Mathematics-related affect is told to predict choices concerning future studies, to correlate with performance, and to be of importance per se. Unfortunately, the affect towards mathematics is frequently reported to be low in several countries, and this contradiction cannot be solved before knowing more about its development. The objective of this study is to increase our knowledge about the timing of the affective factors getting worse, which is crucial for implementing interventions at a correct phase. We investigated a longitudinal data covering Finnish students' affect during comprehensive school years (n=3502). As a result, it was found that enjoyment of mathematics is most likely to decrease during primary school years, whereas self-efficacy is most likely to decrease during lower secondary school years. [For the complete proceedings, see ED597799.]