Teacher Recruitment in Sweden Over the Last Two Decades: How Has Entering Teachers’ GPA Changed Over Time?

Abstract

The question about what constitutes teaching quality is widely discussed in many countries and Sweden is no exception. Teaching quality has been linked to individual characteristics assumed to be related to student learning that are not necessarily associated with specialised training for the craft of teaching. One of these are the standards for entry to the profession. This chapter highlights teachers’ academic performances. More specifically, it explores newly recruited teachers’ grade point average over a period of over 20 years. The findings are based on register data and are analysed with descriptive statistics. The findings demonstrate how newly recruited teachers’ school grade point average (GPA) has decreased the past decades but also that some quite striking differences exists depending on teachers’ certification status. Implications of the results are discussed in relation to the possible effects on student achievement.

1 Introduction

Substantial differences in teacher effectiveness have been observed for quite some time (Darling-Hammond et al., 2005; Goldhaber & Anthony, 2007; Hanushek & Rivkin, 2012; Johansson & Myrberg, 2019; Myrberg et al., 2018; Nye et al., 2004; Rockoff, 2004). Nye et al. (2004), for example, estimated that some 10% of the variance in student achievement can be explained by the teaching quality. However, results on teacher effects are still far from conclusive and it has been claimed that teacher competence is a personal trait, little affected by education and/or that it cannot be measured by observable variables (Hanushek, 1986, 1997, 2011; Kane et al., 2008; Rivkin et al., 2005). Indeed, for some the ability to teach is a diffuse trait that cannot be predicted or particularly prepared for (e.g., Chingos & Peterson, 2011). Teaching quality has been variously defined as knowing subject matter, getting high grades or test scores, being compliant and obedient, or being enthusiastic in the classroom (Darling-Hammond, 2021). These are individual qualities assumed to be related to student learning that are not necessarily associated with specialised training for the craft of teaching. However, Darling-Hammond’s investigation of successful school-systems¹ around the world suggests that they do not operate on this belief. Quite the contrary, these school-systems believe that there is a distinct body of knowledge that every teacher can demonstrate and that teachers can learn to improve their performance. Darling-Hammond identified several characteristics that these school-systems had in common. One central aspect was the clear standards that outlined what teachers are expected to know and be able to do. These standards relate to the framework of teacher knowledge that Shulman (1986, 1987) described in his seminal works. Besides the standards there were other noteworthy aspects of the teacher educations of successful school-systems. For example, teacher education in these school-systems appealed to top-performing students, and attrition rates were low, both as regards entrance to the teacher education and to the profession as such. For example, in Finland entrance to preparation is highly competitive where only 10% of the applicants are admitted to preparation for primary teaching. Moreover, applicants must complete an examination that require them to read and interpret research on teaching.

In Sweden, it is quite a different situation. Teacher status has decreased in the past decades and teacher education no longer appeals to top-performing students. The declining teacher status has been under intense scrutiny by Swedish media and some years ago it was reported that student teachers were admitted to the teacher education with the lowest possible result on the SweSAT test (Örstadius, 2013). In fact, since the early 1990s has teacher students’ final grades from upper secondary education declined to a significant greater extent than for other comparable groups (Alatalo et al., 2021; Bertilsson, 2014). In comparison to students in other higher education programs, student teachers have increasingly lower grades (UKÄ, 2017). One further observation is that the early dropouts from teacher education are extensive compared to other higher education programs, and it is the students with the lowest grades from upper secondary school that are dominating the dropouts (UKÄ, 2017).

At the same time as the academic achievement of the applicants decreased, there have been a number of teacher education reforms intended to raise the quality of teacher education. The many reforms that aimed for improved teacher quality have emerged during an era of expanding educational accountability including measurement and surveillance of teacher classroom behavior. While the intentions have been to raise teacher quality, the status of the profession has been on the decline, and stress and decreased job satisfaction are also increasingly observed. In order to better understand the development of teacher education in Sweden a brief background will follow.

