The Influence of Typologies of School Leaders on Teacher Retention: A Multilevel Latent Class Analysis

• In Event: 69.056 - Roundtable Session 31
  In Roundtable Session: 69.056-8 - Turnover in Schools: New Measures, New Contexts

Abstract

While school leadership literature has searched for practices with the largest effect on outcomes, we know little about how these behaviors vary by context. Further, recent shifts to include teachers in leadership have prompted a need for purposeful distinction between teacher and principal perceptions and roles. Person-centered statistics offer a means to study differences in how teachers and principals perceive leadership by context, how these perceptions interact, and the extent to which this interaction influences teacher decisions, such as whether or not to remain at their current school. In past literature, measures of school leadership are often represented by teacher perceptions, or if other stakeholder perceptions are utilized, including principals, they are aggregated at the school level (see Leithwood and Jantzi, 2008; Robinson, Lloyd and Rowe, 2008). However, principals, who monitor a two-way relationship with teachers, or the interaction between their own perceptions and teacher perceptions, may have a greater influence on teacher outcomes and teacher retention. A principal with this awareness may influence teachers to feel as though their needs are met (Slater and Teddlie, 1992) and that they are a part of a professional community that shares values and expectations (O’Reilly et al., 1991). Teachers who feel as though they are a good fit with their principal and other teachers are more likely to stay (O’Reilly et al., 1991; Smith and Ingersoll, 2004).

This study applies a two-level latent class analysis with a cross-level interaction (Muthén and Muthén, 1998-2010) to the 1999-2000 Schools and Staffing Survey (SASS), n=35,560 teachers across n=7,310 U.S. public schools. SASS includes a comprehensive set of leadership measures, not found in other surveys, collected when U.S. schools restructured to include teachers in leadership. Multilevel latent class analysis helps to identify different types of teachers and principals in leadership, the distribution of teacher types across principal types, and to test the extent that these types predict teacher retention decisions.

Teacher and principal types were best defined by the survey items that addressed their own role, not the role of the other. This provides evidence that these measures are representative of individual roles or experiences rather than part of a collective view of the school leadership as argued in past literature when aggregates are used. The highest and lowest responding teacher types were evenly distributed across principal types. So, no matter the principal type, there appeared to be teachers who responded high and low about leadership. Examining the congruency between teacher perception of principals and the principal perception of their own leadership is an emerging domain in the leadership literature (Goldring et al., 2012; Grissom and Loeb, 2009). Teacher types who reported the lowest principal-directed leadership were more likely to leave their school regardless of principal type. Further, as expected from previous connections between shared instructional leadership behaviors and variables related to teacher retention (e.g. Hulpia, Devos and Rosseel, 2009; Skaalvik and Skaalvik, 2010), teachers who reported high leadership with principals who also viewed high leadership were less likely to move schools the following year.

Author

• Angela Urick, The University of Oklahoma