Eye tracking methodology for studying teacher learning: a review of the research

ABSTRACT

Eye tracking methodology offers insights into human attention, decision-making processes, and user behaviours. In the context of teacher learning, data generated from eye-tracking technology has the potential to provide important information about teachers’ behavioural patterns and cognitive processes that may or may not be occurring during learning experiences. This review analyses existing studies that use eye tracking methodology for studying teacher learning. The authors accessed three different databases (ERIC, PsycNet, and EBSCOHost) for their review. An iterative process of reviewing and coding the articles led to an in-depth review of 10 recently published articles. Emergent themes resulted from the in-depth review and include information processing, multimedia learning, learning tools and resources, and expert-novice knowledge construction. The reviewed articles also highlight how eye tracking can be used to determine teachers’ engagement with learning material, reading patterns, and sense-making strategies. Additional study characteristics, including context, data sources, eye tracking device and metrics were also reviewed. Results of this review shed light on how educational researchers can effectively use eye tracking methodology to investigate teacher learning at different career stages. This, in turn, can inform educational stakeholders about how to optimize learning opportunities and environments.     

The use of eye movements in the study of multimedia learning

This commentary focuses on the use of the eye-tracking methodology to study cognitive processes during multimedia learning. First, some general remarks are made about how the method is applied to investigate visual information processing, followed by a reflection on the eye movement measures employed in the studies published in this special issue. It is argued that global eye movement measures indexing attentional and encoding processes during the entire learning period should preferably be complemented with more fine-grained analyses that are either time-locked to important events taking place in an animation or that by other means provide information about the time course of learning. As nicely documented in the present set of studies, it is also of importance to complement the eye-tracking data with offline measures indexing the end product of learning. Such a complementary approach is likely to yield important new insights into the process of multimedia learning.     

Students’ Learning Activities While Studying Biological Process Diagrams

Process diagrams describe how a system functions (e.g. photosynthesis) and are an important type of representation in Biology education. In the present study, we examined students’ learning activities while studying process diagrams, related to their resulting comprehension of these diagrams. Each student completed three learning tasks. Verbal data and eye-tracking data were collected as indications of students’ learning activities. For the verbal data, we applied a fine-grained coding scheme to optimally describe students’ learning activities. For the eye-tracking data, we used fixation time and transitions between areas of interest in the process diagrams as indices of learning activities. Various learning activities while studying process diagrams were found that distinguished between more and less successful students. Results showed that between-student variance in comprehension score was highly predicted by meaning making of the process arrows (80%) and fixation time in the main area (65%). Students employed successful learning activities consistently across learning tasks. Furthermore, compared to unsuccessful students, successful students used a more coherent approach of interrelated learning activities for comprehending process diagrams.     

Eye-tracking verification of the strategy used to analyse algorithms expressed in a flowchart and pseudocode

The results of qualitative and quantitative investigations conducted with individuals who learned algorithms in school are presented in this article. In these investigations, eye-tracking technology was used to follow the process of solving algorithmic problems. The algorithmic problems were presented in two comparable variants: in a pseudocode and flowchart. The eye-tracking data resulted in both qualitative (films registering the gaze paths) and quantitative measures, which allowed the detection and interpretation of the differences in the task-solving strategies between those who found the correct answer and those that did not. The results confirmed a hypothesis that use of the formal notation characteristic of a programming language for presenting algorithms is often a practical difficulty in the process of solving even simple tasks. This study opens a new direction of research; they show that eye-tracking technology can be used to optimise the educational process of learning programming.     

