In the search for factors that contributes to successful school achievement there are particularly two which seem to hold substantial explanatory power. One is student socioeconomic background which consistently has been found to influence school achievement (e.g Coleman et al., 1966; Hanushek, 1986; 1989; Adams, 1990; Härnqvist, 1992; Snow et al., 1998). However, socio-economic background may be regarded as a composite of several capital dimensions. The concept of cultural capital as described by Bourdieu (e.g. Bourdieu 2002), is a theoretical hypothesis that makes it possible to explain the unequal scholastic achievement of children originating from different social classes by relating academic success to the distribution of cultural capital between classes and class fractions. Recent international studies of the so called SES-effect has shown that this cultural capital factor is not only influential at the individual level of students, but also at the class level, where the socio-economic composition of the class has is highly correlated with achievement differences between classes (e.g Yang, 2003). The other factor that has shown to have quite a substantial influence on student achievement is teacher quality or teacher competence (e.g Darling-Hammond 2000; Gustafsson & Myrberg, 2002). However, most of this research is based on data from the US, and related to these studies are research findings that show that disadvantages students have less access to high quality teacher competence than socially favoured students (Nye, Konstantopolous och Hedges, 2004). This result thus indicate that there is double selection effects works in favour of those already favoured by their home background. Students of similar background tend to cluster within schools and/or classes at the same time as highly qualified teachers tend to seek employment in schools/classes where the social composition of the class/school holds more advantaged students. The purpose of the current study is to examine the effect of teacher competence and student cultural capital on reading achievement at individual and school level in a representative sample of grade 3 students in Sweden.

Data was selected from the IEA PIRLS 2001 (Progress in international Reading Literacy Study). IRT-scores of reading achievement (TOTRead) along with information given by the students, their teachers and parents in the questionnaires that follows the PIRLS study was used in the analyses (Gonzalez & Kennedy, 2003), along with some country specific information regarding teachers educational background (Rosén, Myrberg & Gustafsson, 2005). By applying two-level structural equation modelling technique using Mplus 5( Muthén & Muthén, 2007) under the Streams modelling environment (Gustafsson & Stahl, 2005) differences in reading achievement the overall reading achievement variation was decomposed into variation at the student level and variation at the school level. The influence student cultural capital was then estimated at both the student-level and at the class-level. Separate measurement models for student cultural capital (Ccap) and teacher competence (TeachC) was fitted to the data and the factors were specified in accordance to previous research and theory, and resulted in a models with excellent model fit. Parents highest educational level together with “Number of books at home” was used as indicators of the Ccap factor, while “Adequate teacher certificate”, “Teaching experience in grade 3” and “the occurrence of reading methodology in the teacher educational background” served well as indicators of TeachC). The measurement models were then merged into a two-level model and in a sequence of models the relations to reading achievement as well as possible interaction effects between Ccap and TeachC was investigated.

The result of the structural models showed a positive relation between Ccap and TOTRead at the student level of about .36. The result at the student level matches previous research perfectly. A the class level the relation between Ccap and TOTread were much stronger (.84) explaning about 70% of the variance of the between differences in reading achievement. In addition there were a positive relation between TeachC and TOTRead of .28, indicating that teacher competence has an explanatory power. Together these two factors explaines about 80% of the variance of between class differences in reading performance. Last but not least, no significant relation was found between Ccap and TeachC at the between level. The conclusion from this analysis is that Sweden does not suffer from such double selection effects as has been found in the USA