Skip to content
Main menu hidden.

Turnover among novice school leaders in Sweden

Research project Sweden has the highest turnover of school leaders of all countries in the OECD. This has negative impact on the development of schools in Sweden. Currently there is a knowledge gap about the factors that drive turnover, and especially about the factors that affect turnover of novice school leaders.

The purpose of the project is to investigate which factors can predict the turnover of novice school leaders in Sweden.

Head of project

Tobias Richard
Doctoral student

Project overview

Project period:

2021-09-01 2023-08-31


Umeå School of Education
Department of Sociology
Centre for Principal Development

Participating departments and units at Umeå University

Centre for Principal Development, Department of Sociology

Project description


In Sweden, school leaders change workplaces more often than in the rest of the OECD. On average, a Swedish school leader has been at his or her current school for 2-3 years. In comparison, the other countries in the OECD have an average of 5 years. High turnover of school leaders has negative consequences for the school. Research shows that it takes 5-7 years of continuous leadership to improve a school with mediocre quality. Schools with a high turnover of school leaders thus rarely get the chance to become high effective schools. This has several explanations. Firstly, with each change of school leader, schools with a high turnover of school leaders need to restart their development work. This causes the school's development of structure, culture and leadership to suffer. Second, high turnover of school leaders is related to high turnover of teachers, which in itself has a negative impact on students' knowledge development. Thirdly, it takes a lot of time and resources to work with recruiting a new school leader.

The complicated work environment for school leaders in Sweden has been known for a long time. High pressure is created on school leaders as Swedish schools are regulated at both municipal and state level and as there are several different actors within the school who have expectations and opinions about how school leaders should conduct their work. Such a context becomes particularly challenging for novice school leaders.

Due to the negative effects of school leader turnover, and the special vulnerability of novice school leaders, it is of high priority (politically, socially and scientifically) that it be investigated in order to be reduced. The purpose of the research project is therefore to investigate and map factors that can predict the turnover of novice school leaders in a Swedish context. The implication of such a survey may be that our knowledge of factors that predict school leader turnover can be used to reduce school leader turnover in Swedish schools. This in turn can be beneficial when working on improving the conditions for school leadership in Sweden.

Purpose and research questions

The purpose of the project is to investigate which factors can predict the turnover of novice school leaders in Sweden. Previous research has shown that a number of different factors are related to the turnover of school leaders. The factors from previous research, on which this study will be based, can be divided into four areas.

  1. Conditions from the principal organizer. This level addresses the conditions that the school leader feels he receives from his principal organizer.
  2. School characteristics. There are several studies that show strong connections between characteristics associated with the school and turnover of school leaders. These characteristics can be, for instance, the school's aggregated merit value (how high grades the students have), the students' socio-economic background and the proportion of newly arrived students.
  3. The work environment. The immediate environment around school leaders is important for the turnover of school leaders in an international perspective. In a Swedish context, studies have shown that approximately 25% of all active school leaders are at risk of burnout. This is probably related to whether or not a school leader chooses to stay in the workplace.
  4. Individual factors. Characteristics such as gender, age and educational background have in previous studies shown to have an effect on school leader turnover. There is some evidence that personality could also be associated with the turnover of school leaders.

Common to the four levels is that they are all completely or partially unexplored in relation to school leader turnover in a Swedish context. It must then be taken into account that Sweden has almost twice as high a turnover of school leaders as the rest of the OECD. How the above listed factors affect the turnover of novice school leaders in a Swedish context is therefore important to investigate.

The project's scientific issues are the following:

In what way and to what extent is the turnover of novice school leaders in Sweden related to:

A) the school leader's perceived conditions from the principal organizer?

B) school characteristics?

C) the school leader's work environment?

D) the individual characteristics of the school leader?

Method, procedure and analysis

The project will consist of both a quantitative part and a qualitative part. We begin by describing the quantitative part.

The quantitative part of the project will be carried out as a panel study where data will be collected once every semester for 2 years. The respondents will consist of participants in the Swedish national principal training program. A baseline questionnaire will be constructed based on the various variables to be examined. The baseline questionnaire will consist of 6 different measuring instruments. The different instruments will be woven together into a test battery. For the respondent, it will thus be experienced as one (1) measuring instrument with several different parts. As far as possible, without affecting the reliability of each individual instrument, the questions will be adapted so that the test battery as a whole is perceived as efficient and meaningful without unnecessarily many questions of an overlapping nature. The measurement opportunities after the first baseline survey will use a limited selection of instruments so as not to burden the participants unnecessarily.

For the factors related to research questions A and C, the instruments COPSOQ (Kristensen, Hannerz, Høgh & Borg, 2005), GMSI (Janson & Wernbro, 2017), Principal trust scale (Gareis & Tschannen-Moran, 2004) and The Norwegian principal self‑efficacy for instructional leadership scale (Skaalvik, 2020) are to be used. The two most recently mentioned instruments will be translated by professional translators from English to Swedish. All measuring instruments are standardized instruments for examining people's perceived conditions for work and the work environment. All measuring instruments have also proved to be particularly applicable when it comes to school leaders' perceived conditions and work environment.

For research question B, only public data from the Swedish National Agency for Education will be used. This type of data can show the school's characteristics in terms of parents level of education, distribution between boys and girls, proportion of students born abroad, proportion of students with parents born abroad, proportion of newly immigrated students, merit value, proportion of students who have met the knowledge requirements in all subjects.

For research question D, concerning individual factors in the school leader, a measuring instrument for examining personality will be used together with supplementary information (gender, age, educational background). The measuring instrument that will be used to examine personality is the Mini-IPIP Scales (Donnellan, Oswald, Baird, & Lucas, 2006) which are part of the five-factor theory or Big Five. Mini-IPIP is a standardized measuring instrument that has shown good validity and reliability.

All factors of the data collected will be handled as independent variables and related to the dependent variable “movement pattern”. That is, how can the different levels of factors predict whether the school leader changes workplace or ends his or her work as a school leader. This data will then be analyzed with multivariate methods. This means that several variables will be woven together to create an explanatory model that predicts the turnover of novice school leaders in Sweden.

The qualitative part of the survey will be carried out within the same time frame as the collection of panel data. To get a deeper insight into the school leaders' experience and what is the basis for their decision to change jobs or end their position as school leaders, about 20 school leaders will be interviewed in depth when they change workplaces. The interviews will be conducted individually with the participants who have either changed jobs or who have completed their service as school leaders and who want to stand for an interview. The interviews are conducted as semi-structured based on an interview guide created for this study specifically. The purpose of the qualitative study is to be able to examine on a deeper level the school leaders' own stories about their change of workplace. The interview will be set up based on a chronological structure where different phases of the school leader's time at the workplace will be examined starting from the recruitment phase and to the end of the held position. The interview guide contains specific questions for each "time period", which gives different angles of approach in the conversation. In the interview context, other angles will also be taken into account if the interviewee has something to say that goes beyond the framework of the interview guide.

Latest update: 2021-09-09