We find a positive relationship between math attitude and students' math scores using data obtained from PISA 2012 and a 2SLS model. Math attitude is approximated by three subjective measures: parental attitude and student instrumental motivation, which assess beliefs about math importance for the job market, and student math anxiety. The presence of one family member in a math-related career is our instrumental variable. Regardless of the proxy that is used for math attitude, an increase of one standard deviation increases the student score by at least 40 points, the equivalent of one year of schooling.