Despite the use by instructors of particulate nature of matter (PNOM) diagrams in the general chemistry classroom, misconceptions on stoichiometry continue to prevail among students tasked with conceptual problems on concepts of limiting and excess reagents, and reaction yields. This dissertation set out to explore students’ understanding of stoichiometry at the microscopic level as they solved problems that using PNOM diagrams. In particular, the study investigated how students coordinated symbolic and microscopic representations to demonstrate their knowledge of stoichiometric concepts, quantified the prevalence and explained the nature of stoichiometric misconceptions in terms of dual processing and dual coding theories, and used eye tracking to identify visual behaviors that accompanied cognitive processes students used to solve conceptual stoichiometry problems with PNOM diagrams.

Interviews with students asked to draw diagrams for specific stoichiometric situations showed dual processing systems were in play. Many students were found to have used these processing systems in a heuristic-analytic sequence. Heuristics, such as the factor-label method and the least amount misconception, were often used by students to select information for further processing in an attempt to reduce the cognitive load of the subsequent analytic stage of the solution process.

Diagrams drawn by students were used then to develop an instrument administered over a much larger sample of the general chemistry student population. The robustness of the dual processing theory was manifested by response patterns observed with large proportions of the student samples. These response patterns suggest that many students seemed to rely on heuristics to respond to a specific item for one of two diagrams given for the same chemical context, and then used a more analytic approach in dealing with the same item for the other diagram. It was also found that many students incorrectly treated items dealing with the same chemical context independently of each other instead of using a more integrative approach.

A comparison of the visual behaviors of high-performing subjects with those of low-performers revealed that high performers relied heavily on the given diagrams to obtain information. They were found to have spent more time fixating on diagrams, looked between the chemical equation and the diagram for each problem more often, and used their episodic memory more heavily to collect information early on than low performers did. Retrospective think-alouds used with eye tracking also revealed specific strategies, such as counting and balancing of atoms and molecules across both sides of a diagram, as well as comparing ratios between atoms and molecules in a diagram with those given in a balanced equation, used by students to analyze PNOM diagrams.