The objective of this study is to examine the structural relationships among variables that predict the mathematical ability of young children, namely young children's mathematical attitude, exposure to private mathematical learning, mothers' view about their children's mathematical learning, and mothers' mathematical attitude. To this end, we studied 387 pairs of mothers and their four-, five- and six-year-old children (mean: 66 months old). We conducted tests on the children and questionnaire surveys on the mothers. We used the structural equation modelling technique. In conclusion, young children's mathematical abilities can be predicted based on their own and their mothers' positive attitudes towards math and their mothers' constructivist views about mathematical learning directly and indirectly. Furthermore, exposure to private mathematical learning is not in any way related to young children's mathematical abilities when all of these variables are considered. (author abstract)