3.2 Analysis of Preference Data

A common request in social survey research is to ask the respondents (hereafter called evaluators) to rank order a given set of alternatives. Let E denote the set of evaluators:

E = {e1, e2, . . . , ei, . . . , em}

Consider a set of alternatives denoted by

A = {a1, a2, . . . , aj, . . ., an}

The element of ai may represent various objects, factors influencing a process, possible outcomes of a decision, different plans for achieving a goal, etc.

Preference data are individual opinions expressing preferences for a set of alternatives being written in the form of total order or partial order.

Total order: There are N alternatives; rank them in order of preference.

Partial order (Fixed): Among the N alternatives consider the five, which you consider as most important and rank them in order of preference.

Partial order (Variable): Among the N alternatives, rank in decreasing order of importance those, which you practice. Here, the number of alternatives compared is not fixed.

The basic problem in analysis of preference data is how to aggregate the individual responses and provide the summary statistics of the preference data.