View PDF VersionPrevious Article
 
DOI: 10.1039/C8RP00189H (Paper) Chem. Educ. Res. Pract., 2019, 20, 633-649

The impact of students’ educational background, interest in learning, formal reasoning and visualisation abilities on gas context-based exercises achievements with submicro-animations

Jerneja Pavlin , Saša A. Glažar , Miha Slapničar and Iztok Devetak *
University of Ljubljana, Faculty of Education, Kardeljeva pl. 16, 1000 Ljubljana, Slovenia. E-mail: iztok.devetak@pef.uni-lj.si

Received 23rd July 2018 , Accepted 14th May 2019

First published on 23rd May 2019

Abstract

The purpose of this paper is to explore and explain students’ achievements in solving context-based gas exercises comprising the macroscopic and submicroscopic levels of chemical concepts. The influence of specific variables, such as interest in learning, formal-reasoning abilities, and visualisation abilities, is a significant factor that should be considered when explaining students’ achievements with context-based exercises. Seventy-nine students of three age groups (12, 16, and 23) participated in the study. Questionnaires, tests, and a semi-structured interview including computer-displayed context-based exercises were used to collect data. In addition, an eye-tracker was used to determine the exact location of the participants’ points of gaze. The results show that students on average answered correctly from 40 to 79% of all questions in the context-based exercises. The context-based exercise related to air compression is indicated as being difficult for students. In students’ explanations of different levels of chemical concepts, representation difficulties are detected in all three age groups of students. Students’ achievements in solving context-based gas exercises do not depend on interest in learning chemistry and visualisation abilities. However, statistically significant differences exist in total fixation duration on the correct submicrorepresentation animation between students with different formal-reasoning abilities. The results serve as a starting point for the planning of different context-based exercises and problems comprising the chemistry triplet with 3D dynamic submicrorepresentations.

Introduction

Learning about science, technology, engineering and math (STEM education) is essential in contemporary society. It is evident that scientific and technological innovations have become increasingly more significant; for instance, we face the benefits and challenges of both globalisation and a knowledge-based economy. To be successful in this new information-based and highly technological society, students need to develop their capabilities in STEM to levels much beyond what was considered satisfactory in the past (Osborne and Dillon, 2008; Scientix, 2016).

One step to achieving this is that students have to know the world around them. Air and water are the most common substances in students’ daily lives. This is also reflected in the contents of the curricula for school science subjects. The students learn about the states of matter using the examples of water and air and this content is presented in the science curricula at different levels of Slovenian primary and lower secondary education (Bačnik et al., 2009; Planinšič et al., 2009; Bačnik et al., 2011; Balon et al., 2011; Verovnik et al., 2011). Evident also from various reviews of learning materials, the most common example of presenting these topics are water and air as a substance. The present study includes two context-based exercises related to the states of water and air.

According to this, the purpose of this paper is to explore students’ knowledge about water in a gaseous state and air on the macroscopic and submicroscopic levels and identify the influence of students’ previous education, different mental abilities, and interest on their context-based exercise achievements. Particular attention is put on the identification of attention allocation at correct animation in these tasks. For the purposes of this paper, the term ‘exercise’ is used for tasks that were presented on the screen during the research since, according to Johnstone and El-Banna (1986) and Randles and Overton (2015), tasks comprising information familiar to the students, but merely presented in a different way, are defined as exercises (and not as problems).

Theoretical background

Chemistry learning by applying the chemistry triplet

The complexity of teaching and learning chemistry can be explained by the representation of concepts at three levels: the macroscopic, the (sub)microscopic, and the symbolic levels, which might be thought of as corners of a triangle, in which no form of representation is superior to another, but each complements the other (Johnstone, 1982; Johnstone, 1991). However, this original model has been further developed, and additional aspects of chemistry learning by applying the chemistry triplet are presented extensively by Talanquer (2011) and Taber (2013). The macroscopic level of chemical concepts is illustrated by the observation of chemical phenomena. At the submicroscopic level, the interpretation of the observations by the particles’ interaction is explained. To present chemical concepts at the particulate level, submicrorepresentations (SMRs) can be used and can be presented as static or dynamic modes of representations (Devetak and Glažar, 2010). It is also important to emphasise that SMRs can be presented as one particle even if a presented molecule is more complex (lower levels of chemical education) or that SMRs illustrate the actual structure of the molecules (higher levels of chemical education). These SMRs could be translated into established symbols (symbolic level of concept representation), such as symbols of elements, formulas and equations, mathematical equations, and various graphical and schematic representations (Johnstone 2001; Levy and Wilinsky, 2009; Devetak, 2012; Taber, 2013). Johnstone (2001) explains that chemical concepts are difficult to learn and often misunderstood by students. This might arise from students’ inability to connect the three concepts’ representational levels by overloading their working memory capacity during the learning process (Johnstone and El-Banna, 1986).

Several researchers report misconceptions of chemical concepts at the submicroscopic level, including the states of matter (Bunce and Gabel, 2002; Chiu et al., 2002; Mulford and Robinson 2002; Kind, 2004; Devetak et al., 2009a, 2009b; Slapničar et al., 2017). From the literature review, it is evident that students might impute macro properties of substances to the properties of individual submicro-entities from which one can identify their developed erroneous integration of the macroscopic and submicroscopic levels of particulate matter representations (Harrison and Treagust, 2000; Nicoll, 2001; Chiu et al., 2002; Taber and García-Franco, 2010; Eilam and Gilbert, 2014).

In addition to the fact that students have difficulties with understanding SMRs, teaching about the world of particles is also difficult, because particle theory is abstract, and the use of visualisation material is necessary and complex to apply in the classroom (Johnstone, 2001; Exerciseer and Dalton, 2006; Lin et al., 2016; Cheng and Gilbert, 2017). However, SMRs can significantly help to develop an adequate understanding of chemical concepts and correct possible misconceptions. SMRs are essential elements of various chemical learning materials, and it is often presumed that teachers are able to successfully implement them in chemistry lessons (Ferk Savec et al., 2016).

Students’ age and level of education influence their ability to represent the states of water. Older students less often use macro representations for explaining particulate 3D animations of states of water (Slapničar et al., 2017). The common misconception identified in groups of students of different ages is also that particles of water change size, depending on the state of matter or temperature. However, it was reported that 39% of the 14 year olds specified increasing the particle size, starting from the solid state through the liquid state, and up to the gaseous state of matter. Another misconception reported by the studies of Harrison and Treagust (2000), Frackson et al. (2014) and Slapničar et al. (2017) is that while the matter cools the particles are reduced in size. Results of research also showed that students think that when matter is warmed or cooled the particles of which the matter is composed are heated or cooled. The study of Tóth and Kiss (2006) reports how students aged 13 to 17 identify the distribution of particles in different states of matter. They were most successful (71% of correct explanations) in identifying the distribution of particles in the solid state of matter. The smallest percentage of the students (58%) chose the correct schematic of the distribution of particles in the liquid state. Correia et al. (2018) stress the importance of visualisation in gas behaviour on the submicroscopic level. Yuriev et al. (2017) and Slapničar et al. (2017) highlighted that students have to be skilled in problem solving and in observing two or more variables at the same time in 3D dynamic SMRs. The animation used in the study of Slapničar et al. (2017) presenting compressing the air in the air pump included the purposely-changed parameters: (a) reduction/increase in the number of particles per area unit, (b) reduction/increase in the speed of particle movement, or (c) reduction in the volume of the individual particle of the substance. It tests the higher cognitive levels of the students’ understanding. This context-based exercise is also used for the purposes of this paper.

Visualisation of chemical concepts

The use of molecular representations plays an essential role in chemistry education because it can enable students to visualise phenomena that cannot be directly observed due to the size of the entities in these processes (e.g., atoms, ions, molecules, and sub-atomic particles) (Phillips et al., 2010). For the purpose of this paper, visualisation abilities, as suggested by Gilbert (2004), following the literature analysis by Barnea (2000), are described as the abilities of students to recognise and manipulate visual objects.

Wu and Shah (2004) debated the role of visualisation in chemistry learning and suggested that one of the key characteristics of visualisation tools should be to provide multiple representations (e.g., static 2D or 3D SMRs, computer animations of particles, physical models, etc.). These representations should be implemented in a way that enables students to visualise the connections between representations and relevant concepts. The visual representations in science ought to make related referential connections visible, present the dynamic and interactive nature of chemistry and promote the transformation between 2D and 3D thinking, without additional cognitive loads during the process of learning by making information clear. It was reported that students expressed difficulties in constructing relevant information from the dynamic visual representations because the processes were presented too quickly. Thus, the degree of familiarity students had with the interactive controls seemed to introduce an extra cognitive load, and they seemed to lack the spatial ability to perceive the visualisations completely (Gilbert 2005; Gilbert 2008; Phillips et al., 2010).

Visualisation abilities are positively related to achievements in STEM, but this relationship is influenced by exercise demands and learner strategies (Hinze et al., 2013). However, Raiyn and Rayan (2015) report that a positive correlation exists between students’ visual-spatial abilities and their performance in problem-solving, and good 3D visualisation tools enhance students’ understanding of molecular structures. Students who were exposed to SMRs during the educational process more adequately understand the nature of particle interactions compared to those who learned the same concepts only by reading textbooks (Tien et al., 2007; Kelly and Jones, 2008). Animations increased students’ conceptual understanding by helping students create dynamic mental models of particulate phenomena (Williamson and Abraham, 1995). Russell et al. (1997) showed that when SMRs are used students seem to better correlate the symbolic, macroscopic and submicroscopic levels of representation as their conceptual understanding.

