A Rubric for the Detection of Students in Crisis

Abstract

For some students, standardized tests serve as a conduit to disclose sensitive issues of harm or distress that may otherwise go unreported. By detecting this writing, known as crisis papers, testing programs have a unique opportunity to assist in mitigating the risk of harm to these students. The use of machine learning to automatically detect such writing is necessary in the context of online tests and automated scoring. To achieve a detection system that is accurate, humans must first consistently label the data that are used to train the model. This paper argues that the existing guidelines are not sufficient for this task and proposes a three-level rubric to guide the collection of the training data. In showcasing the fundamental machine learning procedures for creating an automatic text classification system, the following evidence emerges as support of the operational use of this rubric. First, hand-scorers largely agree with one another in assigning labels to text according to the rubric. Additionally, when this labeled data are used to train a baseline classifier, the model exhibits promising performance. Recommendations are made for improving the hand-scoring training process, with the ultimate goal of quickly and accurately assisting students in crisis.

MTLHealth: A Deep Learning System for Detecting Disturbing Content in Student Essays

“…Some prior work has directly explored the problem of detecting disturbing content in student responses. Earlier research at ACT produced a disturbing content pipeline based on an ensemble of multiple non-neural machine learning methods trained on selected Reddit posts (Burkhardt et al, 2017). A study by the American Institutes for Research built a classifier for a large internal dataset of constructed response and compared the performance of several varieties of recurrent neural network architectures.…”

“…The older version only uses Reddit sources, while MTL-Health is able to make predictions on varying input domains and problem formations. (Burkhardt et al, 2017). Whereas the old system used an ensemble of non-neural machine learning models all trained on one dataset, MTLHealth builds an ensemble that incorporates information from several different datasets into its predictions.…”

“…This paper describes a study conducted at ACT with the goal of building a robust deep learning model to automatically flag disturbing content in student responses. Based in part on previous definitions, ACT considers disturbing content to include references to: (1) Attempts at or thoughts of suicide/self-harm, (2) violence or plans to harm others, (3) Physical, sexual, or emotional abuse (4), bullying or social isolation, and (5) substance abuse (Burkhardt et al, 2017) (Ormerod and Harris, 2018).…”

The use of annotations to explain labels: Comparing results from a human‐rater approach to a deep learning approach

“…These alerts are identified via humans or automated engines and routed to schools for intervention to support the student. A recently published study described a process used to develop a three-tier rubric for labeling student writing as 'normal', 'concerning', or 'alert' and illustrated that both humans and automated engines can classify crisis alerts reliably (Burkhardt et al, 2021). While the classification of responses into crisis alert categories is critically important, understanding the rationale behind the classification is similarly important.…”

Machine Learning Literacy for Measurement Professionals: A Practical Tutorial

“…Even before the pandemic, the ML revolution has already impacted most, if not all, disciplines. For example, in educational measurement, researchers and practitioners have successfully applied ML in various areas, including content validity (Anderson et al., 2020), item/test development (Rafatbakhsh et al., 2021), test security (Ferrara, 2017), marking/scoring (Ercikan & McCaffrey, 2022), and crisis prediction (Burkhardt et al., 2020), to name a few. Meanwhile, various new ML‐based educational disciplines have emerged (e.g., educational data mining, learning analytics, and computational psychometrics; von Davier et al., 2021).…”