Exploring students’ mental models of the greenhouse effect through factor and cluster analysis of drawings

In this study, we investigated high school and university students (N = 174) mental models about the greenhouse effect by adopting a combined approach based on principal component analysis and non-hierarchical cluster analysis to analyze student-generated drawings. A revised eight-item rubric was used to code the drawings. Principal component analysis identified two general mental models corresponding to the following physical mechanisms: (1) incoming radiation changes after being emitted by the Earth’s surface, and (2) incoming radiation remains unchanged through the atmosphere, partly reflected and trapped. K-means cluster analysis suggested six different patterns of drawings. Two patterns were consistent with Model 1, correctly distinguishing between incoming solar radiation and outgoing emitted radiation and representing energy flow and greenhouse gas interactions. The other four patterns referred to Model 2, often misrepresenting radiation as trapped or unchanged after reflection from Earth’s surface. Chi-square analysis showed significant differences between the two groups of students, with university students more often producing drawings aligned with scientifically adequate models. These findings highlight the need for the design of teaching sequences to explicitly address the physical mechanisms underlying the greenhouse effect in order improve scientifically informed views about climate change.