The influence of students’ formal reasoning abilities and interest on their achievements in chemistry

To correctly understand 3D animated SMRs in learning chemistry and solve context-based exercises, students have to achieve the formal operational stage, which is usually developed around the age of 12 (Devetak and Glažar, 2010). Prior to that age, students are in the concrete operational stage. The fact that some students’ intellectual development and abilities never reach the formal operational stage must be taken into account (Labinowicz, 1989). Devetak and Glažar (2010) showed that there are statistically significant correlations between formal reasoning abilities and students’ chemical knowledge on a submicroscopic level. Moreover, those students who are formal reasoners are more successful in solving problems. On average, 28% of the students’ achievements variance on items that demand reading 2D SMRs can be explained by the TOLT (test of logic thinking) score (Devetak and Glažar, 2010). Similar results were also reported by Valanides (1996) identifying that formal operation scores correlate significantly with performance in chemistry and Lewis and Lewis (2007) reporting that TOLT might predict chemistry exam success, based on a large sample study.

Any learning of abstract chemical concepts, even if the students have adequately developed formal reasoning abilities, will be fairly minimal, if a student can make little sense of a lesson, and has no interest in paying attention (Taber, 2014). Stipek (1998) claims that highly intrinsically motivated students are more successful in learning new concepts and show a better understanding of the specific content. Juriševič et al. (2008) studied the intrinsic motivation of pre-service primary school teachers for learning chemistry in relation to their achievements. Their finding was that students are more or less equally motivated for chemistry as for any other subject. However, the intrinsic motivation plummets as the level of abstraction in individual subjects, such as chemistry and mathematics, increases. They also showed that the correlation between the level of motivation and the achievements scores is moderate but statistically significant. That motivation can significantly influence students’ achievement scores in science was also revealed in the studies of Patrick et al. (2007) and Cavas (2011).

Eye-tracker method in science education

Students’ motivation for learning chemistry, formal reasoning abilities, and visualisation abilities can influence students’ learning processes. To understand these processes, individuals’ eye movements can be measured and after careful consideration might be used for interpreting processes during solving the tasks, since the direction of the human gaze is closely linked to the focus of attention as individuals process the visual information that is being observed (Just and Carpenter, 1980; Rayner, 2009; Hyönä 2010).

Several studies have been conducted using eye-tracking technology in different fields, for example, to examine how students process text, data diagrams, relevant photographs, explanatory keys, SMRs etc. (Slykhuis et al., 2005; Mason et al., 2013; Havanki and VandenPlas, 2014; Ho et al., 2014; Ferk Savec et al., 2016; Yen and Yang, 2016; Torkar et al., 2018). More concretely, Havanki and VandenPlas (2014) present eye tracking as a tool that has been recently applied in chemistry education research. While Yen and Yang (2016) introduce problems in science education that can be explored with eye-tracking techniques, and the methods that should be well thought out while conducting research and interpretation of data. Their studies investigated how participants’ pre-knowledge and extra indications in the material lead attention allocation because the material with science contents usually consists of multiple representations (e.g., text, illustrations). Slykhuis et al. (2005) studied the ability to determine how studied pre-service science teachers attend to science-related photographs. Therefore pre-service science teachers were shown a PowerPoint presentation that contained photographs: complimentary, most highly integrated with the text, and decorative, and the least integrated with the text. They report that students devoted more attention to the highly relevant photographs. The study investigating the online process of reading and the offline learning from an illustrated science text was conducted by Mason et al. (2013); they examined whether the use of a concrete or abstract picture to illustrate a text is effective. The adoption of eye-tracking methodology to trace text and picture processing was done. Results of the study show that the abstract illustration promoted more efficient processing of the text. The processing of text and data was also studied by Ho et al. (2014). Their 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. Results of the study show overall that students spent more time reading the textual than the graphical information. Students’ use of an explanatory key while solving chemistry tasks based on submicroscopic representations was the research problem of the study by Ferk Savec et al. (2016). Different explanatory keys were used in the study; coloured and black-and-white, and pictorial and textual. Eye-fixation patterns and students’ verbal explanations indicated that the presence of colour in the key does not influence students’ task solving. In contrast, students spent more time and fixated more frequently on the key while solving tasks using a textual key in comparison to a pictorial key.

A similar study was conducted by Torkar et al. (2018) on the understanding of the relation between external representations at the macroscopic, microscopic, and submicroscopic levels on the example of water balance in plants. In their study, the eye-tracking results show that students with correct answers spent less time observing the biological phenomena displayed at the macroscopic and submicroscopic levels than those with incorrect answers did.

Aims and research questions

The present research aims to investigate the understanding of Slovene students, at different levels of education (lower secondary school, upper secondary school, and undergraduate education level), of states of matter presented with SMRs. Firstly, we focused on the descriptions that students used for interpretation of the SMRs of gaseous states of matter regarding Johnstone's model (1991). The context-based exercises used in this study comprise only the macroscopic and submicroscopic levels of concept representation; the symbolic level was not used. Secondly, the focus is also on exploring the differences between specific groups of students such as: (1) the students’ educational level; (2) the success in selecting the correct 3D submicroscopic animation; and (3) the level of motivation, formal reasoning and visualisation abilities in total fixation duration (TFD) on the whole screen image (SI) and on the specific areas of interest (AOI) during solving context-based exercises.

With regard to the research aims, the following research questions can be addressed:

RQ1: Does the educational level impact achievement on two context-based exercises?

RQ2: Does the educational level impact how individuals allocate attention while solving two context-based exercises?

RQ3: Are the differences between students who correctly or incorrectly choose the SMR in the context-based exercise significant with regard to their levels of formal-reasoning abilities, visualisation abilities, motivation, and attention allocation in the context-based exercises?

RQ4: Are there differences in visual attention on the AOI representing the correct animation in the context-based exercises between students with different levels of formal-reasoning abilities, visualisation abilities, and motivation?

Method

A quantitative research approach with descriptive and non-experimental methods was used in this research.

Participants

Seventy-nine students of three different age groups (primary education, upper secondary education, and university education levels) participated in this research, performed in Slovene, their native language. The 1st group consisted of 30 12.0 years-olds (IQR = 0.0 year); 29 students aged 16.0 years (IQR = 1.0 year) were in the 2nd group; the 3rd group had 20 pre-service two subject teachers (chemistry and biology) aged 23.0 (IQR = 2.0 years). Students were from the Ljubljana region and voluntarily participated in the study. Consent was obtained for primary and upper secondary students from school boards, teachers, and parents, in accordance with the judgment of the Ethics Committee for Pedagogy Research of the Faculty of Education, University of Ljubljana. The students were selected from a mixed urban population. All participants had normal or corrected-to-normal vision, and all were competent readers. To ensure anonymity, each student was assigned a code.

Instruments

Different instruments were used to gather data to answer the research questions, such as context-based gas exercises, a test of logical thinking, science motivation questionnaire, a visualisation ability test, and eye-tracking apparatus.
Context-based gas exercises. The analysis of students’ achievements in the two context-based exercises is presented in this paper. Exercises consist of several questions labelled with Q exercise number.question number (e.g., the first question in the first exercise is labelled ‘Q 1.1’). These specific context-based exercises are two out of 11 science exercises that were studied in the research entitled ‘Explaining Effective and Efficient Problem Solving of the Triplet Relationship in Science Concepts Representations’ (J5-6814), financed by the Slovenian Research Agency (ARRS). The exercises were developed according to the Slovenian students’ achievements on TIMMS (Trends in International Mathematics and Science Study), PISA (The Programme for International Student Assessment), and the Slovenian national external assessment for chemistry, physics, and biology but modified for the purpose of this research. The animations of particulate level of chemical concepts or animated SMRs (just the term animations is used in the further text), were designed by science educators; also authors of this paper, and according to their developed ideas, the computer professional finalized them. The animations were developed for the purposes of this research only and they are not available for general use. Time, in which participants viewed the animations, was not limited. If participants needed more time to solve the exercise, the animations started all over again. However, participants did not have the ability to control the animations. The text of context-based exercises was in the Slovene language. For the purposes of this paper, the context-based exercises’ texts were translated into English (see Fig. 1 and 3).
image file: c8rp00189h-f1.tif
Fig. 1 Screen images (SI W1 and SI W2) of the water context-based exercise (image of a kettle with boiling water from https://www.videoblocks.com/video/boiling-water-in-kettle-slow-motion-oemikdf). The areas of interest in SI W2 are in grey rectangles.

First context-based exercise: water vapour

The first exercise starts with a photo from daily life (of a kettle of boiling water) and later requires the selection of the correct animation and explanation on a submicroscopic level. Due to the length of the task and recommendations for determining AOIs to ensure clean fixation assignment, it has two screen images labelled SI W1 and SI W2 (see Fig. 1). More precisely, in the water context-based exercise, the students were asked to identify water in a gaseous state in the displayed photo and, from three animations, to select the appropriate distribution of water molecules in the gaseous state at the particulate level. Animations were displayed on the computer screen simultaneously and intent to present submicroscopic level of different states of water (variables: speed of particles, number of particles per unit). Animations lasted for 12 seconds (one cycle) and started all over again if needed. In the part of the study with the eye tracker, the focus is on submicroscopic representations. Therefore, only the 2nd screen image of the water context-based exercise (SI W2) was divided into five specific AOIs; photo (a dash-dotted line grey rectangle), Animations 1, 2 and 3 separately (a full line grey rectangle), and other (a dashed line grey rectangles).