1.1 Teacher Education in Sweden

In the Swedish school-system, teacher education has been reformed several times in recent decades. A nine-year long compulsory school was implemented in 1962, which resulted in a new teacher education system. Candidates opted for one of four strands aimed at, respectively, primary school grades (grades 1–3), middle school grades (grades 4–6) or towards specific subjects in secondary school grades (grades 7–9). With only minor changes, this organization lasted for some 20 years. In 1988, a new teacher education system was introduced which allowed candidates aiming to teach in compulsory schools the choice between a strand directed to primary and middle grades 1–7 and a strand directed towards the upper grades 4–9. Although the former stage-system (1–3, 4–6, or 7–9) was formally abolished in 1988, in reality the system was retained by many municipalities. As a result, teachers are not always adequately specialized for the grades they are teaching. With the beginning of 1988, and as part of a neo-liberal turn in Swedish politics, teachers were made more exchangeable and subject knowledge got an increasingly obscure position. Then again in 2001 a new teacher education reform was launched. The teacher education then went through further changes towards an increasing flexibility of teachers. The teacher education program that was launched in 2001 aimed to create a new pedagogical teacher identity where specific and well-defined subject knowledge no longer was stressed. The education became less demanding with respect to content studies and there was less emphasis on the importance of studies preparing for teaching in specific grades. A teacher could be certified to teach grades 6–12, to mention one example. In 2011, yet another teacher education reform was implemented. The pendulum had then turned towards an increasing focus on content knowledge and more specificity with respect to grade level. For example, the flexibility of the previous teacher education system with respect to teachers’ subject combinations was abolished and more focused content areas were stressed (e.g., math-science combination). Teacher candidates were now to educate towards grades 1–3, 4–6 and 7–9 again. While it is challenging to quantify how the quality of teacher education has changed between the different teacher education systems, there is a possibility to shed light on the recruitment pattern to the teacher profession in Sweden using teachers’ own grades. This chapter aims to describe the recruitment of teachers in Sweden during the past few decades with respect to the candidates’ academic achievement. The present investigation will mainly focus on newly recruited teachers’ own school grades. The research questions are:

• 1. How has the recruitment pattern in Swedish teacher education changed during the last two decades, with respect to teachers’ average grade (GPA) levels from school?

• 2. How do certified and uncertified teachers’ school GPA level differ?

2 Data and Method

To investigate characteristics of newly recruited teachers, data from the Swedish teacher register provided by Statistics Sweden was used. In this data, the complete population of teachers in Swedish schools is present, including detailed information about, for example, their position, their teacher education, and their certification status. In addition to the teacher register data information from The Gothenburg Educational Longitudinal Database (GOLD), which includes information about all individuals born after 1971, was added. A unique component of both registers is that it is stored by personal identification number, which facilitates a link between the teacher register and the national database GOLD, which also uses the personal identification number system. GOLD comprise rich information about individuals born after 1971, for example on their scholastic achievement. Information on GPA was added to teacher register data. Since the grading system has changed several times between 1996 and 2016, as well as grades being subject to inflation, grades were equated into percentile scores. Basically, to be in the 50th percentile means to have an average GPA in Grade 9. This study relies mainly on descriptive statistics such as mean comparisons to shed light on the general trends of teachers’ grade levels over time.

3 Results

In the following the teachers’ own GPA from grade 9 will be high-lighted in order to provide a picture of the recruitment pattern to the teacher profession in Sweden. Since information on GPA only is available for teachers born 1972 and later, focus is placed on specific birth cohorts or ages in the analyses. The data is cumulative in nature and more teachers are added each year. In 1996, these were just around 1500 since most teachers were older than 24. The most common ages to enter the teacher work force is 24–28 during the time-period. Some age groups were therefore selected for further analysis. In Fig. 4.1, GPA for newly recruited teachers is presented for each year, 1996–2016. To achieve comparability, different age groups (24–26 year olds and 27–29 year olds) were included.

Fig. 4.1

GPA for newly recruited teachers

Notably, the GPA decreases over time. A newly recruited teacher in 2010 is in the 65th percentile on average while in 1998 the same age group were in the 75th percentile. The results also suggest that teachers who join the profession earlier in life (24–26) have higher GPAs. Typically, the 24–26 year olds go from an upper secondary education to teacher education while the other age groups might have joined another profession or education before starting their teacher education. It is also worth noting that GPA mainly decreases for the teachers in the age group 24–26, and only up to about year 2005. The picture that emerges suggests that prerequisites have decreased more for those who have teaching as a first career choice. However, to also investigate the GPA by birth cohorts, those born in 1972, 1977, 1982 and 1987 were selected for further scrutiny. The GPA development for these cohorts is presented in Fig. 4.2.

