RESUMEN
When students learn a scientific theory that conflicts with their earlier naïve theories, the newer and more correct knowledge does not always replace the older and more incorrect knowledge. Both may coexist in a learner's long-term memory. Using a new speeded reasoning task, Shtulman and Valcarcel (2012) showed that naïve theories interfere with retrieving scientific theories. Although mathematics learning is a central aim of schooling and a vital prerequisite for success in life, no study has tested whether Shtulman and Valcarcel's (2012) findings generalize to mathematical subdomains such as algebra, geometry, and probability. Additionally, it is unclear how the interference strength relates to domain-specific and domain-general competencies. We investigated these questions using the speeded reasoning task with new mathematical items in a sample of 62 university students. Solution rates and reaction times indicated interference between naïve and scientific mathematical theories. Additionally, interference strength was inversely related to mathematical achievement and unrelated to general inhibitory control. After controlling for general inhibitory control, mathematical achievement was still substantially related to interference strength. These findings indicate that interference strength reflects domain-specific achievement rather than domain-general inhibitory control.