Animation 1 presented water as a liquid, animation 2 as a gas, and animation 3 as a solid. Animations differed in the amount of water molecules on the presented area and in the way they move. The animation which correctly illustrates the gaseous state of water is labelled with number 2 (Fig. 2). Students were also asked to justify their decision. This exercise tested the level of students’ recall of knowledge about aggregate states of water.


image file: c8rp00189h-f2.tif
Fig. 2 Six consecutive screenshots (from 1 to 6) of the water context-based exercise SMRs. The elapsed time between the two screenshots is two seconds.

Second context-based exercise: air pump

The second context-based exercise starts with two photos of objects from daily life (of an air pump and an airbed) mentioned in the introduction text with the aim of providing a meaningful context for the exercise (Fig. 3). Later, the exercise focuses on the process of air compression in the air pump (warming of air in a pump) and requires the selection of the correct animation and explanation on a submicroscopic level. Animations of the air-pump context-based exercise intent to present the submicroscopic level of the process of compressing air (variables: size of particles, number of particles per unit, speed of particles). Animations were 16 seconds long (one cycle) and started all over again if needed. The stimuli are designed similarly as for the water context-based exercise presented in Fig. 1. However, the designed air pump context-based exercise is more complex than the water context-based exercise due to the less common (at least in school) context that reflects the length of the exercise. However, its questions are placed on three screen images; labelled SI A1, SI A2, and SI A3 (see Fig. 3). SI A1 and SI A2 include photos of the airbed and/or just the air pump according to the content of the text/question. In the study part with the eye tracker, we focused on animations. To assure enough spacing between AOIs with SMRs to capture the best eye-tracking data, they were placed on the third screen image of the air pump context-based exercise (SI A3). The SI A3 was divided into four specific AOIs; Animations 1, 2 and 3 separately (a full line grey rectangle), and other (a dashed line grey rectangles).
image file: c8rp00189h-f3.tif
Fig. 3 Screen images (SI A1, SI A2, and SI A3) for the air-pump context-based exercise (image of air pump from http://www.ideo.si). The areas of interest in SI A3 are in grey rectangles.

More precisely, based on the photo and students’ everyday experience, they were asked to connect the process of compressing and heating of air with the selection of the appropriate animation. They were also asked to justify their selection. Usually, the use of the pump as presented in the second context-based exercise had not been presented to the students at school. The fact is that the topic is not introduced as a learning goal in science subject curricula. However, students learn how substances are changed during different processes. According to primary school physics curricula (grade 8), students learn that at higher temperature particles move faster (Verovnik et al., 2011). In the animation, the air at the particulate level was represented by nitrogen molecules (blue) and oxygen molecules (red), whereby the volume ratio and size of the molecules were considered. SMRs do not take into account other gases in the unpolluted and dry air (such as argon and carbon dioxide) due to their low concentrations in the air. Students were asked to pay attention to the speed of movement of the particles and to the number of particles in the given space (density) (see Fig. 4) Animation 1 takes into account the increasing number of particles per area unit during the compression and higher speed of particles due to the air heating during the compression process. Animation 2 presents the common misconception that particles shrink and move faster when the air is compressed. Animation 3 illustrates the slower motion of particles after the air compression and testing the misconception that particles move slower in smaller spaces. The animation that correctly illustrates the submicroscopic level of the process of compressing air is labelled with number 1 (Fig. 4).


image file: c8rp00189h-f4.tif
Fig. 4 Eight consecutive screenshots (from 1 to 8) of the air-pump context-based exercise SMRs. The elapsed time between the two screenshots is two seconds.
Test of logical thinking. The test of logical thinking (TOLT) test is a multiple-choice paper-pencil test evaluating five reasoning abilities relevant to the teaching of science (Tobin and Capie, 1984). The test contains 10 problems that require some consideration and the use of problem-solving strategies in different areas (i.e., controlling variables, as well as proportional, correlational, probabilistic, and combinatorial reasoning). Participants got one point for a correct answer and explanation of it (in exercises 1–8) and for the correct combinations and the correct number of them (in exercises 9–10). These points were summed up in a total score (maximum is 10 points), which was used as the main result of the test (Devetak and Glažar, 2010). Students had 38 minutes to solve the test, and it was applied in a group.
Science motivation questionnaire. To measure motivation for science in our study, we used an adapted Slovenian version of the paper-pencil self-evaluation science motivation questionnaire (SMQ) of Glynn et al. (2009). The term “science” included chemistry, biology, and physics. Participants responded to each of the 30 items in the SMQ on a five-point Likert scale ranging from 1-never to 5-always. The questionnaire consists of six five-item scales: (1) intrinsically motivated science learning, (2) extrinsically motivated science learning, (3) relevance of learning science to personal goals, (4) responsibility (self-determination) for learning science, (5) confidence (self-efficacy) in learning science, and (6) anxiety about science assessment. Students could spend 35 minutes to complete the questionnaire. We calculated average responses in all six scales and overall (anxiety was coded in reverse) in order to be able to compare and differentiate all aspects of motivation for science.
Visualisation ability test. The patterns-based approach was used to evaluate students’ visual processing skills (visualisation ability) with the application of the Pattern Comparison Test (PCT) from the PEBL test battery, which is a set of psychological tests for researchers and clinicians. In the PCT, there were 60 pairs of two grid patterns, of which 30 were equal, and 30 were different, displayed on the computer screen (Mueller and Piper, 2014). The participants individually had to compare the stimuli in pairs and respond as quickly as possible, by pressing the certain key on the keyboard, whether the patterns were the same or different. The reaction time and the correctness of the answers were measured. The maximum number of points was 100. Students had 15 minutes to complete the test.
Eye-tracking apparatus. An eye-tracker device can be used to measure eye movements, such as fixations of the gaze to the specific area of the computer screen during a specific activity and saccades, which are eye movements between fixations. To detect students’ visual attention towards different elements of the task on the computer screen, the total amount of time (total fixation duration, TFD) spent in particular areas of interest (AOI) can be measured. This is also defined as the participant's visual attention or attention allocation (Tsai et al., 2012; Havanki and VandenPlas, 2014). To ensure clean fixation assignment, the task displayed on the computer screen must be separated into several carefully and clearly divided AOIs according to the placement elements that are interested from the research problem point of view (Ferk Savec et al., 2016).

The identification of saccades/fixations is based on the motion of gaze during each sample collected. When both the velocity and acceleration thresholds (in our case: 30 degrees per second and 8000 degrees per second squared) are exceeded, a saccade begins; otherwise, the sample is labelled as a fixation.

The screen-based EyeLink 1000 (35 mm lens, horizontal orientation) eye tracker apparatus and associated software (Experiment Builder for the preparation of the experiment and a connection with EyeLink; Data Viewer for obtaining the data and basic analysis) for recordings and analyses of students’ eye movement when solving context-based exercises was used. Data were collected from the right eye (monocular data collection that followed corneal reflection and pupil responses) at 500 Hz (Torkar et al., 2018).

Data collection

TOLT and SMQ were applied a week before the eye-tracker study to the groups of participants in the standards environment. The study with the eye-tracker was conducted from November 2016 to March 2017 and took place in the Laboratory of the Department of Psychology, at the Faculty of Arts, University of Ljubljana. Each participant individually completed the PCT from the PEBL at the same day, but in the different room of the laboratory. After completing the PCT participants entered the eye-tracker room. Participants were not limited in time when solving eleven context-based exercises; it took them approximately 30 minutes to complete them. Before the testing, each participant was individually informed about the purpose of the study, the method used and their role in it. They sat approximately 60 cm away from the screen (distance to the eyes) and had to place their head in a special head-supporting stand, to ensure stability and gather the most optimal recordings. After the initial calibration and validation (through the algorithm of nine points), participants solved all the exercises out loud – verbally provided answers (in the same order for all participants), while the experimenter wrote their answers down. The description of the screen images of the specific exercises, used for the purposes of this paper, is presented in section Context-based gas exercises. The same procedure was conducted with participants in each age group. Their participants’ eye-movements were measured with the eye tracker. All data were collected in the Slovene language.

Data analysis

Participants’ think-aloud responses obtained during their solving the context-based exercises were graded, according to the correct model answer, which was designed on the learning content of current textbooks for primary and secondary schools approved by the Council of Experts of the Republic of Slovenia for General Education. The reliability of grading was ensured by independent grading by two researchers (the two authors of this paper). Altogether, 97% reliability was achieved. To determine how students allocate attention to the various SMRs in context-based exercises, eye-movement measures were obtained (recorded and analysed) with EyeLink 1000 and associated software. Experiment Builder was used for the preparation of the experiment, and a connection with the EyeLink Data Viewer was used for obtaining the data, basic analysis, and drawing of the heat maps. The heat maps’ text presented in Fig. 5 is in the Slovene language because data was collected in Slovene since it is usually appropriate for data to be analysed in the language in which it is collected (Taber, 2018).
image file: c8rp00189h-f5.tif
Fig. 5 Presentations of the density of fixations for students who correctly choose the animation (the correct animation was labelled with 1) (left) and who answered incorrectly (right).

All data were collected in Excel and statistically processed in SPSS (Statistical Package for the Social Sciences). Basic descriptive statistics (medians Mdn and an interquartile ranges IQR) of the numerical variables were determined. The percentage of fixation duration was calculated as the ratio between the total fixation duration on specific AOI and total fixation duration on the screen image. Three groups of students were obtained according to their visualisation abilities, interests, and reasoning abilities. Into Group A (poor abilities) were classified students who scored less than Mdn − 0.5 IQR points, into Group B (average abilities) those who scored between Mdn − 0.5 IQR and Mdn + 0.5 IQR points, and into Group C (superior abilities) students who scored above Mdn + 0.5 IQR points on the test.

