Situational interest is an important motivator elicited by content and/or context rather than personal preferences (Hidi, 2000). In physical education, situational interest motivates learners to become physically active (Chen et al., 2002). However, its direct impact on learning is found insignificant (Shen et al. 2003). The impact gap is hypothesized due to curriculum incoherence in physical education where learning is often overlooked and rarely assessed. This research was designed to examine the predictability of situational interest for both in-class physical activity and learning achievement. The participants were 3rd, 4th, and 5th graders (N = 161) randomly selected from approximately 6,700 pupils in 30 elementary schools. The pupils and schools were participating in a large physical education curriculum intervention study and assigned to an experimental or a control group through a stratified randomization procedure. A health-science-based physical education curriculum was taught in the experimental schools using a constructivist approach. The control schools followed their school curriculum consistent with the national physical education standards. Both groups were represented in the sub-sample for this study. Situational interest and its sources (Novelty, Challenge, Attention Demand, Exploration, and Instant Enjoyment) were measured using the Situational Interest Scale. In-class physical activity was measured in Vector Magnitude (VM) counts using accelerometers. Achievement was determined using arithmetic difference between pre- and post-tests on concepts of heart, healthy exercise zone, and exercise benefits. Descriptive analysis showed that experimental pupils demonstrated higher achievement than control pupils (21% vs. 2% gain, p < .001). No differences were found in activity levels (p = .44) or situational interest dimensions (p > .05) except Challenge which was rated higher by the control group (p = .01). Regression analysis revealed that Instant Enjoyment was the only predictor for VM counts (R2 = .37, â = .37), indicating, that “fun” experiences motivated children to become physically active. Valid predictors for achievement included the curriculum (â = -.24, experimental = 1, control = 2), situational interest (â = .40), challenge (â = -.16), attention (â = .19), instant enjoyment (â = .23), and VM counts (â = .19). These variables accounted for 55% of achievement. The findings suggest that motivating children to learn demands more than “fun”. The content should be situationally interesting, encourage children to be attentive and active, and provide manageable challenges and instant enjoyment. The findings imply that the functions of situational interest should be emphasized in developing a coherent curriculum.Keyword(s): curriculum development, elementary education