Escala de heteroavaliação de estilo de liderança de diretores escolares: construção e evidências de validade

Abstract

The role of the school principal has repercussions on the well-being of the team, especially the teaching staff. These repercussions are observed in professional development, health, and teacher turnover. In addition to these, there are also impacts on students' educational outcomes. In Brazil, the leadership of the principal is strongly associated with school management guidelines. Adding to the scenario is the theoretical question regarding the multiple definitions of leadership and the complexity of achieving a consensus on its meaning. Nationally, these mentioned factors reverberate in the scarcity of studies and measures that focus on the investigation and evaluation of educational leadership as an independent construct from school management. Furthermore, there is an absence of instruments that allow the evaluation of teachers regarding the leadership context to which they are subjected. Therefore, this study aimed to construct and present initial evidence of validity for the School Principals' Leadership Style Hetero-Assessment Scale (EHELDE). The adopted theoretical framework was the Managerial Grid, a model that establishes leadership styles based on the leader's behavior regarding a focus on people and results. The scale items were constructed through a literature review and the constitutive and operational definitions of the construct. Evidence of content validity was verified through expert judges and semantic analyses by participants from the target audience. For data collection, a questionnaire containing EHELDE, the Depression, Anxiety, and Stress Scale (DASS-21), and sociodemographic questions was used. The research data consisted of a sample of 1904 teachers from Brazilian basic education. R and Rstudio applications were used for data analysis. Exploratory (EFA) and confirmatory (CFA) factor analyses were performed. Convergent validity evidence was analyzed through a linear regression between depression, derived from DASS-21, and leadership styles. Reliability was tested using Cronbach's alpha and McDonald's omega. Additionally, a k-means clustering procedure was used to determine scale cutoff points. As a result, at the end of EFA, two factors were presented, focusing on people and focusing on results, which explained 74.8% of the total variance. In the resulting CFA model, adequate fit indices were obtained (robust RMSEA = 0.07; CI: 0.05-0.06, p < 0.001; SRMR = 0.03; robust CFI = 0.96, robust TLI = 0.96). The instrument also showed evidence of convergent validity and satisfactory internal consistency (people focus: α = 0.98, ω = 0.98; results focus: α = 0.92, ω = 0.92). In defining cutoff points, values of 5.81 for people focus and 6.4 for results focus were obtained. With the presented evidence of validity, it is concluded that the application of the instrument can be used for the promotion of public policies and educational leadership training. Similarly, for the development of new research on organizational behavior, teacher health, and, more specifically, on the impact of the principal's leadership on the school's work context.