Recent work in education focuses on how characteristics of the structure and organization of schools influence students’ academic development. The aim of these studies is to identify the direct and indirect effect of the school on a variety of factors related to the academic achievement and the psychosocial adaptation of students, after taking into account some of their individual characteristics, the influence of the social environment (peers, family) and economic status (Rutter et Maughan 2002; Teddlie et Reynolds, 2000). Research on school effects relates mainly to one of the two following levels: the school- and the classroom-effect. These two trends of research try to determine how the enrolment in a particular school or a particular classroom makes a difference in the students’ success, beyond their personal and social characteristics (Bressoux, 2006; Lee, 2000). Results on the school effect with regards to academic achievement show variations between different countries (from 1% to 20% of explained variance), presumably due to differences in the educational systems. General consensus is that classroom effects outweigh school effects (Odden, Borman, & Fermanich, 2004; Teddlie & Reynolds, 2000). An important question is whether schools and classrooms have as much influence on motivational variables as on achievement. The aim of this study was to document the school- and the classroom-effect on the motivational profile, engagement and academic achievement of students from the province of Québec (Canada).

Method

The total sample includes 3645 high school students (aged from 12 to 17 years old) nested in 175 classrooms (92 in Language Arts and 83 in Mathematics) from 24 schools. These schools are located in a variety of areas (urban, semi-urban, rural) and represented a broad range of socioeconomic levels.

Questionnaires developed by Ntamakiliro, Monnard and Gurtner (2000) were used to assess the students’ general interest (INT) (4 items, α = .86), utility of school (UTIL) (4 items, α = .71), ego orientation (EGO) (4 items, α = .88), task orientation (TASK) (3 items, α = .64), perceived competence in language arts (PCLA)(5 items, α = .91) and in mathematics (PCM) (5 items, α = .93). The quantity of efforts that the student is ready to devote in school work was assessed separately for Language Arts (EFFLA) and mathematics (EFFMA) (3 items each, α = .84 for LA and .93 for maths). Finally, school achievement was measured by students’ self-reports of their final marks in language arts (ACHLA) and mathematics (ACHM) using an ordinal scale (from “less than 35%”, “36-40%”, “40 to 45%”, etc.).

Results and discussion

The data were analyzed using hierarchical linear modeling (HLM) techniques, in which students were nested within schools and within classrooms for data analysis (Bryk, Raudenbush & Congdon, 2000). The first step in HLM involves fitting a null model (a model with no predictors or control variables) and examining the variance of the dependent measures partitioned into their within- and between- schools or classrooms levels. The proportion of the total variance attributable to individual factors or higher level factors was called the intraclass correlation (ICC) (see table 1).

Results indicated that 7.2 % or less of the variability in the motivational variables was between classrooms whereas variation between schools never accounted for more than 3.2% of the variance. Since many studies have showed that motivation declines during the high school years, it appears that differences between classrooms can be explained by variations between grade levels. Therefore, in the second step of the analysis, we included the grade level as statistical control. The results show that between 20 and 40% of the classroom effect was related to grade level, with the exception of the perception of utility, for which near by 90% of the effect was related to variation between grade levels. For academic achievement, variations between classrooms are much higher than for motivational variables. Furthermore, for this variable, less than 10% of classroom effect was due to variations between academic levels.

These results are in line with other studies in other countries showing less school effects on motivational and affective variables than on achievement (e.g. Battistich, Solomon, Kim, Watson & Schaps, 1995; Konu, Lintonen & Autio, 2002: Opdenakker & Van Damme, 2000). These results suggest that differences in motivation were more related to the individual characteristics than to schools or classrooms characteristics. This does not imply that schools or classrooms have no impact but that they presumably use similar practices or interventions.