72 relevant expertise (Liu, 2021; Spillane, 2006; Tam, 2019). This means that the person who is asked for advice may perform a leadership role (Sinnema et al., 2020; Yukl, 2002), when he/she exerts influence on someone’s knowledge and skills (Moolenaar et al., 2011). However, until now data gathering and analyses in studies on distributed leadership are largely dominated by aggregation approaches using self-perception questionnaires (D’Innocenzo et al., 2016; Hulpia et al., 2009; Joo, 2020; Liu & Werblow, 2019; Sun & Xia, 2018). These methods do not regard each individual relation but focuses on distributed leadership on team level, since the questionnaires ask team members for perceptions of their team (D’Innocenzo et al., 2016). As previously introduced, there are various reasons for combining distributed leadership theory with a social network perspective, such as that the perspective includes the informal processes, studies each team member’s perception and all relations between teachers and school principals within a school team. Therefore, in this study, we follow the growing number of scholars that call for combining the social network perspective with distributed leadership theory (Cullen-Lester & Yammarino, 2016; D’Innocenzo et al., 2016; Naumov et al., 2020; Rodway & Farley-Ripple, 2020, Sinnema et al., 2020). We empirically explore how to apply the perspective to study the collective, dynamic, and relational aspects, and in this way develop a more comprehensive picture of distributed leadership. The question arises how to measure all three core aspects of the multi-faceted concept of distributed leadership (collective, dynamic, and relational). The social network perspective includes several measures that might represent various aspects of interaction and thus leadership (see for an overview Borgatti et al., 2013; Gest & Kindermann, 2012). Previous social network studies on distributed leadership mostly included one or two social network measures (e.g., Liou et al., 2014; Mehra et al., 2006) and have been largely based on quite small samples (e.g., De Lima, 2008). In more detail, previous studies on distributed leadership mostly studied graphical sociograms, without including network measures (Mehra et al., 2006; Pitts & Spillane, 2009; Sinnema et al., 2020), or utilized merely one measure to capture one aspect of distributed leadership, mostly density (Carson et al., 2007). Prior studies that utilized two measures, mostly density and centralization (Liou et al., 2014), studied a hypothetical dataset (Mayo et al., 2003), or solely included informal leaders (De Lima, 2008) or utilized it for role identification (Apkarian & Rasmussen, 2020; Smith et al., 2018) and in smaller samples of two schools (De Lima, 2008; Liou et al., 2014; Warfield, 2009) or five schools (Apkarian & Rasmussen, 2020; Brown et al., 2020). Concepts close to distributed leadership are more often studied by a combination of network measures, such as studies on collaboration of teachers (Moolenaar et al., 2012; Sinnema et al., 2020; Smit et al., 2021), research-based practices in networks (Farley-Ripple &Yun, 2021), and leadership of formal leaders or leadership teams (Hooge et al., 2019; Liou & Daly, 2018a, 2020, 2020a;
RkJQdWJsaXNoZXIy MTk4NDMw