Then, for each respondent, and for each factor, a factor score was derived by computing the mean across each item contributing to that factor. Finally, these factor scores were entitled Motiv 1 (social motivation), Motiv 2 (relaxation motivation), Motiv 3 (sport/competency motivation) and Motiv 4 (intellectual motivation). They were used as inputs for a cluster analysis.
Second, a cluster analysis resulted in limited senior travelers segments, which were identified and labeled (Punj and Stewart, 1983). More precisely, a hierarchical clustering on 90 randomly selected people (among the 300 individuals of the final sample) resulted in two groups. We then performed a series of non-hierarchical analyses on the whole sample (the first one involved k 1⁄4 2 groups). After these successive analyses, the solution generated with four segments, was the most satisfying one regarding the convergence criterion. This solution is presented in Table IV.