We congratulate the newly awarded Doctor of Technological Sciences in Informatics Engineering, Vaida Masiulionytė-Dagienė, who successfully defended her doctoral dissertation, Modelling of Computational Thinking Automated Assessment, at the Faculty of Mathematics and Informatics of Vilnius University (VU MIF).
Pictured (left to right): doctoral defence committee member Prof. Dr Tomas Blažauskas, doctoral defence committee chair
Prof. Dr Julius Žilinskas, Dr Vaida Masiulionytė-Dagienė, academic supervisor Assoc. Prof. Dr Tatjana Jevsikova, doctoral defence committee members Dr Anita Juškevičienė, Acad. Prof. Habil. Dr Gintautas Dzemyda, Prof. Dr Mart Laanpere. Photo by Gintautas Tamulevičius
In her dissertation, the author examines automated methods for assessing computational thinking, based on the analysis of pupils' solutions to interactive tasks.
From Assessing Answers to Analysing Thinking
According to Dr V. Masiulionytė-Dagienė, automatically assessed tasks and tests are becoming increasingly common. Automated assessment – partial or full – is used in national pupil attainment checks, lower secondary education attainment checks, and the first part of the national school-leaving examination.
"Such systems allow large numbers of pupils to be assessed quickly, but they rarely reveal how a pupil thought or what path they took to reach their answer. The results of such checks are not always objective," says the researcher.
Analysing the solution process makes it possible to better understand where pupils encountered difficulties, what skills they lack, and what kind of support would be most beneficial.
The Journey Matters as Much as the Destination
Using data collected from pupils' interactions with tasks in the international Bebras competition, the researcher examined how behavioural patterns – click sequences, time spent on solutions, and other activity indicators – reflect different thinking and problem-solving strategies, and how these can be used to assess the solution process.
The study found that pupils' approaches to solving tasks can be grouped into meaningful categories, reflecting different logical approaches to the same problems.
The dissertation proposes a unified framework that allows these strategies to be classified automatically and provides teachers with greater insight into pupils' thinking processes. Machine learning methods — clustering and neural network-based classification – were used to identify distinct behavioural patterns.
"The study showed that pupils can arrive at the same result in different ways, and an incorrect answer does not necessarily mean that the pupil did not understand the task," says Dr V. Masiulionytė-Dagienė.
An Unexpected Finding
One of the findings that surprised the researcher concerned pupils who performed a large number of actions whilst solving a task, yet whose actions showed no discernible consistent strategy.
"It appeared that the actions were carried out rather randomly, yet a correct answer was nonetheless obtained in the end," says Dr V. Masiulionytė-Dagienė.
In the researcher's view, it would be interesting to examine such cases in greater depth to determine whether they represent instances of chance success, or whether some underlying pattern of thinking might be identified.
How Process Assessment Could Transform Learning
According to the researcher, process assessment in this study was focused on the development of computational thinking, but a similar logic applies in other fields where not only the final result matters, but also how it is achieved: "In sport, for example, it is not only the result that counts, but also the manner of training, as poor habits can over time hinder progress or increase the risk of injury."
Dr V. Masiulionytė-Dagienė notes that the primary aim of assessing the solution process is to provide teachers with richer information about pupils' thinking, and to enable them to give more targeted feedback, help pupils choose more effective methods, discuss alternative approaches, and draw attention to the most important principles.
In this way, assessment becomes not merely a check on knowledge, but also a tool for developing computational, logical, and critical thinking.
The dissertation was prepared between 2021 and 2025 at Vilnius University.
Academic supervisor: Assoc. Prof. Dr Tatjana Jevsikova (VU)
Doctoral defence committee:
Prof. Dr Julius Žilinskas – committee chair (VU)
Prof. Dr Tomas Blažauskas (KTU)
Acad. Prof. Habil. Dr Gintautas Dzemyda (VU)
Dr Anita Juškevičienė (VU)
Prof. Dr Mart Laanpere (Tallinn University, Estonia)
19 June 2026