Rankings depend on two choices:
- the choice of indicators and 
- the choice of weights attached to them

The second choice is vital as one usually measures 'apples and oranges' and the essence of ranking is to produce final, overall scores. If, for example, the regional quality of life is a matter of socio-economic benefits and costs, the final score may well depend on the relative weight of costs and benefits. 

The two choices provide an opportunity for potential manipulation. A sensitivity analysis of alternative sets of indicators and weights may reveal such manipulations. Drewe & Rosenboom (1994) have demonstrated the sensitivity of rankings with regard to the choice of weights.

In 2006, I have attended a conference in Montreal on the copmpetitivess of cities (Drewe, 2006). In 2004, Canadian Business had ranked 'the best places in Canada to set up shop'. Montreal/Laval ended up last: number 41 out of 41 major urban centers. As a joke, I have offered my services to produce another ranking with Montreal/Laval as number one...

Issues of ranking are twofold. They do not only relate to the method applied, but also to the way they are used.