Abstract
Assessment and feedback category is a fundamental component for testing student learning, huge debates about assessment and feedback have been a hot topic for researchers, the goal is to advance assessment and feedback practice to improve students’ performance.Researchers conducted studies to analyse assessment and feedback category as it has gained lowest students’ satisfaction in higher education, they tried to detect category’s issues that negatively affect students’ attainment, they used focused group and traditional surveys (multiple choice and ratings questions) where they can gauge the students’ satisfaction about different aspects of the category.
Data mining methods have proven to be a powerful tool to study and analyse phenomenons in different sectors including education. Novel frameworks were developed to analyse different aspects of the learning process, most of them focused on using structured data. In the last decade the research trend has moved towards unstructured data(e.g text), the goal was to free students from the worded questions that are designed by surveys.
In this thesis a text mining framework has been developed to analyse students’ comments that related to assessment and feedback category.
Two datasets were collected from students, each of which includes textual feedback. A qualitative analysis of the two corpora revealed two reliable corpora tailored to assessment and feedback category. Sentiment analysis models indicated poor student satisfaction about learning in general and about assessment and feedback in particular. The most frequently reported assessment issues are: assessment design, clarity of the assessment, marking criteria, poor feedback, and lack of mock exams. Several algorithms were used to create models for automatic detection of the assessment issues, with the best results being obtained by the Random Forest and Support Vector Machine algorithms. To shed more light on the category, the National Student Survey data was analysed; the outcome showed that the category has the lowest student’s satisfaction at national, institutional and department levels; it has also showed the overall student satisfaction about assessment and feedback category is lower than student satisfaction regarding each aspect of assessment and feedback. The research implications include: to inform the institution decision makers of the findings, to publish the results and the extracted datasets for further research.
Keywords: Assessment and Feedback, Sentiment Analysis, Educational Data Mining, Text Mining, Word Cloud, Student’s Evaluations.
| Date of Award | 3 Apr 2023 |
|---|---|
| Original language | English |
| Awarding Institution |
|
| Supervisor | Mohamed Bader-El-Den (Supervisor), Ella Haig (Supervisor) & Alexander Gegov (Supervisor) |