“…Following Bollen (1989), by inspecting modification indices scholars can determine whether the data suggests alternative theoretical models. Taking into account the information reported by the LMTEST, and based on theoretical arguments, we re-specify the initial model by classifying the QM practices into two dimensions: social, soft or instrumental QM; and technical, hard or core QM (Yong & Wilkinson, 2001;Rahman, 2004;Lewis, Pun, & Lalla, 2006;Bou, Escrig, Roca, & Beltrán, 2009;Fotopoulos & Psomas, 2009;Gadenne & Sharma, 2009;Zu, 2009;Psychogios, 2010;Zairi & Alsughayir, 2011). In this final model QM was conceived as a latent construct that accounts for the correlation between the hard and soft dimensions (Figure 2), a correlation that captures a widespread idea in QM literature: that soft or hard QM issues cannot be managed in isolation because both dimensions are needed for successful QM implementation (Hackman & Wageman, 1995;Sun, 1999;Zu, 2009).…”