chapter  1
1 Pages

Methods and statistical Treatments

Finally, we used a statistical “multiclass analysis” enabling us this time to refl ect on this data on all the agreements without any preference and respecting the integrity of the agreement as a unique situation, combining different themes and characteristics. The principle was simple; the aim was to search out “families” of agreements, through a classifi cation method. The objective was to obtain a limited number of categories (by eliminating small and dispersed categories) with a set of relatively homogeneous agreements 2 while distinguishing these from other categories of agreements. 3 The goal was to get categories of agreements which the sociologist is able to interpret. The interpretation will be then on the number of categories and the number of agreements per category, based on themes found within these categories. 4 These themes have fi rst a relationship to the context negotiation (section of the legislation carried, type of offensive or defensive employment agreement, year of signature, signatory-union delegate, mandated employeeand branch of industry and fi rm size). These themes were to subsequently affect the negotiated rules: monthly pay and annual bonuses, evolution of