This study examines the relative importance of di erent factors inuencing students' performance in first year university mathematics. Characteristics (motivation, actions and beliefs) of students at three universities in Sweden (n=1007) were measured twice; when entering the university and at the end of the rst year. Principal Component Analysis (PCA) and Orthogonal Projection to Least Squares (OPLS) analysis were used for the identication of constructs and analysis of the predictive power of the constructs, respectively. Four important constructs were found which we label Self-ecacy, Motivation type, Study habits and Views of mathematics.
When entering university, these four constructs predicted 14% of the variation in students' level of success at university math, as measured by their grades in university mathematics, while 37% could be predicted at the end of their rst year at the university. When comparing the four constructs, self-ecacy is the best predictor, yet, by it self, only explaining 5% of the variation when measured in the beginning of the university studies, but 21% at the end of the first year. Upper secondary grades alone predicted 17% of the variation in first year university grades. Together, students' upper secondary grades and the four constructs predicted 21% and 43% of students' university mathematics grades, when constructs pertained to upper secondary school and university, respectively.
Study habits were found to be more important for predicting university achievement for the third of the students' with the lowest upper secondary grades. Relying on the textbook and frequent interaction with peers during the university studies, at the expense of using internet-based resources for learning, contributed positively to university achievement for this same group. The association between students' view of mathematics and their university performance was low for students with intermediate and high upper secondary grades, and non-existing for students with lower grades.