Testing the social process model on selection through expert analysis

For a long time, academics have treated the social process side of personnel selection in a stepmotherly way. Based on a literature review we presented the Social Process model on Selection (SPS model), reflecting applicants' expectations and valuing of procedural and product characteristics of the selection encounter (Derous & De Witte, 2001). In this paper, eight characteristics of the SPS model are tested through expert analysis. Expert analysis is seen as a preliminary step in the development of a questionnaire. Data were generated through Q-sorting: Thirty experts in selection sorted 69 items into mutually exclusive piles. Additionally, experts were asked to label each pile according to the underlying meaning of the items. Classic ordinal multidimensional scaling was applied on Q-sorts, revealing a two-dimensional configuration, comprising six regions (stress=.18 and R-2=.85). Corroborative evidence on six clusters was found on the basis of additive tree modelling. To estimate the degree of consistency in labelling, the Consistency in Meaning index (CIM index) was developed. Although experts were not informed about the research goal, both the graphic solution and labelling task came close to the theoretical categories of the SPS model. Based on the results of the expert analysis, the adapted SPS model is presented