A decision-maker’s preference values – or ‘point’ or ‘utility’ values – represent the relative importance of the criteria (or attributes) to him or her.
Each preference value combines two effects with respect to the relevant criterion and level:
- the relative importance (weight) of the criterion, and
- the ‘degree of performance’ of the level on the criterion.
Relative importance (weight) of the criterion
The relative importance or weight of a criterion is represented by the value corresponding to its highest level, e.g. usually bolded in 1000minds reports.
These weights on the criteria sum across the criteria to a total of 100%, i.e. 1, which means they can be interpreted in relative terms.
e.g. If the highest level for criterion A = 22% and the highest level for criterion B = 11%, then criterion A is twice (i.e. 22/11) as important as criterion B. It can also be said that “criterion A’s importance is 22%” and “criterion B’s importance is 11%”.
A criterion’s relative importance depends on how its levels are defined, i.e. the broader and more salient the levels, the more important will be the criterion.
e.g. If you were ranking houses to buy, then a “view” criterion would be revealed (by its preference values) as being more important if its highest level were “spectacular” rather than just “OK”.
Performance on the criterion
The value for each level on a criterion represents the level’s ‘degree of performance’ on the criterion.
The lowest level corresponds to the minimum (e.g. ‘worst’) possible on the criterion, and the highest level to the maximum (e.g. ‘best’) possible. Levels between these two extremes correspond to some fractional value of the maximum possible.