Learning to reason through organic reaction mechanisms is challenging for students because of the volume of reactions covered in introductory organic chemistry and the complexity of conceptual knowledge and reasoning skills required to develop meaningful understanding. However, understanding reaction mechanisms is valuable for students because they are useful for predicting and explaining reaction outcomes. To identify the features students find pertinent when explaining reaction mechanisms, we have collected students’ written descriptions of an acid-catalysed amide hydrolysis reaction. Students’ writing was produced during the implementation of Writing-to-Learn assignments in a second semester organic chemistry laboratory course. We analysed students’ written responses using an analytical framework for recognizing students’ mechanistic reasoning, originally developed with attention to the philosophy of science literature. The analysis sought to identify the presence of specific features necessary for mechanistic reasoning belonging to four broad categories: (1) describing an overview of the reaction, (2) detailing the setup conditions required for the mechanism to occur, (3) describing the changes that take place over the course of the mechanism, and (4) identifying the properties of reacting species. This work provides a qualitative description of the variety of ways in which students included these features necessary for mechanistic reasoning in their writing. We additionally analysed instances of co-occurrence for these features in students’ writing to make inferences about students’ mechanistic reasoning, defined here as the use of chemical properties to justify how electrons, atoms, and molecules are reorganized over the course of a reaction. Feature co-occurrences were quantified using the lift metric to measure the degree of their mutual dependence. The quantitative lift results provide empirical support for the hierarchical nature of students’ mechanistic descriptions and indicate the variation in students’ descriptions of mechanistic change in conjunction with appeals to chemistry concepts. This research applies a framework for identifying the features present in students’ written mechanistic descriptions, and illustrates the use of an association metric to make inferences about students’ mechanistic reasoning. The findings reveal the capacity of implementing and analysing writing to make inferences about students’ mechanistic reasoning.

This systematic review of 46 published articles investigates the constructs employed and the meanings assigned to writing in writing-to-learn assignments given to students in science courses. Using components of assignments associated with the greatest learning gains-meaning making, clear expectations, interactive writing processes, and metacognition-this review illuminates the constructs of writing that yield conceptual learning in science. In so doing, this article also provides a framework that can be used to evaluate writing-to-learn assignments in science, and it documents a new era in research on writing to learn in science by showing the increased rigor that has characterized studies in this field during the past decade.

Student misconceptions are an obstacle in science, technology, engineering, and mathematics courses and unless remediated may continue causing difficulties in learning as students advance in their studies. Writing-to-learn assignments (WTL) are characterized by their ability to promote in-depth conceptual learning by allowing students to explore their understanding of a topic. This study sought to determine whether and what types of misconceptions are elicited by WTL assignments and how the process of peer review and revision leads to remediation or propagation of misconceptions. We examined four WTL assignments in an introductory biology course in which students first wrote about content by applying it to a realistic scenario, then participated in a peer-review process before revising their work. Misconceptions were identified in all four assignments, with the greatest number pertaining to protein structure and function. Additionally, in certain contexts, students used scientific terminology incorrectly. Analysis of the drafts and peer-review comments generated six profiles by which misconceptions were addressed through the peer-review process. The prevalent mode of remediation arose through directed peer-review comments followed by correction during revision. It was also observed that additional misconceptions were elicited as students revised their writing in response to general peer-review suggestions.