The Mann–Whitney U and Kruskal–Wallis non-parametric tests were used to explain the relationship between the correctly solved questions, including animation at the sub-microscopic level and formal-reasoning abilities, visualisation abilities and interest as well as attention allocation of the context-based water and air pump exercises due to the small sample size and non-normal distribution (Pallant, 2011). The Mann–Whitney U test is used to test for differences between two independent groups on a continuous measure that compares medians. It converts the scores on the continuous variable to ranks across the two groups. It then evaluates whether the ranks for the two groups differ significantly. As the scores are converted to ranks, the actual distribution of the scores does not matter. The Kruskal–Wallis test allows comparing the scores on a continuous variable for three or more groups. Scores are converted to ranks, and the mean rank for each group is compared. Post hoc Dunn's tests are used to determine the differences between different pairs of groups (Pallant, 2011). Statistical hypotheses were tested at a 5% alpha error rate. To describe whether the effects have a relevant magnitude, effect size measure eta squared η2 was used to describe the strength of a phenomenon. Benchmarks to define small (0.01), medium (0.06), and large (0.14) effects were provided by Cohen (1988).

Results and discussion

The results and discussions are presented according to the research questions.

Educational level impact on achievements on two context-based exercises

The first research question deals with the differences in the understanding of concepts presented in two context-based exercises that include animations at the sub-microscopic level of the gaseous state of water and air between students at different educational levels. Table 1 shows the percentage of students who correctly answered different questions of the two context-based exercises in different age groups of students (12-, 16- and 23-years old).
Table 1 The percentage of students of different age groups who answered correctly on different context-based exercises and the average achievements in the water and air context-based exercise
Context-based exercises Primary school students [%] Upper secondary school students [%] University students [%]
SI W1 1.1 steam 100 100 100
1.2 water 87 83 80
1.3 water molecules 13 48 55
1.4 gaseous 97 100 100
SI W2 1.5 correct selection 97 100 100
1.6 correct justification 13 7 40
Average 68 73 79
SI A1 2.1 air 80 90 75
SI A2 2.2 reduced volume and a higher temperature 47 38 90
SI A3 2.3 correct selection 30 55 75
2.4 correct justification 3 38 65
Average 40 55 76

The first context-based exercise covers the gaseous state of water. The science curricula prescribe the learning content on the states of water for 11-year-olds; therefore, 12-year-olds should have basic knowledge about this (Skvarč et al., 2011). Students were able to identify steam (Q 1.1). Around 80% of students answered Q 1.2 (water forms steam) correctly. Question Q 1.3 requires an answer at the sub-microscopic level – water molecules – since asking what forms the substance (water) which was an answer on the previous question. Common incorrect responses to the question, what does the substance in the (red) marked box consist of, were: water, hydrogen and oxygen, two oxygens and one hydrogen. It is evident that 13% of 12-year-old students and half of the students of age groups 2 and 3 (16 and 23 years old) are not familiar with the fact that water molecules form steam. These findings further support the idea that students do not correlate SMR with the explanation at the submicroscopic level.

All of the students answered correctly that the gaseous state (Q 1.4) of water is presented on SI W1 in the photo and they all chose the correct animation at the particulate level representing the steam of water in SI W2 (Q 1.5). Students had to justify their selection of the animation at Q 1.6. It was expected that students would elaborate on their answer at the sub-microscopic level. As a correct answer to question Q 1.6, it was also considered that at the particulate level water molecules in the gaseous state move freely, there are mostly no attractions between them (even that water molecules in the gaseous state form aggregates because of the hydrogen bonding), and that the speed of the molecules is high. The distances between the molecules are large, and they are not arranged in an orderly way, and they are moving in the whole area at their disposal (Slapničar et al., 2017). Table 1 shows that students of all age groups have difficulties in justifying their answers, taking into account the model answer presented above.

From the students’ achievements on the context-based exercise with the air pump, it is evident that the percentage of students who solved all four questions of the exercise correctly increased with the level of education. However, the Slovene science curriculum in primary school introduces matter at the sub-microscopic level in Grade 6 using 2D static one-particle SMRs. More complex 3D dynamic SMRs are applied in Grade 8 in chemistry and physics courses (Bačnik et al., 2009; Planinšič et al., 2009; Bačnik et al., 2011; Verovnik et al., 2011). At about the same time, the transition from the concrete to the formal operational stage appears (Labinowicz, 1989).

The results show that more than 75% of students correctly answered that the airbed was filled with air using an air pump (Q 2.1); 25% of 23 year-old students have difficulties with question Q 2.1. The next question in the second context-based exercise (Q 2.2) deals with the properties of air changing during the pumping of the airbed. It was expected that participants would suggest that the volume of air was reduced, while the air was simultaneously heated. Less than 50% of 12 year-old and 16 year-old students answered correctly, whereas 90% of the 23 year-olds gave the correct answer.

Question Q 2.3 required students to select the correct animation of the air at the particulate level when compressed in an air pump. Table 1 shows that around one third of 12 year-old students chose the right animation, 55% of 16 year-old students and 75% of 23 year-old-students did so.

The differences in achievements of solving both context-based exercises are statistically significant between the different age groups of participants with medium and large effect sizes (water: Kruskal–Wallis χ2(2) = 7.4, p = 0.025, η2 = 0.071; pump: Kruskal–Wallis χ2(2) = 21.1, p ≤ 0.001, η2 = 0.251). However, it was identified in the study of Slapničar et al. (2017) that students have difficulties in observing three animations with two variables changing at the same time for representing unknown exercises. Therefore, effort should also be made to allow 12 year-olds to slowly and gradually develop their representational competence (Kozma and Russell, 2005; Gilbert, 2008; Phillips et al., 2010).

The students justified their selection of the correct animation (Q 1.4) by their observations of the faster movement of particles and their greater number per unit volume. It is evident that justification is a difficult exercise for students of all three age groups. The percentage of students who gave the correct justification increases with the level of education up to 65% with pre-service science teachers (Group 3).

The analysis of students’ answers which require a justification of the choice of animation (water vapour SMR – Q 1.6; air in the air pump SMR Q 2.4) showed that students argue their answers by using different levels of the representation of chemical concepts (Table 2).

Table 2 Levels of correct and incorrect answers to Q 1.6 and Q 2.4 of different age groups
Exercise Level Incorrect answers Correct answers
Primary school students (Group 1) [%] Upper secondary school students (Group 2) [%] University students (Group 3) [%] Primary school students (Group 1) [%] Upper secondary school students (Group 2) [%] University students (Group 3) [%]
Q 1.6 Macroscopic 14 7 0 0 0 0
Sub-microscopic 28 55 55 3 7 30
Macroscopic and sub-microscopic 45 31 10 10 0 10
Sub-microscopic and symbolic 0 0 5 0 0 0
Q 2.4 No answer 10 7 0 0 0 0
Macroscopic 18 7 0 0 0 0
Sub-microscopic 44 21 10 3 17 15
Macroscopic and sub-microscopic 27 27 25 0 21 50

It is evident that more than half of the students from Group 1 answered Q 1.6 with macro or a combination of macro and submicroscopic levels, whereas more than half of the students in Groups 2 and 3 answered on the sub-microscopic level. The answer to Q 2.4 was given only on the macro level by 18% of Group 1 students. This is expected because the students in the first group are the youngest, and they are not used to applying the submicroscopic level of science concepts in their explanations (Haluk, 2011).

The differences in justifying the correct animation of the sub-microscopic representation in the first and second context-based exercises between groups are statistically significant with a medium effect size for the first context-based exercise and large effect size for the second exercise (Kruskal–Wallis χ2(2) = 9.4, p = 0.009, η2 = 0.097; Kruskal–Wallis χ2(2) = 28.4, p ≤ 0.001, η2 = 0.347). The correct interweaving of the macro and sub-microscopic levels of representation of chemical concepts is extremely important; therefore, students gradually build their knowledge at the macroscopic, submicroscopic, and symbolic levels during schooling (Chittleborough, 2014) as seen from the results of this study. Several studies have shown that macroscopic properties are often assigned to particulate matter, thus revealing students’ misconceptions regarding the macro and sub-microscopic levels of particulate matter representations (Nicoll, 2001; Chiu et al., 2002; Bucat and Mocerino, 2009; Slapničar et al., 2017).