Eye tracking system for enhanced learning experiences

Nowadays, we are living in a world where information is readily available and being able to provide the learner with the best suited situations and environment for his/her learning experiences is of utmost importance. In most learning environments, information is basically available in the form of written text. According to the eye-tracking technology, eye movements, scanning patterns and pupil diameter are indicators of thought and mental processing involved during visual information extraction. Hence, learners can be supported and guided throughout their learning journey by the real-time information of the precise position of gaze and of pupil diameter. A proper interpretation and consequently an efficient monitoring or supervision of learners’ eye movements by different methods of eye tracking may lead to an enhanced learning process and experience. This research gives an insight of the various existing techniques that contribute to the improvement of the learning mechanism through proposing a real time monitoring system using image processing and eye detection techniques. To portray such a robust mechanism, a multipronged eye tracking approach has been envisioned. The system is deployed with a first identification of the user’s eyes followed by the detection of the iris and pupil movements of the latter. Subsequently, the information about the eyes and pupil movements were analysed and graphs were generated that helps in determining the interest and behaviour of the user. To evaluate the accuracy of the system, some user tests and various scenarios in different application domains have been performed by computing the tracking error rate and as a result it has been noticed that these tests yield to an acceptable efficiency rate and a True Acceptance Rate (TAR) of around 75 %. Moreover, the proposed system is a low cost system and can be compatible with any computer or laptop equipped with an ordinary web camera.     

Mining students' learning patterns and performance in Web-based instruction: a cognitive style approach

Personalization has been widely used in Web-based instruction (WBI). To deliver effective personalization, there is a need to understand different preferences of each student. Cognitive style has been identified as one of the most pertinent factors that affect students' learning preferences. Therefore, it is essential to investigate how learners with different cognitive styles interact with WBI programs. This paper presents an empirical study, which examines the effects of cognitive styles on students' learning patterns and the effects of learning patterns on their learning performances. Riding's cognitive style analysis was used to identify the students' cognitive styles. Data mining, especially a clustering technique, was used to analyze the results. It was found that field independent students frequently used an alphabetical index whereas field dependent students often chose a hierarchical map. Such learning patterns also have great effects on their performance, especially for field dependent students.     

All good readers are the same,but every low-skilled reader is different: an eye-tracking study using PISA data

PISA results show that a considerable number of 15-year-old pupils after 8 to 10 years of schooling have a low level of functional reading literacy, as defined in the PISA framework. While PISA results help identify the level of reading competency, they do not reveal what might be the reasons why some students fail to solve the tasks. One way to explore the difficulties students encounter while solving PISA reading tasks is to track their eye movements during reading. The main aim of this study was to explore the similarities and differences in eye movement patterns between students with high and low scores on PISA reading tasks. A sample of 92 students took part in the pre-test, which was based on PISA items, and administered to identify groups of students with high and low PISA reading scores. Based on student pre-test results, 20 students were selected for the main, eye-tracking test—10 participants with low scores and 10 with high scores. The eye-tracking test consisted of four different released PISA reading tasks, three of them continuous and one non-continuous. The continuous items were followed by one multiple-choice question each at L1, L2, and L3 levels of difficulty. The non-continuous text was followed by three multiple-choice questions (also L1–L3).Three main findings were the following: regarding saccadic amplitudes, the reading was found to be less fluent for the low-skilled group; according to the heat maps, they had difficulty finding the relevant material; and taking into account standard deviations of eye-tracking measures, the variability was found to be larger in this group. Taken together, the findings of this study show that the PISA results differentiating low- and high-scoring groups go hand in hand with insights from more fine-grained eye-tracking measurements.     

Instructor fluency leads to higher confidence in learning,but not better learning

Students’ judgements of their own learning often exceed their knowledge on a given topic. One source of this pervasive overconfidence is fluency, the perceived ease with which information is acquired. Though effects of fluency on metacognitive judgments have been explored by manipulating relatively simple stimuli such as font style, few studies have explored the effects of fluency on more complex forms of learning encountered in educational settings, such as learning from lectures. The present study manipulated the fluency of a 31-min video-recorded lecture, and measured its effects on both perceived and actual learning. In the fluent condition, the instructor used non-verbal gestures, voice dynamics, mobility about the space, and appropriate pauses. In the disfluent condition, the same instructor read directly from notes, hunched over a podium, rarely made eye contact, used few non-verbal gestures, spoke in monotone pitch, and took irregular and awkward pauses. Though participants rated the fluent instructor significantly higher than the disfluent instructor on measures of teaching effectiveness and estimated that they had learned more of the material, actual learning between the two groups did not differ as assessed by a memory test over the lecture contents given immediately (Experiment 1) or after a 1-day delay (Experiment 2). This counterintuitive result reveals an “illusion of learning” due to fluency in lecture-based learning, a very common form of instruction.     

