Abstract Peer feedback has been widely used in computer-supported collaborative learning (CSCL) setting to improve students’ engagement with massive courses. Although the peer feedback process increases students’ self-regulatory practice, metacognition, and academic achievement, instructors need to go through large amounts of feedback text data which is much more time-consuming. To address this challenge, the present study proposes an automated content analysis approach to identify relevant categories in peer feedback based on traditional and sequence-based classifiers using TF-IDF and content-independent features. We use a data set from an extensive course (N = 231 students) in the setting of engineering higher education. In particular, a total of 2,444 peer feedback messages were analyzed. The CRF classification model based on the TF-IDF features achieved the best performance. The results illustrate that the ability to scale up the automatic analysis of peer feedback provides new opportunities for student-improved learning and improved teacher support in higher education at scale.

Essay writing has become one of the most common learning tasks assigned to students enrolled in various courses at different educational levels, owing to the growing demand for future professionals to effectively communicate information to an audience and develop a written product (i.e. essay). Evaluating a written product requires scorers who manually examine the existence of rhetorical categories, which is a time-consuming task. Machine Learning (ML) approaches have the potential to alleviate this challenge. As a result, several attempts have been made in the literature to automate the identification of rhetorical categories using Rhetorical Structure Theory (RST). However, RST do not provide information regarding students’ cognitive level, which motivates the use of Bloom’s Taxonomy. Therefore, in this research we propose to: i) investigate the extent to which classification of rhetorical categories can be automated based on Bloom’s taxonomy by comparing the traditional ML classifiers with the pre-trained language model BERT, ii) explore the associations between rhetorical categories and writing performance. Our results showed that BERT model outperformed the traditional ML-based classifiers with 18% better accuracy, indicating it can be used in future analytics tool. Moreover, we found a statistical difference between the associations of rhetorical categories in low-achiever, medium-achiever and high-achiever groups which implies that rhetorical categories can be predictive of writing performance.

Brazilian universities have included essay writing assignments in the entrance examination procedure to select prospective students. The essay scorers manually look for the presence of required Rhetorical Structure Theory (RST) categories and evaluate essay coherence. However, identifying RST categories is a time-consuming task. The literature reported several attempts to automate the identification of RST categories in essays with machine learning. Still, previous studies have focused on using machine learning algorithms trained on content-dependent features that can diminish classification performance, leading to over-fitting and hindering model generalisability. Therefore, this paper proposes: (i) the analysis of state-of-the-art classifiers and content-independent features to the task of RST rhetorical moves; (ii) a new approach that considers the sequence of the text to extract features – i.e. sequential content-independent features; (iii) an empirical study about the generalisability of the machine learning models and sequential content-independent features for this context; (iv) the identification of the most predictive features for automated identification of RST categories in essays written in Portuguese. The best performing classifier, XGBoost, based on sequential content-independent features, outperformed the classifiers used in the literature and are based on traditional content-dependent features. The XGBoost classifier based on sequential content-independent features also reached promising accuracy when tested for generalisability.

Teaching skills for problem solving is usually accomplished on the basis of good examples of problems and corresponding sound solutions. By studying well-constructed examples the student learns how to analyze and decompose non-elementary problems and learns how to provide well-organized solutions.The tools we demonstrate support the teacher in presenting problems in an effective way and help the student in solving them. The teacher chooses a problem and provides a solution within Access or Excel, usually reducing the original problem to a collection of simpler, logically related sub-problems. The system thoroughly analyzes the teacher's solution and provides feedback about its structure as well as many automatically generated solution hints for the student. The teacher may add his own suggestions and establishes the form and content of the problem's presentation. Essentially, the teacher can specify which aspects of his own solution should be visible to the student. In this way, the difficulties for the student to solve the task can be largely controlled.The problem is proposed as a (possibly incomplete) set of sub-problems whose mutual relations may be left partially unspecified. In the same vein, some of the suggestions may be hidden in the initial problem presentation. The student can ask for hints during his solution attempts, and receives them at the price of penalties in the final evaluation. The results that the teacher's solution produces for the different sub-problems are supplied to the student (just the results, not the solutions). This provides three main benefits. The first one is motivational: the teacher's result is a clearly visible goal to reproduce and, by simple comparison, provides immediate feedback about the correctness of the student's solution attempts. The second benefit stems from the fact that the student is allowed to face the collection of sub-problems in a more flexible way. In fact, he can exploit the teacher's results to solve a particular sub-problem, independently from the sub-problems that he has (or has not) already solved. Finally, since the teacher's hidden solutions provide results that are assumed to be reliable, if the student uses them instead of his own results, error propagation is totally prevented.The system uses the teacher's results to automatically check the correctness of the student's results by comparison, and by considering different data samples the system infers the correctness of the student's solution. Moreover, since correctness is established by comparing results, the system will accept any solution that produces the same results as those arising from the teacher's solution, regardless of how the former are obtained. Experimentation with the system at the Italian Naval Academy has given good evidence that non-elementary problems can be proposed in a working context where students are stimulated to elaborate personal comprehension and to develop original solution techniques.The engineering of the system has been funded by the AICA-CRUI project "IT4PS - Information Technology for Problem Solving".

This article presents a device which is suitable for a practical and efficient implementation of Double Adaptive Quadrature.The device includes local error estimates and attempts to detect the presence of numerical difficulties in the integrand function.If a family of rules with suitable properties is chosen, then this can be achieved without affecting the overall computational cost.Extensive numerical testing has been performed on a comprehensive set of functions showing the effectiveness of the device and its efficiency.