How can valid and reliable automatic formative assessment predict the acquisition of learning outcomes?

Background: Sound learning design should be based on the constructive alignment of intended learning outcomes (LOs), teaching and learning activities and formative and summative assessment. Assessment validity strongly relies on its alignment with LOs. Valid and reliable formative assessment can be...

Descripción completa

Detalles Bibliográficos
Publicado en:Journal of Computer Assisted Learning Vol. 40; no. 6; pp. 2616 - 2633
Autores principales: Divjak, Blaženka, Svetec, Barbi, Horvat, Damir
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=180899658&site=ehost-live
header:
  @attributes:
    shortDbName: ccm
    uiTerm: 180899658
    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:
        180899658
        175495159
        180899658
        180899658
        10.1111/jcal.12953
        180899658
      ppf: 2616
      ppct: 17
      formats:
      tig:
        atl: How can valid and reliable automatic formative assessment predict the acquisition of learning outcomes?
      aug:
        au:
          Divjak, Blaženka
          Svetec, Barbi
          Horvat, Damir
        affil: Faculty of Organization and Informatics, University of Zagreb, Varaždin, Croatia
      sug:
        subj:
          Outcomes of Education
          Educational Measurement
          Formative Evaluation Research
          Automation
          Student Knowledge Evaluation
          Human
          Prediction Models
          Reliability and Validity
          Random Forest
          Mathematics
          Data Quality
          Factor Analysis
          Learning Environment
          Data Collection
          ROC Curve
          Models, Educational
          Textbooks
          Motion
          Presenteeism
          Funding Source
      ab: Background: Sound learning design should be based on the constructive alignment of intended learning outcomes (LOs), teaching and learning activities and formative and summative assessment. Assessment validity strongly relies on its alignment with LOs. Valid and reliable formative assessment can be analysed as a predictor of students' academic performance, but the question is how significant its predictive power is, and what other elements can affect predictions. Objectives: Our aim was to investigate the predictive power of formative assessment for summative assessment, measuring the acquisition of LOs. Methods: We analysed formative assessment results (quizzes, homework), together with log data (video and other material use, class attendance), to determine the most influential predictors and establish a reliable predictive learning analytics model. We used the Random Forest algorithm. The model is based on the data from two university mathematical courses, delivered at different years and levels of study, incorporating 813 students in two consecutive years. Results and Conclusions: Our results show that formative assessment, together with previous summative assessment, is a stronger predictor of summative assessment results than other data on students' engagement. The study pointed to the importance of completeness and quality of data, and clear links between assessment and LOs when making predictions of student results. It suggested that predictions are less reliable for the lowest and the highest performing students. It was noted that other factors can also affect predictions, like the level of LOs, or factors not easily extracted from digital data, like the learning environment and individual students' strategies. Lay Description: What is currently known about this topic?: When educators design lessons, they aim for students to acquire intended learning outcomes.To know if students are making progress, educators use a mix of regular checks, like quizzes (formative assessment) and major exams (summative assessment).Predictive learning analytics may contribute to sound learning design (LD). What does this paper add?: In our study, we looked into whether these regular checks can predict how a student might perform on high‐stake exams.We collected data from students taking two university maths courses, including their scores on quizzes, their attendance, and their use of study materials.Using a machine learning technique called the Random Forest algorithm, we built a model to predict student exam scores.Our results indicate that formative assessment is the best predictor of the success in summative assessment. Implications for practice/or policy: We found that how a student performs on regular checks is a good sign of how they might do on the high‐stake exam and careful preparation of assessment and whole LD help student learning.Assessment, formative as well as summative, needs to be aligned with intended learning outcomes and only in this case results can inform quality assurance processes and the restructuring of LD.There are other personal factors that can influence a student's performance, which are not easy to spot just from the digital data we have.The complexity of what is being taught plays a role in predicting students' final scores.
      pubtype: Academic Journal
      doctype:
        research
        tables/charts
        Journal Article
      ougenre: Article
    language: English
    refInfo:
    holdings:
      @attributes:
        islocal: N