An examination of cognitive processing of multimedia information based on viewers' eye movements

This study utilized qualitative and quantitative designs and eye-tracking technology to understand how viewers process multimedia information. Eye movement data were collected from eight college students (non-science majors) while they were viewing web pages containing different types of text and illustrations depicting the mechanism of atmospheric pressure and the formation of sea and land winds. The results showed that participants' eyes were fixated more often on the text than on the illustrations. Breaking the instructional multimedia into small successive segments did not seem to increase the number of eye fixations on the illustrations. Participants alternated their eye fixations on related components of the illustrations while focusing on verbs or sentences representing the kinematics of the weather systems. Text seemed to provide more detailed explanations than pictures and therefore served as the main resource for understanding the content. Scan paths revealed that participants were likely to be distracted by decorative icons in the illustrations. The decorative icons also created a split-attention effect on students' cognitive processing. Eye-tracking technology was found to be a valuable tool for achieving the objectives of this study, and it provided insights into students' cognitive processes during multimedia learning. Discussion and suggestions for instructional designers and future studies are provided.     

The Influence of Different Representations on Solving Concentration Problems at Elementary School

This study investigated the students’ learning process of the concept of concentration at the elementary school level in Taiwan. The influence of different representational types on the process of proportional reasoning was also explored. The participants included nineteen third-grade and eighteen fifth-grade students. Eye-tracking technology was used in conducting the experiment. The materials were adapted from Noelting’s (1980a) “orange juice test” experiment. All problems on concentration included three stages (the intuitive, the concrete operational, and the formal operational), and each problem was displayed in iconic and symbolic representations. The data were collected through eye-tracking technology and post-test interviews. The results showed that the representational types influenced students’ solving of concentration problems. Furthermore, the data on eye movement indicated that students used different strategies or rules to solve concentration problems at the different stages of the problems with different representational types. This study is intended to contribute to the understanding of elementary school students’ problem-solving strategies and the usability of eye-tracking technology in related studies.     

网络课程的可用性测试初探

随着网络教育的蓬勃发展,如何更有效地建设网络资源成为重要而迫切的问题。阐述了网络课程设计的可用性测试方法,并将视线追踪技术引入到网络课程的可用性测试中,获取学习者客观的眼动数据,分析人在网页浏览中的视觉特性,为网络课程设计者提供人的视觉参数。目的是使网络课程的设计更符合学习者的生理与心理方面的需要,使学习者的学习效率更高。     

The integration of information in a digital,multi-modal learning environment

The aim of the reported eye-tracking study was to investigate whether learners integrate information presented on several pages within a digital learning environment and whether the underlying processes differ between a purely text-based and a multi-modal digital learning environment. Participants (N = 97) learned about the development of tornados. Picture presentation (yes vs. no) and presentation of inconsistent information between pages (yes vs. no) was varied between-subjects. Results showed that inconsistent information led to more intense gaze behavior. This indicates that learners re-activated information presented on previous pages to integrate it with the actual information, enabling them to detect the inconsistencies. Furthermore, the data pattern for text-related gaze behavior was not influenced by picture presentation, indicating that cognitive processes in purely text based or multi-modal learning environments are much the same. Comprehension outcomes did not differ between groups. The implications of these results are discussed.     

