View in EDS

Applying temporal chunk signals analysis to measure programming competence by the transcription of Java program code

Saved in:
Bibliographic Details
Title: Applying temporal chunk signals analysis to measure programming competence by the transcription of Java program code
Authors: Albehaijan, Noorah Abdullah
Publisher Information: University of Sussex, 2022.
Publication Year: 2022
Collection: University of Sussex
Subject Terms: QA0076.6 Programming
Description: This thesis investigates the basis for a novel method of quickly and efficiently assessing programming comprehension. It investigates the feasibility of assessing learners' mental chunk structures, and their temporal chunk signals, as a way of measuring their competence. The focus is on the Java programming language. The thesis investigates the feasibility of chunk-based measures in two different simple transcription tasks: view display, where stimulus is visible at all times; and hide and show, where the stimulus is only made visible when a participant presses a special button. University computer science students and faculty are the target group. Chunking theory is utilised to define three chunking measures of competence and to anticipate how they would vary across participants with different degrees of Java competence. The measures are as follows: (1) the number of characters transcribed per view (or the number of views) of the Java program code; (2) the time spent writing between the views; and (3) the duration of pauses before writing each written character. Ninety-six participants participated in the three experiments, transcribing on graphics tablets in experimental settings, and evidence of chunking's essential role in transcription tasks was revealed. Significant relationships were discovered between the chunking measures of competence and independent measures of Java competence (Java familiarity scores and students' final test marks (for the third experiment)). The third experiment included a longitudinal post-test component spanning three months of learning, in which changes to the mean scores in characters per view, writing-times, and pauses reflected the students' amount of learning.
Document Type: Electronic Thesis or Dissertation
Language: English
Access URL: https://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.856236
Accession Number: edsble.856236
Database: British Library EThOS
Description
Abstract:This thesis investigates the basis for a novel method of quickly and efficiently assessing programming comprehension. It investigates the feasibility of assessing learners' mental chunk structures, and their temporal chunk signals, as a way of measuring their competence. The focus is on the Java programming language. The thesis investigates the feasibility of chunk-based measures in two different simple transcription tasks: view display, where stimulus is visible at all times; and hide and show, where the stimulus is only made visible when a participant presses a special button. University computer science students and faculty are the target group. Chunking theory is utilised to define three chunking measures of competence and to anticipate how they would vary across participants with different degrees of Java competence. The measures are as follows: (1) the number of characters transcribed per view (or the number of views) of the Java program code; (2) the time spent writing between the views; and (3) the duration of pauses before writing each written character. Ninety-six participants participated in the three experiments, transcribing on graphics tablets in experimental settings, and evidence of chunking's essential role in transcription tasks was revealed. Significant relationships were discovered between the chunking measures of competence and independent measures of Java competence (Java familiarity scores and students' final test marks (for the third experiment)). The third experiment included a longitudinal post-test component spanning three months of learning, in which changes to the mean scores in characters per view, writing-times, and pauses reflected the students' amount of learning.