Influence of educational level on attention allocation while solving two context-based exercises

The second research question refers to the differences between students at different educational levels in total fixation durations on the whole screen image and the specific area of interest representing animations in the context-based exercises. Table 3 presents the total fixation duration on the whole screen image for water and air context-based exercises. A Kruskal–Wallis test shows that there are statistically significant differences between different age groups of students and time spent on the specific screen image. The post hoc tests to test pairwise comparisons were conducted to show which group is statistically different to which.
Table 3 TFD in seconds spent on the screen image of the exercise with water (SI W1 and SI W2) and air (SI A1, SI A2, and SI A3) in the gaseous state. Students of Group 1 usually spent more time on the certain SI than students of Groups 2 and 3. Whether differences between age groups are statistically significant was verified using the Kruskal–Wallis test
Screen image Primary school students (Group 1) Upper secondary school students (Group 2) University students (Group 3) Kruskal–Wallis test
Mdn [s] IQR [s] Mdn [s] IQR [s] Mdn [s] IQR [s] χ 2 P η 2
Results of post hoc Dunn's tests:a Group 1–2: P = 0.303; group 2–3: P = 0.087; group 1–3: P ≤ 0.001; 1 = 2, 2 = 3, 1 > 3.b Group 1–2: P = 0.003; group 2–3: P = 0.502; group 1–3: P ≤ 0.001; 1 > 2 = 3.c Group 1–2: P = 0.060; group 2–3: P = 0.209; group 1–3: P = 1.000; 1 = 2 = 3.d Group 1–2: P = 0.002; group 2–3: P = 1.000; group 1–3: P = 0.002; 1 > 2 = 3.e Group 1–2: P = 0.094; group 2–3: P = 1.000; group 1–3: P = 0.003; 1 = 2, 2 = 3, 1 > 3.
SI W1 33.1 15.6 28.8 11.2 20.7 11.0 13.528 0.001 0.152
SI W2 31.2 15.6 22.7 11.1 19.3 9.8 21.305 ≤0.001 0.254
SI A1 18.6 5.8 15.5 7.1 18.6 9.5 6.134 0.047 0.054
SI A2 39.3 38.8 29.8 11.9 26.3 31.2 15.749 ≤0.001 0.182
SI A3 66.3 40.3 56.0 39.4 49.9 30.0 8.048 0.018 0.080

It can be identified that the total fixation duration on the screen images decreases with the students’ age. The differences between groups are statistically significant; size effects are from small to large (η2 in Table 3). However, the second screen image of the water exercise (SI W2) and the third screen image of the air pump exercise (SI A3) included three different animations (one of them correct), and the example presented in SI W2 (boiling water) was more or less known to the students for all three age groups. In contrast, the example under SI A3 (air pump) was less known to them, as can be anticipated from the science curricula at different school levels. It is obvious that students need more time to solve context-based exercises comprising examples that present novel stimuli (Susac et al., 2014). Similar results were obtained by other researchers dealing with problem solving (Gilbert, 2005; Gilbert, 2008; Phillips et al., 2010).

The second screen image of the water context-based exercise and the third screen image of the air pump exercise were divided into more than three AOIs (Animation 1, Animation 2, Animation 3, photo or/and other) (Fig. 1 and 3). Each animation is the separate AOI. In Tables 4 and 5, the total fixation durations and percentages of fixation duration for students of different age groups for water and air pump context-based exercises are presented. According to the known duration of animations and the total fixation duration on the specific animation, the participants viewed each individual animation shorter time on average, than it is needed for the animations one cycle from start to finish (see the above description of the context-based gas exercises).

Table 4 Total fixation duration and percentage of fixation duration of different age groups of students spent on the AOIs on water context-based exercise. The correct animation is written in bold. The results of the Kruskal–Wallis test and post hoc tests are added
Primary school students (Group 1) Upper secondary school students (Group 2) University students (Group 3) Kruskal–Wallis test
Mdn [s] IQR [s] Mdn [s] IQR [s] Mdn [s] IQR [s] χ 2 P η 2
Results of post hoc Dunn's tests:a Group 1–2: P = 0.005; group 2–3: P = 1.000; group 1–3: P = 0.001; 1 > 2 = 3.b Group 1–2: P = 0.037; group 2–3: P = 0.196; group 1–3: P ≤ 0.001; 1 > 2 = 3.c Group 1–2: P = 1.000; group 2–3: P ≤ 0.001; group 1–3: P ≤ 0.001; 1 = 2 > 3.d Group 1–2: P = 0.027; group 2–3: P = 0.148; group 1–3: P ≤ 0.001; 1 > 2 = 3.e Group 1–2: P = 0.077; group 2–3: P = 1.000; group 1–3: P = 0.039; 1 = 2, 2 = 3, 1 > 3.f Group 1–2: P = 0.660; group 2–3: P = 0.025; group 1–3: P = 0.001; 1 = 2 < 3.g Group 1–2: P = 0.938; group 2–3: P = 0.301; group 1–3: P = 0.031; 1 = 2, 2 = 3, 1 > 3.h Group 1–2: P = 0.002; group 2–3: P ≤ 0.001; group 1–3: P = 0.002; 1 < 2; 1 > 3; 2 > 3.
Total fixation duration [s]
Animation 1 5.8 8.2 2.2 2.2 2.3 2.1 15.948 ≤0.001 0.184
Animation 2 7.3 9.0 4.7 4.1 7.5 7.7 2.544 0.280 0.007
Animation 3 2.8 4.3 1.7 2.0 0.8 1.8 17.491 ≤0.001 0.204
Photo 2.9 4.3 2.2 3.9 0.4 0.7 21.431 ≤0.001 0.256
Other 10.7 8.3 7.2 5.0 5.3 3.7 19.390 ≤0.001 0.229
Percentage of fixation duration [%]
Animation 1 19.5 18.3 10.0 10.5 11.5 9.5 7.761 0.021 0.076
Animation 2 22.5 21.2 27.0 20.0 43.5 23.7 14.410 0.001 0.163
Animation 3 7.0 10.5 8.0 8.0 3.5 11.5 6.601 0.037 0.061
Photo 9.5 14.0 10.0 17.0 1.5 3.8 16.523 ≤0.001 0.191
Other 34.0 11.0 32.0 18.5 31.0 21.5 0.490 0.783 0.020

Table 5 Total fixation durations and percentages of fixation duration of different age groups of students spent on the AOIs on the air pump context-based exercise. The number of the correct animation is written in bold. The results of the Kruskal–Wallis test and post hoc tests are added
AOI Primary school students (Group 1) Upper secondary school students (Group 2) University students (Group 3) Kruskal–Wallis test
Mdn [s] IQR [s] Mdn [s] IQR [s] Mdn [s] IQR [s] χ 2 P η 2
Results of post hoc tests:a Group 1–2: P = 0.149; group 2–3: P = 1.000; group 1–3: P = 0.069; 1 = 2 = 3.b Group 1–2: P = 0.111; group 2–3: P = 1.000; group 1–3: P = 0.075; 1 = 2 = 3.c Group 1–2: P = 0.055; group 2–3: P = 1.000; group 1–3: P = 0.019; 1 = 2; 2 = 3; 1 < 3.
Total fixation duration [s]
Animation 1 8.6 14.6 13.7 14.8 13.7 7.7 2.123 0.346 0.003
Animation 2 13.8 19.1 9.8 12.9 10.5 10.9 4.417 0.110 0.031
Animation 3 8.0 8.6 6.4 5.2 4.0 6.1 6.048 0.049 0.053
Other 28.2 23.2 20.6 11.9 19.3 16.0 6.503 0.039 0.059
Percentage of fixation duration [%]
Animation 1 11.4 21.3 26.6 18.3 25.4 16.6 9.122 0.010 0.094
Animation 2 19.5 24.0 18.1 18.9 19.4 11.7 0.255 0.880 0.023
Animation 3 11.8 17.5 9.4 6.0 9.5 5.3 1.497 0.473 0.007
Other 40.6 24.3 36.7 20.1 39.5 16.1 0.232 0.891 0.023

It is evident that not all animations were equally interesting for the students while solving the exercises, and there are statistically significant differences in some cases (see Tables 4 and 5). Students’ total fixation durations were the longest on the animation presenting the molecules of water in gaseous states. The other two animations presented the molecules of water in the solid and liquid states. This example is used several times during the lessons and reflects the memorisation of the student. It also confirms the fact that students spend significantly longer looking at their chosen answer than the rejected answers with familiar stimuli (Susac et al., 2014). However, from the results shown in Table 1, it can be concluded that choosing a correct animation of water in the gaseous state was not at all difficult for the students of all three age groups because almost all of them chose the correct animation.

The context-based exercise with the air pump included three animations in which the air at the particulate level was represented by nitrogen molecules and oxygen molecules, taking into account the volume ratio and the size of molecules. Two variables were changing at the same time during the animation. The correct animation was Animation 1 and included the changing of the speed of the particles and the number of particles in the given space. The percentage of fixation durations’ size order is similar for all animations due to the novel stimuli since students have to carefully observe each animation before making a decision. However, the percentage of duration fixation on Animation 1 was statistically significantly higher for students of Groups 2 and 3. In contrast, for Group 1 students, the percentage of fixation duration is the largest in Animation 2, which might indicate an already developed misunderstanding that particle sizes changes during heating (Table 1). On the other hand, the information provided by the animations does not give students any key information to decide about the correct particle behaviour in the air pump due to their limited knowledge about changes of air properties during compression.

Students’ achievement in selecting the correct SMR and their educational background, interest, formal reasoning abilities, and visualisation abilities

The third research question deals with the differences between students’ who correctly or incorrectly choose the animation in the context-based exercises, which is significant regarding their levels of formal-reasoning abilities, visualisation abilities, motivation, and attention allocation on the specific AOIs representing animations in the context-based exercises. The third research question refers only to the second context-based exercise with the air pump because almost all students correctly selected the animation of particles in the water context-based exercise. The division of students into two groups is thus meaningless. Therefore, only the air-pump context-based exercise will be analysed further, using two groups of students.

The participants were divided into two groups based on the correctly (1) and incorrectly (2) chosen animation in the air pump context-based exercise (Q 2.3); two similar groups regarding the number of students in each group were obtained. Table 6 presents the results of the Mann–Whitney U test showing the differences between the specific groups of students in formal reasoning abilities, motivation, visualisation, total fixation duration, and percentage of fixation duration on animations.

