Educational Psychology, Vol. 20, No. 1, 2000 A Structural Model of Attitudes Towards School Subjects, Academic Aspiration and Achievement MAHER M. ABU-HILAL, Department of Psychology, Faculty of Humanities and Social Sciences, The United Arab Emirates University, UAE ABSTRACT The aims of this study were to test the structure of attitudes to school subjects and to test the causal relationship between attitudes, level of aspiration (goal) and achievement with SEM procedures. A total of 280 high school students participated in the study. Items measuring attitude towards each subject matter clustered around one factor indicating that each school subject is unique, but related, to other subjects. These distinct, but related, factors seemed to be underlined by a higher-order construct: `attitudes to school’. The structural analysis revealed that `attitudes to school’ in¯uence achievement, but only indirectly. Level of aspiration had signi®cant direct effect on achievement and mediated between `attitudes’ and achievement. The relationship between attitude and behaviour has been of concern to researchers in education and psychology. Speci®cally, much research has been conducted in the past few decades that addresses the relationship between attitude to school and achievement behaviour. However, the causal relation between these two constructs has yet to be clearly de®ned. Also, much of the work on the relationship between attitudes towards school subjects and achievement has been done with maths and science and, to a lesser extent, with language. Many theorists and researchers have simply assumed that achievement is strongly related to attitudes towards school and school subjects. The reviews conducted on studies addressing attitudes to school subjects and achievement indicated that the relationship between the two constructs is not conclusive (see Aiken, 1970; Wilson, 1983, for more). Some researchers have found that the two constructs are signi®cantly correlated (e.g. Abu-Hilal, 1992; Ethington, 1991; Ethington & Wol¯e, 1986; Marsh, 1989; Marsh & Yeung, 1998; Scho®eld, 1982). Other researchers, in contrast, have found that the two constructs are weakly or non-correlated (e.g. Abu- Hilal & Atkinson, 1990). Methodologically speaking, several researchers have expressed concern about the 0144-3410/00/010075-10 Ó 2000 Taylor & Francis Ltd