (9 + 8 + 6 + 5) ÷ 4 = 7
Result: With your weights Offer B scores 7.7 against 7.2 for Offer A, the reverse of the simple means (7 against 7.75), because B is stronger on the two factors you care about most.
Where it is used
Your contract pays different rates for different shifts and you want one figure to compare with a salaried job. Last month you worked 96 hours at 16.00, 40 evening hours at 19.00 and 24 weekend hours at 24.00. The weights are the hours: (96 × 16 + 40 × 19 + 24 × 24) ÷ (96 + 40 + 24) = (1,536 + 760 + 576) ÷ 160 = 2,872 ÷ 160 = 17.95 an hour. The simple mean of the three rates, 19.67, would overstate it, because most of your hours are at the lowest rate. The 17.95 figure is the one to carry into the hourly to annual formula.
Common mistakes
- Dividing by the number of values instead of by the total of the weights, which gives a figure far outside the range of the values.
- Pairing values with the wrong weights after reordering one list but not the other.
- Choosing or adjusting the weights after seeing the scores, so that the calculation merely confirms the offer you already preferred.
- Scoring factors on different scales, such as pay out of 10 and commute out of 5, which quietly changes how much each one counts.
- Using a simple mean of pay rates, grades or percentages when the groups behind them are of unequal size.
Questions and answers
Do the weights have to add up to 100?
No. The formula divides by whatever the weights total, so 5, 3 and 2 work as well as 50, 30 and 20. Adding up to 100, or to 1, simply makes each weight readable as a share of the whole.
How do I choose weights for comparing job offers?
List the factors that matter to you, then share out 100 points among them before looking at any scores. If two factors feel equal, give them equal points. Asking "would I give up a point of this for a point of that?" is a quick test of whether the weights reflect your priorities.
What happens if a weight is zero?
That value drops out of the calculation completely: it adds nothing to the top or to the bottom. This is a tidy way to see how the result changes without one factor. At least one weight must be above zero, or there is nothing to divide by.
Is a grade point average a weighted average?
Usually, yes. Where courses carry different credit values, each grade is multiplied by its credits and the total is divided by the total credits. Institutions differ in the details, so check how yours calculates it before quoting a figure on a resume.