Fig. 4.2

GPA for newly recruited teachers in different birth cohorts

The results demonstrate quite clearly what was seen in the previous graph; that the earlier teachers enter the profession, the better GPA they had. The pattern is quite similar for all four birth cohorts, but the older cohorts had typically better GPA in ages 24–28. When newly recruited teachers are around 30 or older the GPA lie about the 55th–50th percentile. In 2016 all teachers observed here are somewhat older than the typical entry age, some are 44 (born in 1972), others are 29 (born in 1987). However, their GPA tend to be much the same, about the 50th percentile. This is thus slightly lower than was shown in Fig. 4.1, which indicated that younger teachers were about the 60th percentile in 2016.

On the whole, the results suggest that the more able students from compulsory school have not chosen teacher education to a high extent in recent years. The picture that emerges shows also that the recruitment to the teaching profession have gotten more homogeneous in recent past, at least in terms of grade levels. However, the grade levels are lower in the end of the period than they have been before. In the later years, there has been a large recruitment of uncertified teachers to the compulsory school, which may have led to decreases in the overall GPA. Therefore, an additional analysis to shed light on the grade development for certified and uncertified teachers respectively was conducted.

First the general trend for teachers’ certification status was analysed. A large share of Swedish teachers do not hold any teacher qualifications. In Fig. 4.3 below, teachers are classified into two groups: those with a certification and those without. To be certified means that teachers have a training in education. Certification does not take into account degree of specialization and a teacher might not teach in the grades of subjects (s)he holds a training for. In the analysis of certification for two samples of teachers, the population of teachers working in grades 7–9 (Secondary) as well as teachers working in Grades 1–6 (Primary) were explored.

Fig. 4.3

The share of uncertified teachers in Sweden 1996–2016 divided on primary and secondary education

Figure 4.3 shows the share of uncertified teachers in the work force in secondary school (grades 7–9) and primary school (grades 1–6) respectively. This trend fluctuates somewhat across years, the general trend being that there were a higher proportion of certified teachers in the beginning of the period. In fact, the share of uncertified teachers has doubled during the time-period. It may also be noted that, teachers in primary school are certified to higher degree than is the population of teacher in secondary school 7–9. In the beginning of the 2000s the share of uncertified teachers was high, and a likely explanation of this is the large students’ cohorts, and that many uncertified teachers were then hired. More teachers were hired in response to the larger student populations; however, many of these teachers did not have an adequate teacher education. The share of certified teachers has been shown to be especially low in private schools as compared with public schools. In the beginning of the 1990s, Sweden introduced a voucher system that made it more attractive to start new private, or independent, schools. The private schools in Sweden are tax-financed and the economic conditions are about the same as for public schools. Since the introduction of the voucher system new private schools has been introduced at an increasing rate. Much of the decision-making is delegated to school level even though more strict regulations have been formulated in recent past, for example, since 2011, a teaching license is required to assign grades. However, the teaching license have only had limited influence on the general teacher certification level. Figure 4.4 presents the share of certified teachers in public and private schools.

Fig. 4.4

Proportion of certified teachers in public and private schools

As can be seen in Fig. 4.4, the proportion of certified teachers is clearly lower in the private schools. The difference is about 20% during the first decades but decreases somewhat after the teaching license requirement in 2011

Based on the analyses above it could be concluded that there seems to be a need for certified teachers in Sweden. To hold a teacher training should naturally be considered as an advantage compared to have none. However, certified teachers’ pre-requisites in terms of own GPA-levels need not to be higher than those of uncertified teachers. It should be noted that uncertified teachers may come from other professions that typically require higher GPA for higher-education admission than is required to enrol in teacher education. To shed light on this, an analysis of the GPA levels was carried out for certified and uncertified teachers respectively. The GPA was studied for three groups of teachers in the age of 24–26 and results are presented in Fig. 4.5. Those certified first time they teach, those who never (up to 2016) become certified, and all teachers.

Fig. 4.5

GPA for certified and uncertified teachers

Notably, there is a general decline for all groups, thus mirroring the pattern previously shown. However, one may note that group with certified teachers has substantially higher GPA than the uncertified group. For teachers in the 24–26 years of age, who hold a certification when they start teaching, the grades decline over time but their grades are still clearly above the average (50th) percentile. Certified teachers have around the 65th–70th percentile in the last decade. The uncertified teacher group ends at about the 50th percentile. The findings are relevant to the discussion regarding the quality of certified and uncertified teachers.

