In many European countries there is a shortage of teachers both in general and in specific fields. Moreover, recruiting quality teachers is internationally a widespread problem (Cooper & Alvarado, 2006). There is vast evidence that teacher competence is an important factor influencing student achievement. High quality teachers have been estimated to have as large effects on achievement as those stemming from students’ social background (Darling-Hammond, 1999; Hanushek, 2003; Wayne & Youngs, 2003; Rivkin, Hanushek & Kain, 2005). However, the meaning of the concept teacher competence is unclear. Formal teacher education is one aspect of competence. Assessed competence is another aspect that typically is measured by students´ or parents´ evaluations of teachers´ qualities. The third important aspect is manifest competence, that is, efficiency that can be linked to student achievement. It is, however, likely that these three aspects are intertwined. To sort out the relevant aspects of teacher competence and their relative value is therefore of great interest. There are indications that there has been a successive decline in the attractiveness of the teaching profession in Sweden during several decades. The formal qualifications of the entrants also have declined. (Björklund, Edin, Fredriksson & Kreuger, 2004; Börjesson, 2006). This may have caused a less stringent selection of students to the teacher education programs, with regard to academic preparation. However, no systematic analyses have been conducted of these changes over time. One main aim of the paper is to investigate changes in the academic level and profile of the students admitted to teacher education programs for cohorts born between 1948 and 1977. The analyses also focus on the characteristics of the students who completed teacher training, and who were subsequently employed as teachers. However, there is also another road to the teaching profession in Sweden, which in particular has opened up since the early 1990s when the Swedish school system was decentralized and deregulated. Then it became possible to get employment as a teacher without a teacher education, and particularly so in independent schools. Typically teachers were hired after they had been temporarily employed as teachers for one or more spells. Hence, these teachers too have been through a process of selection, but little is known who have been recruited, and on the basis of which criteria the selection has been made. It is thus of great interest to investigate the consequences of these two ways of recruiting and hiring teachers for the level and profile of the competence of the teachers of the Swedish comprehensive school, and changes over time.

Data comes from the two databases GOLD (Gothenburg Longitudinal Educational Database and ETF (Evaluation Through Follow-up). In GOLD there is information about everyone born between 1972 and 1984, and in ETF there is information about approximately 10 % samples born in 1948, 1953, 1967, 1972, and 1977. In these databases there is individual-level information on demographics (gender, date of birth, land of birth etc), labour market outcomes (education, salaries etc) and school achievement (programs, grades, etc). The study also uses data from the teacher register that covers, among other variables, teachers´ year of birth, gender, experience, type of education and occupational status. Analyses are made with multivariate techniques.

Initial analyses based on the ETF data support the hypothesis that there has been a decline in the level of previous school achievement for students with a teacher education born between 1948 and 1972. The analyses will be further elaborated to investigate recruitment to and completion of teacher education, as well as employment as a teacher, and will be extended to include the GOLD data as well.