Table 6 Results of Mann–Whitney U test for students divided into two groups, with correctly and incorrectly chosen animations in the air pump context-based exercise
Incorrect (n = 39) Correct (n = 40) Mann–Whitney U test
Mdn IQR Mdn IQR U P η 2
TOLT – test of logical thinking; SMQ – science motivation questionnaire; PCT – visualisation ability test.
TOLT (10 pts possible) 4.0 6.0 7.0 3.0 489.500 0.015 0.103
SMQ (5 pts possible) 3.5 0.54 3.5 1.1 718.95 0.975 0.005
PCT (100 pts possible) 47.0 28.0 52.0 27.0 675.000 0.303 0.013
Total fixation duration on AOI with correct animation 1 [s] 8.7 11.6 14.3 12.1 987.000 0.042 0.052
Percentage of fixation duration on AOI with correct animation 1 [%] 11.8 17.3 27.3 18.7 1112.000 0.001 0.134
Total fixation duration on AOI with incorrect animation 2 [s] 14.3 19.3 9.0 12.0 530.000 0.014 0.076
Percentage of fixation duration on AOI with incorrect animation 2 [%] 19.2 22.9 18.8 16.0 631.500 0.145 0.027
Total fixation duration on AOI with incorrect animation 3 [s] 6.9 8.5 5.7 5.9 608.000 0.092 0.036
Percentage of fixation duration on AOI within correct animation 3 [%] 10.8 12.0 9.3 6.6 686.500 0.359 0.011

Formal reasoning abilities have a statistically significant medium effect size on the achievements in selecting the correct animation in the air pump context-based exercise (Q 2.3) (Fig. 5). Similarly statistically significant differences appear in attention allocation on the area with correct animation with achievements in Q 2.3. It was expected that only students with higher formal reasoning abilities would select the correct 3D SMR, which is in accordance with the findings by Devetak and Glažar (2010, 2014) about statistically significant correlations between formal reasoning abilities and students’ knowledge, especially on the sub-microscopic level. The achievements in selecting the correct animation in the air context-based exercise are not statistically significantly different among students with different motivation levels. However, Juriševič et al. (2008) suggested that the correlation between the level of motivation and the chemical knowledge is moderate but statistically significant. Q 2.3 referred to an animation for which visualisation ability might come to the forefront (Phillips et al., 2010). It was not identified that students’ visualisation abilities (identified through the PCT – a test of visual processing and pattern recognition) play an essential role in selecting the correct animation in this specific context.

Students’ attention allocation to the correct animation and their educational background, interest, formal reasoning abilities, and visualisation abilities

The fourth research question relates to differences in attention allocation on the AOI representing the correct animation in the context-based exercises between students with different levels of formal-reasoning abilities, visualisation abilities, and motivation.

The Kruskal–Wallis test was used to determine the statistically significant differences in total fixation duration on the area of interest with correct animation in the water and air pump context-based exercises (Q 1.5 and Q 2.3) between students of different formal reasoning abilities, motivation, and visualisation. Students were divided into three groups (poor, average and superior abilities), taking into account the Mdn and 0.5 IQR of the testing variable (see Tables 7 and 8). The results show that there are no statistically significant differences between the three groups of students in attention allocation on AOIs with the correct animations (Q 1.5 and Q 2.3) and motivation as well as visualisation ability. However, there are statistically significant differences in the groups of students with the air pump context-based exercise in total fixation duration on the area of interest with correct animation at air pump context-based exercises (Q 2.3) and formal reasoning abilities. The magnitude of the differences in the medians is large; in contrast, this is not significant for the water context-based exercise (Q 1.5). From Tables 7 and 8, it is evident that the percentage of fixation duration on the AOI with correct animation for water and air pump exercises is statistically significantly different between groups of students with different levels of thinking.

Table 7 Results of Kruskal–Wallis test regarding exploring differences in total fixation duration on AOI with correct animation and percentage of fixation duration on AOI with correct Animation 2 for the water context-based exercise (Q 1.5) versus independent variables: formal reasoning abilities (TOLT), motivation (SMQ), and visualisation ability (PCT). Three groups regarding Mdn − 0.5IQR (Group A), Mdn ± 0.5IQR (Group B) and Mdn + 0.5IQR (Group C) on each tested independent variable were formed
Testing variable Group A Group B Group C Kruskal–Wallis test
Mdn IQR Mdn IQR Mdn IQR χ 2 P η 2
Thinking ability (TOLT) Total fixation duration 7.2 9.5 6.9 6.4 12.8 — 0.014 0.993 0.026
Percentage of fixation duration 24.0 23.5 31.0 25.0 43.0 — 2.351 0.309 0.005
Motivation (SMQ) Total fixation duration [s] 6.6 3.6 7.0 7.3 7.3 12.3 1.699 0.428 0.004
Percentage of fixation duration [%] 23.0 18.5 27.5 24.3 42.0 32.0 3.572 0.168 0.021
Visualisation ability (PCT) Total fixation duration [s] 7.4 5.1 6.9 8.7 7.0 8.4 0.299 0.861 0.022
Percentage of fixation duration [%] 22.0 18.0 30.0 22.0 35.5 29.3 2.673 0.263 0.008

Table 8 Results of Kruskal–Wallis test regarding exploring differences in attention allocation on AOI with correct animation and percentage of attention allocation on AOI with correct Animation 1 for the air pump context-based exercise (Q 2.3) versus independent variables: formal reasoning abilities (TOLT), motivation (SMQ), and visualisation ability (PCT). Three groups regarding Mdn − 0.5IQR (Group A), Mdn ± 0.5IQR (Group B) and Mdn + 0.5IQR (Group C) on each testing independent variable were formed
Testing variable Group A Group B Group C Kruskal–Wallis test
Mdn IQR Mdn IQR Mdn IQR χ 2 P η 2
Results of post hoc tests:a Group A–B: P = 0.014; group B–C: P = 1.000; group A–C: P = 0.402; A < B = C.b Group A–B: P = 0.001; group B–C: P = 0.826; group A–C: P = 0.065; A < B = C.
Thinking ability (TOLT) Total fixation duration [s] 6.8 11.0 13.9 14.2 26.7 — 8.771 0.012 0.089
Percentage of fixation duration [%] 11.1 17.3 26.6 20.2 40.8 — 15.502 ≤0.001 0.178
Motivation (SMQ) Total fixation duration [s] 16.2 20.3 12.1 14.1 10.7 11.6 1.585 0.453 0.005
Percentage of fixation duration [%] 27.7 18.8 21.2 18.9 17.5 25.2 0.802 0.670 0.016
Visualisation ability (PCT) Total fixation duration [s] 10.7 14.1 10.2 15.7 15.5 9.5 2.302 0.316 0.004
Percentage of fixation duration [%] 15.8 20.5 15.0 24.0 25.5 15.1 2.842 0.241 0.011

The context-based water exercise is cognitively less demanding; only one particle type is presented. The second exercise is cognitively more demanding; students have to use higher cognitive strategies to select the correct animation and describe the observation accurately. The animation comprises two different types of particles (oxygen and nitrogen molecules) moving quite quickly. Gilbert (2005, 2008) and Phillips et al. (2010) report difficulties in constructing relevant information from the dynamic visual representations because the information was presented too quickly.

Conclusions

This paper aimed to explore and explain students’ achievements in solving context-based gas exercises comprising the macro and submicroscopic levels of chemical concepts. The influence of other independent variables, such as motivation, formal-reasoning abilities, visualisation abilities, are also important factors that should be considered when explaining students’ ability to solve exercises. Following this assumption, the above variables were measured, and four sets of conclusions are presented.

The impact of educational level on students’ achievements on two context-based exercises

It can be concluded from the first set of results that there are even for advanced students still some gaps in students’ knowledge and understanding concepts related to the gaseous states of water and air. Students’ average achievements in solving the water exercise (73% of all points) were higher than solving the air pump exercise (57% of all points). Students of all age groups had difficulties solving the air pump exercise. This might indicate that students are less familiar with changes of particle behaviour during air pumping in comparison with the water exercise. Twenty-three-year-olds are more familiar with the explaining of science concepts by interweaving submicroscopic and macroscopic levels.

The impact of educational level on individuals’ attention allocation while solving two context-based exercises

The second set of findings is related to the differences between students at different educational levels in the duration of total fixations on the whole screen image and the specific areas of interest representing animations in the context-based exercises. It can be summarised that total fixation duration on animations comprising examples with novel stimuli (i.e., air pumping) is longer. The percentage of fixation duration that students from different age groups spend on the correct animations (area of interest) is significantly different. These findings suggest that the students’ age (years of chemistry education) influence the fixation duration on the correct animation.

Differences between students who correctly or incorrectly chose the SMR in the context-based exercises

The third part of the results indicates the differences between students who correctly or incorrectly chose the 3D dynamic SMR in the context-based exercise comprising an air pump in their levels of formal-reasoning abilities, visualisation abilities, motivation, the total fixation duration, and the fixation count on the specific AOIs. The major finding is that students’ formal reasoning abilities and the total fixation duration on specific areas of interest may play an important role in the selection of the correct animation. It is also evident that visualisation abilities and motivation do not play a role in choosing the correct animation.

Differences in visual attention on the AOI representing the correct animation

The last set of results relates to the differences in the duration of total fixation on the area of interest representing the correct animation in the context-based exercises between students with different levels of formal-reasoning abilities, visualisation abilities, and motivation. One of the more significant findings to emerge from this research is that the percentage of duration of students’ total fixation on the correct animation is statistically significantly different regarding their formal reasoning abilities only when students had to solve the more cognitively demanding context-based exercise with an air pump; this indicates that this task is for students more of a problem than an exercise.