It should also be noted that the group of uncertified teachers is unbalanced across time; the share is larger in the end of the period, indicated by the GPA drop in 2015–2016 for all teachers. A likely explanation for this is that many in the group of uncertified teachers are teacher candidates who did not yet receive their license but nevertheless been working in schools as teachers. The total GPA levels for teachers in Swedish compulsory school might also be affected by the entry age to the profession. In the beginning of the time-period, it was more common to start at the age of 24 and 25 than it was some years later in. One reason is due to the fact that the new teacher education 2001 was one semester longer for teachers preparing to teach grades 1–7.

4 Discussion

The picture that emerges from the register data is that the recruitment pattern of the teacher profession has changed during the past decades. Newly recruited teachers have an increasingly lower GPA from compulsory school. It is difficult to tell how this has affected students’ performance levels but a speculation is that it has contributed to the declines in Sweden’s results in international comparisons.

Research has demonstrated that teacher’s own schooling is important for developing both CK and PCK competencies (Kleickmann et al., 2013). Kukla-Acevedo (2009) found that only the overall GPA, not the subject specific college performance for mathematics teachers was predictive of students’ 5th Grade mathematics achievement. In Swedish research, it has been difficult to demonstrate effects on students’ school achievements. Grönqvist and Vlachos (2008) estimated that a decline in teachers’ academic ability, expressed as aptitude test scores and final grades from upper secondary school, were negative for high-performing students, while low-performing students instead were negatively affected by having a teacher with high academic ability. While positive effects of teachers’ academic ability on student achievement have been observed, international evidence is not conclusive. Harris and Sass (2011), for example, showed that elementary and middle school teachers’ college entrance exam scores did not affect teacher productivity.

While it is difficult to say how the decreasing GPA levels have affected student achievement in Sweden, there are reasons to believe that the recruitment to the teacher profession has changed character with respect to the candidates’ pre-requisites. However, this has been a gradual change for many more years than is shown in the present study. A few studies have tried to evaluate teacher knowledge for different teacher cohorts. Alatalo (2011) used a content knowledge test, teachers’ content knowledge in the Swedish language structures and basic spelling rules to examine a sample of about 300 primary-school teachers in Sweden. These teachers had substantial variation in their teacher education and years of teaching experience. The results showed that primary school teachers who qualified before 1988 (born before 1972) achieved the best test results. In another study (Frank, 2009), it was found that teachers educated before 1988 received more education in both basic and remedial reading teaching than subsequent cohorts of students, thus supporting the results of Alatalo (2011). Based on the findings of these two studies, it seems reasonable to conclude that the teachers who were educated more than 30 years ago have a more appropriate education for teaching younger pupils to read. It should also be noted that the teacher education had higher admission demands in the 1980s and that candidates likely had even higher grades than the first cohorts of the present study. However, while candidates’ pre-requisites may play a role for future performances on the job, it might also interact with the quality of the teacher education and its demands.

In the present study, it was found that teachers that were somewhat older than the typical entry age (e.g., >28 years) generally have somewhat lower grades, and the share of new teachers with higher entry age has increased during the past decades. This is not necessarily negative in the sense that these teachers come with other experiences, possibly other backgrounds and different motivation. The current study used GPA from the final grade in compulsory school; however, much life experience takes place between the ages of 16 and 30, and these experiences could contribute to teachers’ knowledge.

The present investigation could not relate student performances to the teachers’ GPA levels. A potential drawback is that teachers from the register cannot be linked directly to their students. However, teacher and student data can be aggregated to school-level and analysed at an aggregated level. The longitudinal design allows for panel analyses where students’ outcomes are measured in 3rd, 6th and 9th grade, as well as for using sophisticated multilevel models. A nice feature of the PIRLS² and TIMSS³ data is that teachers can be linked to their students; however, there is no general ability measures for the teachers in these studies. Moreover, international surveys like TEDS-m⁴ and TALIS⁵ include vast information on teachers in many countries – relating both to teacher knowledge as well as the working conditions. Both these projects are excellent in many ways but there is no link to student achievement, although successful national adaptations have been made (e.g., Baumert et al., 2010). Teacher effectiveness research is a vibrant research field and the interest in teacher quality has been intensified in the past two decades, not least with the numerous research studies accumulating. Still, however, studies including adequate controls like students’ prior achievement, or studies using longitudinal and experimental designs, are rare and should be considered in future research.