(Challenge) Given some data with corresponding weights, what would happen to the weighted average if all the weights are increased by a constant value, say 1? If required, experiment with some data. What do you observe? Justify your answer using algebra.
How Quantities Combine: Understanding Data · Question 32 of 33
(Challenge) Given some data with corresponding weights, what would happen to the weighted average if all the weights…
Experiment: values 10 and 20 with weights 1 and 3.
Increase each weight by 1, to 2 and 4:
The ordinary (unweighted) mean of 10 and 20 is 15.
Observation: unlike doubling the weights, adding a constant to all weights generally changes the weighted average. It moves towards the ordinary mean of the values (from 17.5 towards 15). Adding 1 changes the ratio of the weights from to , so the weights become more equal.
Algebra: let the values be with weights . Write and let be the weighted average, so . Let be the ordinary mean, so . With every weight increased by 1:
The new average is itself a weighted mean of the old weighted average (weight ) and the ordinary mean (weight ). So it lies between them. In the example, .
It stays the same only when (for example, when all the values are equal, or when the weights were already all equal). Adding a constant instead of 1 gives . The larger is, the closer the weighted average gets to the ordinary mean.