Limitations of this research

The major limitation of this research is the small sample in each age group. The selection of contexts on which the exercises are defined should be gender neutral. Context-exercises should be designed in a similar way (e.g., the amount of the text, the number of macro-figures, and the number of variables on a specific animation of particles). The triple nature of chemical concepts should be integrated into the context-based exercises when applicable; exercises discussed in this research were not designed in such a way that the symbolic level is necessary. Screen images should be developed so that eye-tracking measurements are clearly defined. An additional limitation of this study is also students’ knowledge. If students do not know that the temperature is rising on compression of air in the air pump, they might select the wrong animation even knowing that gas molecules move faster when they are heated.

Implications for educational process

The evidence from the research suggests that more emphasis should be given to teaching different examples based on specific contexts. These exercises should be defined on different cognitive levels regarding students’ age. At all levels of education, students should be exposed to applying the triple nature of chemical concepts. When animations of particles are applied, students should learn how to derive the conclusion on the basis of specific and detailed observations.

Further research guidelines

Further work needs to be done to establish whether the symbolic level of chemical concepts influences context-based problem solving from the information-processing point of view. Another possible area for future research would be to investigate why students’ motivation for learning chemistry and visualisation abilities does not reflect their exercise-solving abilities. The issue of teachers’ role in presenting animations of particles when teaching specific chemical concepts is an intriguing one, which could be usefully explored in further research.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This research was supported by the project “Explaining Effective and Efficient Problem Solving of the Triplet Relationship in Science Concepts Representations” (J5-6814), financed by the Slovenian Research Agency (ARRS). We would also like to thank Manja Veldin for useful discussions about the draft paper and the preparation of heat maps.

