For a sample of 367 middle-school students in the USA, learning environment criteria were used to evaluate two alternative science curriculum sequences (generalised/integrated and topic-based) and to investigate the differential effectiveness of the alternative sequences for two different ethnic groups (Caucasian and Hispanic). The six scales used were Personal Relevance, Student Cohesiveness, Teacher Support, Investigation/Involvement, Task Orientation, and Enjoyment of Science. Science was enjoyed more by both Caucasian and Hispanic students experiencing the topic-based sequence (effect size of 0.74 standard deviations). In terms of task orientation, Hispanic students benefitted more from the general curriculum model, but the two curriculum sequences were equally effective for Caucasian students.