自闭症谱系障碍个体社会性注意的眼动特征及干预策略

在过去的十多年中,自闭症谱系障碍个体社会性注意的眼动研究数量激增,很多研究表明,自闭症谱系障碍个体对社会性刺激的注意表现出两个特征:对社会性刺激的注意减少;对社会性刺激的处理方式是非典型的,伴随对眼睛注意的减少和对嘴注意的增加。然而,也有相反的证据。本研究系统地回顾了从婴儿到成人期自闭症谱系障碍个体社会性注意的眼动追踪研究,分析了已有研究结果存在矛盾的可能原因。最后,根据自闭症谱系障碍个体社会性注意的眼动特征,提出了相应的干预策略。     

Verbatim, standard, or edited? Reading patterns of different captioning styles among deaf, hard of hearing, and hearing viewers

One of the most frequently recurring themes in captioning is whether captions should be edited or verbatim. The authors report on the results of an eye-tracking study of captioning for deaf and hard of hearing viewers reading different types of captions. By examining eye movement patterns when these viewers were watching clips with verbatim, standard, and edited captions, the authors tested whether the three different caption styles were read differently by the study participants (N = 40): 9 deaf, 21 hard of hearing, and 10 hearing individuals. Interesting interaction effects for the proportion of dwell time and fixation count were observed. In terms of group differences, deaf participants differed from the other two groups only in the case of verbatim captions. The results are discussed with reference to classical reading studies, audiovisual translation, and a new concept of viewing speed.     

PRIOR KNOWLEDGE AND ONLINE INQUIRY-BASED SCIENCE READING: EVIDENCE FROM EYE TRACKING

This study employed eye-tracking technology to examine how students with different levels of prior knowledge process text and data diagrams when reading a web-based scientific report. Students’ visual behaviors were tracked and recorded when they read a report demonstrating the relationship between the greenhouse effect and global climate change in 2 diagrams and 4 textual sections. Based on the pretest scores, 13 participants were categorized into high and low prior knowledge (PK) groups. Eye-tracking measures including the total reading time, total fixation duration, and total regression number on each area of interest of the 2 groups were compared. A heat map was further used to show the overall visual distribution of each group. In addition, the inter-scanning transitions between the textual and graphical information of the 2 groups were compared and further confirmed by the patterns of the scan paths. The results revealed that overall students spent more time reading the textual than the graphical information. The high PK students showed longer fixation durations and more regressions on the graphics than the low PK students. Meanwhile, the high PK students showed more inter-scanning transitions than the low PK students not only between the text and graphics but also between the 2 data diagrams. This suggests that the high PK students were more able to integrate text and graphic information and inspect scientific data which is essential for online inquiry learning. This study provides eye-tracking evidence to show that low PK students have difficulties integrating scientific diagrams with expository texts and inspecting scientific data diagrams that are commonly shown in websites. Suggestions are made for future studies and instructional design for online inquiry-based science learning.     

How to guide learners' processing of multimedia lessons with pedagogical agents

Two experiments examined the effects of a pedagogical agent's (PA's) pointing gestures, eye gaze, and eye contact on learning processes (measured by learners' eye fixations on relevant elements) and learning outcomes (as measured by retention and transfer test scores) with a multimedia lesson on neural transmission. In Experiment 1, having the PA look at and point to relevant elements as she lectured led to more eye fixations on the relevant portion of the graphic and better retention and transfer test scores. Keeping eye contact with learners tended to improve retention test scores and increased their eye fixations on relevant elements when the PA also looked at and pointed to the graphic. In Experiment 2, the PA's pointing gestures as a stand-alone feature caused better retention test scores and more fixations on relevant elements of the graphic, but eye gaze direction did not. These findings help extend the embodiment principle.     

Ways of Knowing and Patterns of Reasoning: Women and men in adult secondary education

Unlike most studies on gender and learning, which use general theoretical frameworks (and instruments) on student learning, this article investigates the Epistemological Reflection Model which has been developed from a gender perspective. This cognitive development model includes four hierarchically ordered ways of knowing, and two patterns of reasoning within each of these ways of knowing. The patterns of reasoning are assumed to be gender-related and can be described, in general terms, by using the concepts of connectedness and autonomy/separation. To extend the model's validity, the authors investigated this model by making use of interviews in one study and questionnaires in another. In the interview study, the patterns of reasoning turn out to be gender-related, especially with the patterns supposedly more often used by men. In the questionnaire study, on the other hand, the patterns of reasoning do not turn out to be gender-related. In the discussion section, these results are interpreted and the model is discussed in detail.     

