Evaluating Sentence‐BERT‐powered learning analytics for automated assessment of students' causal diagrams.

Background: When learning causal relations, completing causal diagrams enhances students' comprehension judgements to some extent. To potentially boost this effect, advances in natural language processing (NLP) enable real‐time formative feedback based on the automated assessment of students' diagra...

Descripción completa

Detalles Bibliográficos
Publicado en:Journal of Computer Assisted Learning Vol. 40; no. 6; pp. 2667 - 2681
Autores principales: Pijeira‐Díaz, Héctor J., Subramanya, Shashank, van de Pol, Janneke, de Bruin, Anique
Formato: research tables/charts Journal Article
Publicado: Wiley-Blackwell Dec2024
Acceso en línea:Ver este registro en EBSCOhost
fields @attributes:
  recordID: 1
pdfLink:
plink: https://search.ebscohost.com/login.aspx?direct=true&db=ccm&AN=180899661&site=ehost-live
header:
  @attributes:
    shortDbName: ccm
    uiTerm: 180899661
    longDbName: CINAHL Complete
    uiTag: AN
  controlInfo:
    bkinfo:
    dissinfo:
    jinfo:
      jid:
        02664909
        6M1
      jtl: Journal of Computer Assisted Learning
      issn: 02664909
      maglogo: Y
    pubinfo:
      dt: Dec2024
      vid: 40
      iid: 6
      pid: 480
      pub: Wiley-Blackwell
      place: Malden, Massachusetts
    artinfo:
      ui:
        180899661
        176779630
        180899661
        180899661
        10.1111/jcal.12992
        180899661
      ppf: 2667
      ppct: 14
      formats:
      tig:
        atl: Evaluating Sentence‐BERT‐powered learning analytics for automated assessment of students' causal diagrams.
      aug:
        au:
          Pijeira‐Díaz, Héctor J.
          Subramanya, Shashank
          van de Pol, Janneke
          de Bruin, Anique
        affil: Department of Educational Development and Research, School of Health Professions Education, Faculty of Health, Medicine and Life Sciences (FHML), Maastricht University, Maastricht, The Netherlands
      sug:
        subj:
          Data Analytics Methods
          Support Vector Machine
          Neural Networks (Computer)
          Students, High School
          Automation
          Reliability
          Funding Source
          Human
          Semantic Analysis
          Educational Technology
          Machine Learning
          Feedback
          Workload
          Natural Language Processing
      ab: Background: When learning causal relations, completing causal diagrams enhances students' comprehension judgements to some extent. To potentially boost this effect, advances in natural language processing (NLP) enable real‐time formative feedback based on the automated assessment of students' diagrams, which can involve the correctness of both the responses and their position in the causal chain. However, the responsible adoption and effectiveness of automated diagram assessment depend on its reliability. Objectives: In this study, we compare two Dutch pre‐trained models (i.e., based on RobBERT and BERTje) in combination with two machine‐learning classifiers—Support Vector Machine (SVM) and Neural Networks (NN), in terms of different indicators of automated diagram assessment reliability. We also contrast two techniques (i.e., semantic similarity and machine learning) for estimating the correct position of a student diagram response in the causal chain. Methods: For training and evaluation of the models, we capitalize on a human‐labelled dataset containing 2900+ causal diagrams completed by 700+ secondary school students, accumulated from previous diagramming experiments. Results and Conclusions: In predicting correct responses, 86% accuracy and Cohen's κ of 0.69 were reached, with combinations using SVM being roughly three‐times faster (important for real‐time applications) than their NN counterparts. In terms of predicting the response position in the causal diagrams, 92% accuracy and 0.89 Cohen's κ were reached. Implications: Taken together, these evaluation figures equip educational designers for decision‐making on when these NLP‐powered learning analytics are warranted for automated formative feedback in causal relation learning; thereby potentially enabling real‐time feedback for learners and reducing teachers' workload. Lay Description: What is currently known about this topic?: Students' monitoring accuracy of causal relation learning is on average low.Completing causal diagrams improves monitoring accuracy to some extent.Advances in natural language processing (NLP) enable automated diagram assessment.NLP‐powered learning analytics can be used for automated formative feedback. What does this paper add?: Evaluation of the reliability of the automated diagram assessment.Performance comparison of different language technologies and techniques.The accuracy of the automated diagram assessment ranged from 84% to 86%.Human‐computer Cohen's κ surpassed that of human–human (0.89 vs. 0.84). Implications for practice/or policy: The tested technologies can be embedded into digital learning environments.Diagram assessment can be reliably (semi‐)automated.This can reduce teachers' workload and enable real‐time formative feedback.This evaluation enables testing feedback interventions in future work.
      pubtype: Academic Journal
      doctype:
        research
        tables/charts
        Journal Article
      ougenre: Article
    language: English
    refInfo:
    holdings:
      @attributes:
        islocal: N