References

  1. Bačnik A., Bukovec N., Poberžnik A., Požek Novak T., Keuc Z., Popič H. and Vrtačnik M., (2009), Učni načrt, Program srednja šola, Kemija: gimnazija: klasična, strokovna gimnazija, [Curriculum, Program of Secondary School, Chemistry: Gymnasium: Classical, Professional Gymnasium], Ljubljana: National Education Institute Slovenia.
  2. Bačnik A., Bukovec N., Vrtačnik M., Poberžnik A., Križaj M., Stefanovik V., Sotlar K., Dražumerič S. and Preskar S., (2011), Učni načrt, Program osnovna šola, Kemija, [Curriculum, Program of Primary School, Chemistry], Ljubljana: National Education Institute Slovenia.
  3. Balon A., Gostinčar Blagotinšek A., Papotnik A., Skribe Dimec D. and Vodopivec, I., (2011), Učni načrt, Program osnovna šola, Naravoslovje in tehnika, [Curriculum, Program of Primary School, Science and Technology], Ljubljana: National Education Institute Slovenia.
  4. Barnea N., (2000), Teaching and Learning about Chemistry and Modelling with a Computer Managed Modelling System, in Gilbert J. K. and Boulter C. (ed.), Developing models in science education, Dordrecht: Kluwer Academic Publishers, pp. 307–324.
  5. Bucat B. and Mocerino M., (2009), Learning at the Submicro Level: Structural Representations, in Gilbert J. K. and Treagust D. (ed.), Multiple Representations in Chemical Education, Dordrecht: Springer, pp. 11–29.
  6. Bunce D. M. and Gabel D., (2002), Differential Effects in the Achievement of Males and Females of Teaching the Particulate Nature of Chemistry, J. Res. Sci. Teach., 39(10), 911–972.
  7. Cavas P., (2011), Factors Affecting the Motivation of Turkish Primary Students for Science Learning, Sci. Educ. Int., 22(1), 31–42.
  8. Cheng M. M. W. and Gilbert J. K., (2017), Modelling Students’ Visualization of Chemical Reaction, Int. J. Sci. Educ., 39(9), 1173–1193.
  9. Chittleborough G., (2014), The Development of Theoretical Frameworks for Understanding the Learning of Chemistry, in Devetak I. and Glažar S. A. (ed.), Learning with Understanding in the Chemistry Classroom, Dordrecht: Springer, pp. 25–40.
  10. Chiu M. L., Chiu M. H. and Ho C. Y., (2002), Using Cognitive-based Dynamic Representations to Diagnose Students' Conceptions of the Characteristics of Matter, Proc. Natl. Sci. Counc., 12(3), 91–99.
  11. Cohen J., (1988), Statistical Power Analysis for the Behavioral Sciences, 2nd edn, Hillsdale, NJ: Lawrence Erlbaum Associates.
  12. Correia A. P., Koehler N., Thompson A. and Phye G., (2018), The Application of PhET Simulation to Teach Gas Behavior on the Submicroscopic Level: Secondary School Students’ Perceptions, Res. Sci. Technol. Educ., retrieved from  DOI:10.1080/02635143.2018.1487834.
  13. Devetak I., (2012), Zagotavljanje kakovostnega znanja naravoslovja s pomočjo submikroreprezentacij, Analiza ključnih dejavnikov zagotavljanja kakovosti znanja v vzgojno – izobraževalnem sistemu [The Analysis of the Key Factors in Ensuring the Quality of Knowledge in Educational System], Ljubljana: Faculty of Education, University of Ljubljana.
  14. Devetak I. and Glažar S. A., (2010), The Influence of 16-year-old Students' Gender, Mental abilities, and Motivation on their Reading and Drawing Submicrorepresentations Achievements, Int. J. Sci. Educ., 32(12), 1561–1593.
  15. Devetak I. and Glažar S. A., (2014), Learning with Understanding in the Chemistry Classroom Constructing Active Learning in Chemistry: Concepts, Cognition and Conceptions, in Devetak I. and Glažar S. A. (ed.), Learning with Understanding in the Chemistry Classroom, Dordrecht: Springer, pp. 5–23.
  16. Devetak I., Drofenik Lorber E., Juriševič M. and Glažar S. A., (2009a), Comparing Slovenian Year 8 and Year 9 Elementary School Pupils’ Knowledge of Electrolyte Chemistry and their Intrinsic Motivation, Chem. Educ. Res. Pract., 10(4), 281–290.
  17. Devetak I., Vogrinc J. and Glažar S. A., (2009b), Assessing 16-year-old Students’ Understanding of Aqueous Solution at Submicroscopic Level, Res. Sci. Educ., 39(2), 157–179.
  18. Eilam B. and Gilbert J. K. (ed.), (2014), Science Teachers' Use of Visual Representations, Dordrecht: Springer.
  19. Exerciseer R. and Dalton R., (2006), Research into Practice: Visualization of the Molecular Word Using Animations, Chem. Educ. Res. Pract., 7(2), 141–159.
  20. Ferk Savec V., Hrast Š., Devetak I. and Torkar G., (2016), Beyond the Use of an Explanatory Key Accompanying Submicroscopic Representations, Act. Chim. Slov., 63(4), 864–873.
  21. Frackson M., Vivien M. C. and Asiana B., (2014), Comparing Male and Female Pre-service Teachers’ Understanding of the Particulate Nature of Matter, J. Balt. Sci. Educ., 13(6), 821–827.
  22. Gilbert J. K., (2004), Models and Modelling: Routes to More Authentic Science Education, Int. J. Sci. Math. Educ., 2(2), 115–130.
  23. Gilbert J. K., (2005), Visualization: A Metacognitive Skill in Science and Science Education, in Gilbert J. K. (ed.), Visualization in Science Education, Dordrecht: Kluwer Academic Press, pp. 9–27.
  24. Gilbert J. K., (2008), Visualization: An Emergent Field of Practice and Enquiry in Science Education, in Gilbert J. K., Reiner M. and Nakhlem M. (ed.), Visualization: Theory and Practice in Science Education, Dordrecht: Springer, pp. 3–24.
  25. Glynn S. M., Taasoobshirazi G. and Brickman P., (2009), Science Motivation Questionnaire: Construct Validation with Nonscience Majors, Inc. J. Res. Sci. Teach., 46(2), 127–146.
  26. Haluk O., (2011), Turkish Primary Students’ Conceptions about the Particulate Nature of Matter, Int. J. Environ. Sci. Educ., 6(1), 99–121.
  27. Harrison A. G. and Treagust D. F., (2000), Learning about Atoms, Molecules, and Chemical bonds: A Case Study of Multiple–model Use in Grade 11 Chemistry, Sci. Educ., 84(3), 352–381.
  28. Havanki K. L. and VandenPlas J. R., (2014), Eye Tracking Methodology for Chemistry Education Research, in Bunce D. M. and Cole R. S. (ed.), Tools of Chemistry Education Research, Washington, DC: American Chemical Society, pp. 191–218.
  29. Hinze S. R., Williamson V. M., Shultz M. J, Williamson K. C., Deslongchamps G. and Rapp D. N., (2013), When do Spatial Abilities Support Student Comprehension of STEM Visualizations? Cogn. Process. Int. Q. Cogn. Sci., 14(2), 129–142.
  30. Ho H. N. J., Tsai M. J., Wang C. Y. and Tsai C. C., (2014), Prior Knowledge and Online Inquiry-based Science Reading: Evidence from Eye tracking, Int. J. Sci. Math. Educ., 12(3), 525–554.
  31. Hyönä J., (2010), The Use of Eye Movements in the Study of Multimedia Learning, Learn. Instr., 20(2), 172–176.
  32. Johnstone A. H., (1982), Macro- and Micro-Chemistry, Sch. Sci. Rev., 64(227), 377–379.
  33. Johnstone A. H., (1991), Why is science difficult to learn? Things are seldom what they seem, J. Comp. Assist. Learn., 7(2), 75–83.
  34. Johnstone A. H., (2001), Teaching of Chemistry-Logical or Psychological? Chem. Educ. Res. Pract., 1(1), 9–15.
  35. Johnstone A. H. and El-Banna H., (1986), Capacities, Demands and Processes – A Predictive Model for Science Education, Educ. Chem., 23(3), 80–84.
  36. Juriševič M., Devetak I., Razdevšek Pučko C. and Glažar S. A., (2008), Intrinsic Motivation of Pre-service Primary School Teachers for Learning Chemistry in Relation to Their Academic Achievement, Int. J. Sci. Educ., 30(1), 87–107.
  37. Just M. A. and Carpenter P. A., (1980), A Theory of Reading From Eye Fixations to Comprehension, Psychol. Rev., 87(4), 329–354.
  38. Kelly R. M. and Jones L. L., (2008), Investigating Students’ Ability to Transfer Ideas Learned from Molecular Animations of the Dissolution Process, J. Chem. Educ., 85(2), 303–309.
  39. Kind V., (2004), Beyond Appearances: Students’ Misconceptions about Basic Chemical Ideas, 2nd edn, Durham: Durham University, School of Education.
  40. Kozma R. and Russell J., (2005), Students Becoming Chemists: Developing Representational Competence, in Gilbert J. K. (ed.), Visualization in Science Education, Dordrecht: Kluwer Academic Press, pp. 121–145.
  41. Labinowicz E., (1989), Izvirni Piaget [The Piaget Primer: Thinking, Learning, Teaching], Ljubljana: Državna založba Slovenije.
  42. Levy S. T. and Wilinsky U., (2009), Crossing Levels and Representations: The Connected Chemistry (CC1) Curriculum, J. Sci. Educ. Tech., 18(3), 224–242.
  43. Lewis S. and Lewis J., (2007), Predicting at-risk Students in General Chemistry: Comparing Formal Thought to a General Achievement Measure, Chem. Educ. Res. Pract., 8(1), 32–51.
  44. Lin Y. I., Son J. Y. and Rudd J. A., (2016), Asymmetric Translation between Multiple Representations in Chemistry, Int. J. Sci. Educ., 38(4), 644–662.
  45. Mason M., Pluchino P., Tornatora M. C. and Ariasi N., (2013), An Eye-Tracking Study of Learning From Science Text With Concrete and Abstract Illustrations, J. Exp. Educ., 81(3), 356–384.
  46. Mueller S. T. and Piper B. J., (2014), The Psychology Experiment Building Language (PEBL) and PEBL Test Battery, J. Neurosci. Meth., 222, 250–259.
  47. Mulford D. R. and Robinson W. R., (2002), An Inventory for Alternative Conceptions among First-semester General Chemistry Students, J. Chem. Educ., 79(6), 739–744.
  48. Nicoll G., (2001), A Report of Undergraduates’ Bonding Misconceptions, Int. J. Sci. Educ., 23(7), 707–730.
  49. Osborne J. and Dillon J., (2008), Science Education in Europe: Critical Reflections, London: King's College London.
  50. Pallant J., (2011), SPSS Survival Manual: a step by step guide to data analysis using SPSS, 4th edn, Crows Nest, NSW, Australia: Allen & Unwin.
  51. Patrick A. O., Kpangban E. and Chibueze O. O., (2007), Motivation Effects on Test Scores of Senior Secondary School Science Students, Stud. Home Comm. Sci., 1(1), 57–64.
  52. Phillips L. M., Norris S. P. and Macnab J. S., (2010), Visualization in Mathematics, Reading and Science Education, Dordrecht: Springer.
  53. Planinšič G., Belina R., Kukman I. and Cvahte M., (2009), Učni načrt, Program srednja šola, Fizika: gimnazija: klasična, strokovna gimnazija, [Curriculum, Program of Secondary School, Physics: Gymnasium: Classical, Professional Gymnasium], Ljubljana: National Education Institute Slovenia.
  54. Raiyn J. and Rayan A., (2015), How Chemicals’ Drawing and Modelling Improve Chemistry Teaching in Colleges of Education, World J. Chem. Educ., 3(1), 1–4.
  55. Randles C. A. and Overton T. L., (2015), Expert vs. Novice: Approached Used by Chemists when Solving Open-ended Questions, Chem. Educ. Res. Pract., 16(4), 811–823.
  56. Rayner K., (2009), Eye Movements and Attention in Reading, Scene Perception, and Visual Search, Q. J. Exp. Psyhol., 62(8), 1457–1506.
  57. Russell J., Kozma R., Jones T., Wykoff J., Marx N. and Davis J., (1997), Use of Simultaneous-synchronized Macroscopic, Microscopic, and Symbolic Representations to Enhance the Teaching and Learning of Chemical Concepts, J. Chem. Educ., 74(3), 330–334.
  58. Scientix, (2016), New Eurydice Report: Students' Underachievement in STEM a Continuing Problem in Europe, retrieved from http://www.scientix.eu/web/guest/news/.
  59. Skvarč M., Glažar S. A., Marhl M., Skribe Dimec D., Zupan A., Cvahte M., Gričnik K., Volčini D., Sabolič G. and Šorgo A., (2011), Učni načrt, Program osnovna šola, Naravoslovje, [Curriculum, Program of Primary School, Science], Ljubljana: National Education Institute Slovenia.
  60. Slapničar M., Devetak I., Glažar S. A. and Pavlin J., (2017), Identification of the Understanding of the States of Water and Air among Slovenian Students aged 12, 14 and 16 Years through Solving Authentic Exercises, J. Balt. Sci. Educ., 16(3), 308–323.
  61. Slykhuis D. A., Wiebe E. N. and Annetta L. A., (2005), Eye-Tracking Students' Attention to PowerPoint Photographs in a Science Education Setting, J. Sci. Educ. Technol., 14(5–6), 509–520.
  62. Stipek D., (1998), Motivation to Learn: From Theory to Practice, Boston: Allyn and Bacon.
  63. Susac A., Bubic A., Kaponja J., Planinic M. and Palmovic M., (2014), Eye Movement Reveal Students’ Strategies in Simple Equation Solving, Int. J. Sci. Math. Educ., 12(3), 555–577.
  64. Taber K. S., (2013), Revisiting the Chemistry Triplet: Drawing upon the Nature of Chemical Knowledge and the Psychology of Learning to Inform Chemistry Education, Chem. Educ. Res. Pract., 14(2), 156–168.
  65. Taber K. S., (2014), Constructing active learning in chemistry: concepts, cognition and conceptions, in Devetak I. and Glažar S. A. (ed.), Learning with Understanding in the Chemistry Classroom, Dordrecht, Springer, pp. 5–23.
  66. Taber K. S., (2018), Lost and Found in Translation: Guidelines for Reporting Research Data in an ‘Other’ Language, Chem. Educ. Res. Pract., 19(3), 646–652.
  67. Taber K. S. and García-Franco A., (2010), Learning Processes in Chemistry: Drawing Upon Cognitive Resources to Learn About the Particulate Structure of Matter, J. Learn. Sci., 19(1), 99–142.
  68. Talanquer V., (2011), Macro, Submicro, and Symbolic: The many faces of the chemistry “triplet”, Int. J. Sci. Educ., 33(2), 179–195.
  69. Tien L. T., Teichert M. A. and Rickey D., (2007), Effectiveness of a MORE Laboratory Module in Prompting Students to Revise Their Molecular-level Ideas about Solutions, J. Chem. Educ., 84(1), 175–180.
  70. Tobin K. and Capie W., (1984), The Test of Logical Thinking, J. Sci. Math. Educ. Southeast Asia, 7(1), 5–9.
  71. Torkar G., Veldin M., Glažar S. A. and Podlesek A., (2018), Why Do Plants Wilt? Investigating Students' Understanding of Water Balance in Plants with External Representations at the Macroscopic and Submicroscopic levels, Eurasia J. Math. Sci. Technol. Educ., 14(6), 2265–2276.
  72. Tóth Z. and Kiss E., (2006), Using Particulate Drawings to Study 13–17 Years Olds’ Understanding of Physical and Chemical Composition of Matter as well as the State of Matter, Pract. Theor. Syst. Educ., 1, 109–125.
  73. Tsai M. J., Hou H. T., Lai M. L., Liu W. L. and Yang F. Y., (2012), Visual attention for solving multiple-choice science problem: an eye-tacking analysis, Comput. Educ., 58, 375–385.
  74. Valanides N., (1996), Formal Reasoning and Science Teaching, Sch. Sci. Math., 96(2), 99–107.
  75. Verovnik I., Bajc J., Beznec B., Božič S., Brdar U. V., Cvahte M., Gerlič I. and Munih S., (2011), Učni načrt, Program osnovna šola, Fizika, [Curriculum. Program of Primary School. Physics], Ljubljana: National Education Institute Slovenia.
  76. Williamson V. M. and Abraham M. R., (1995), The Effects of Computer Animation on the Particulate Mental Models of College Chemistry Students, J. Res. Sci. Teach., 32(5), 521–534.
  77. Wu H. and Shah P., (2004), Exploring Visuospatial Thinking in Learning, Sci. Educ., 88(3), 465–492.
  78. Yen M. H. and Yang F. Y., (2016), Methodology and Application of Eye-Tracking Techniques in Science Education, in Chiu M. H. (ed.), Science Education Research and Practices in Taiwan, Springer: Singapore, pp. 249–277.
  79. Yuriev E., Naidu S., Schembri L. S. and Short J. L., (2017), Scaffolding the Development of Problem-Solving Skills in Chemistry: Guiding Novice Students out of Dead Ends and False Starts, Chem. Educ. Res. Pract., 18(3), 486–504.
This journal is © The Royal Society of Chemistry 2019
Click here to see how this site uses Cookies. View our privacy policy here.