Towards adaptive open learning environments: Evaluating the precision of identifying learning styles by tracking learners’ behaviours

Open learning represents a new form of online learning where courses are provided freely online for large numbers of learners. MOOCs are examples of this form of learning. The authors see an opportunity for personalising open learning environments by adapting to learners’ learning styles and providing adaptive support to meet individual learner needs and preferences. Identifying learning styles of learners in open learning environments is crucial to providing adaptive support. Learning styles refer to the manner in which learners receive and perceive information. In the literature, a number of learning style models have been proposed. The Felder and Silverman Learning Styles Model (FSLSM) has been selected as the most appropriate model for open learning. In previous studies two approaches have been used to automatically identify learning styles based on the FSLSM. These approaches are known as the data-driven method and the literature-based method. In the literature, the literature-based method has been shown to be more accurate in identifying learning styles. This method relies on tracking learners’ interactions with the provided learning objects based on a set of pre-determined patterns that help in inferring learning styles. The patterns are monitored based on pre-identified threshold values. This paper aims to apply the literature-based method to open learning environments and introduce the optimal patterns and threshold values for identifying learning styles based on the FSLSM. To achieve this aim, a study was conducted whereby a prototype that simulates the open learning environment was developed and piloted on an undergraduate IT course so that learner behaviour could be tracked and data could be collected. Next, different sets of threshold values from the literature were considered along with some updated threshold values considering the context of open learning environments, and the precision of identifying learning styles was calculated. Eighty-three students participated in the study and used the developed prototype. Precision results from different threshold values presented in the literature along with customised threshold values for this study are reported and analysed in this paper. It is shown that threshold values derived from literature and customised to suit open learning environments provide a high level of accuracy in identifying learning styles. The paper presents the first study of its kind in evaluating threshold values and precision in identifying learning styles based on the FSLSM in open learning environments. The results are promising and indicate that the proposed methodology is efficient in detecting learning styles in open learning environments and useful for developing an adaptive framework.     

Teachers’ experiences of using eye gaze-controlled computers for pupils with severe motor impairments and without speech

The purpose of this study is to explore teachers’ experiences of using eye gaze-controlled computers with pupils with severe disabilities. Technology to control a computer with eye gaze is a fast growing field and has promising implications for people with severe disabilities. This is a new assistive technology and a new learning situation for teachers. Using a reflective lifeworld approach, 11 teachers were interviewed twice. The essence of the phenomenon of teaching pupils who use an eye gaze-controlled computer is to understand what the pupil does with the computer and relate this to what the pupil wants to express through the computer. The pupils have emotions, wishes and knowledge that are trapped in their own bodies. The eye gaze computer creates opportunities to get a glimpse of these thoughts to others, and creates hope concerning the pupil’s future possibilities. The teacher’s responsibility to try to understand what is inside the pupil’s trapped body is a motivating factor to integrate the computer in everyday classroom activities. The results give directions for teaching and for implementation of eye gaze computers in the school system, and also suggest improvements that could be made to computers.     

Evaluation of change blindness in multimedia learning environment with cognitive process

ABSTRACT

This study aims to investigate the change blindness and cognitive processes with eye-tracking method in multimedia learning environment. For this purpose, a multimedia animation which had some changes was designed. The eye movements were examined during the process of detecting the changes in multimedia via eye-tracking technics. The research model was defined as controlled experiment and survey methods. Twenty-one ungraduated students participated in the experiment. Attention and perception levels of participants were determined by d2 Attention Test and Group Embedded Figures Test. Change detection numbers of participants were analyzed according to their attention level and field dependence. The appearance of a major object on the scene was the most detected change and the change on the detail object was detected less. According to findings about the attention level and change detection, the participants at high attention level were more successful at detecting change in multimedia. It was observed that field-independent individuals could detect the change much more compared to the field-dependent individuals.