Effect of visual anthropomorphic stories on students’ understanding of the particulate nature of matter and anthropomorphic discourse

Suat Türkoguz; İzel Ercan

doi:10.1039/D1RP00057H

View PDF VersionPrevious ArticleNext Article

DOI: 10.1039/D1RP00057H (Paper) Chem. Educ. Res. Pract., 2022, 23, 206-225

Effect of visual anthropomorphic stories on students’ understanding of the particulate nature of matter and anthropomorphic discourse†

Suat Türkoguz *^a and İzel Ercan *^b
^aBuca Faculty of Education, Department of Science Education, Dokuz Eylul University, İzmir, Turkey. E-mail: suat.turkoguz@gmail.com; Tel: +90 505 698 06 34
^bRepublic of Turkey Ministry of National Education, Ankara, Turkey. E-mail: iercan95@hotmail.com; Tel: +90 534 520 10 87

Received 24th February 2021 , Accepted 22nd October 2021

First published on 29th October 2021

Abstract

This study investigated the effect of visual anthropomorphic stories on students’ understanding of the particulate nature of matter and their level of anthropomorphic discourse. This study employed a quasi-experimental research design with pretest and posttest control groups. Science activities supported by visual anthropomorphic stories were conducted in the experimental group, and the 2018–2019 science curriculum was implemented in the control group. Two measurement tools, which included a ‘two-tier diagnostic test for the particulate nature of matter’ and an ‘anthropomorphic discourse usage-level test for the atomic concept’, were used. Consequently, this study showed that the anthropomorphic discourse scores were statistically significant in favor of the students in the experimental group, while their scientific explanation scores were statistically significant in favor of students in the control group. That is, the science activities supported by visual anthropomorphic stories further developed students' anthropomorphic discourse in the experimental group but could not sufficiently develop their scientific explanations. Additionally, this study showed that the scores of the ‘positive tendency’ taxonomy on the particulate nature of matter for the students in both groups were close to each other, but their scores of the ‘negative tendency’ taxonomy were statistically significant differences in favour of the control group. Namely, science activities supported by visual anthropomorphic stories could help students in well understanding concepts of the particulate nature of matter.

Introduction

The word ‘anthropomorphism’ is derived from two Greek words: ‘anthropos’ and ‘morphe’. ‘Anthropos’ means ‘human’; ‘morphe’ means ‘form’ (Epley et al., 2007; Connell, 2013). Anthropomorphism is the attribution of humanistic features (affective characteristics such as intention, desire, motivation, emotion) to non-human objects (animals, toys, geometric shapes, robotic and technological devices, religious and belief agents, etc.) (Kallery and Psillos, 2004; Nakiboğlu and Poyraz, 2006; Epley et al., 2007; Talanquer, 2007; Byrne et al., 2009; Connell, 2013; Gjersoe and Wortham, 2019). Therefore, anthropomorphism is also defined as a process of inference of humanistic features about the unobservable properties of a non-human object. This inference process requires basic cognitive operations, which include the acquisition of information, the activation or retrieval of stored information, and the application of active information to a specific target (Epley et al., 2008). According to information processing theory, first, data obtained from observations in the macroscopic world are collected in sensory memory; second, these data are processed by comparing them with prior knowledge and experiences in short-term memory with limited capacity, and finally, this structured new knowledge is stored in permanent memory (Senemoğlu, 2004). In this process, students can make scientific or anthropomorphic inferences while transferring newly structured information to long-term memory. These inferences depend entirely on the students’ daily experiences, prior knowledge, and macroscopic observations with the senses. Thus, people can learn some phenomena by experiencing them through their observations, and sometimes they make an anthropomorphic inference for unobservable phenomena. Ultimately, students can choose anthropomorphic discourses as an easy unscientific way of explanation to establish a relationship between the observable macro-world and the unobservable micro-world.

Anthropomorphism is common in the language of daily life (Geerdts, 2016). Examples of anthropomorphic explanations are as follows: ‘plants prefer wet soil’ (Friedler et al., 1993), ‘the male peacock wants to impress the female by opening his tail’, ‘animals are different from plants. They have brains and nervous systems, so they think and wish… They are like human beings’ (Tamir and Zohar, 1991), ‘atoms like to share electrons in covalent bonds’, ‘atoms want to buy or share electrons to complete their octet’, and ‘atoms want to be stable’ (Taber and Watts, 1996; Nakiboğlu and Poyraz, 2006). Verbs such as ‘like’, ‘love’, ‘prefer’, and ‘want’ in children's efforts at scientific explanations may be indicators of weak anthropomorphic discourses.

Taber and Watts (1996) classified anthropomorphism into two parts as weak and strong: metaphorical or weak anthropomorphism that refers to desires, feelings and human abilities to transfer ideas in a way similar to a social being; and strong anthropomorphism, which is teleological and is the explanation for the phenomena about life, and the universe with non-existent desires and feelings in the entities to achieve a target state. Teleology is the explanation of facts about life and the universe based on a purpose, and it differs from causal explanations. Examples such as ‘Atoms want to complete their octet with eight electrons to be stable’, ‘In winter, the seeds sleep to protect themselves from the cold’ and ‘The trees rest in the winter to produce new leaves in the spring’ (Kallery and Psillos, 2004) are purely teleological and strong anthropomorphic discourses (Taber and Watts, 1996; Kallery and Psillos, 2004; Epley et al., 2007). A strong anthropomorphic discourse may facilitate understanding of the phenomenon in daily life. People's use of weak anthropomorphic discourse with strong anthropomorphic discourse can gain familiarity in their daily speech and their metaphorical use can increase over time. The metaphorical or weak anthropomorphic discourse may disappear over time as people age (Taber and Watts, 1996). Considering that the weak anthropomorphic discourse disappears over time, the abstract structure of the atom may be explained metaphorically to students through weak anthropomorphic discourse. If students’ strong anthropomorphic discourse flexibly improves toward weak anthropomorphic discourse, perhaps this discourse may also encourage understanding of science concepts.

Advantages and disadvantages of anthropomorphic discourse

The anthropomorphic discourse has both advantages and disadvantages. The advantages of anthropomorphic discourse from the literature are as follows under three headings: education, sociability and communication. First, in terms of education, the use of anthropomorphic discourse in lectures, media such as TV, the Internet, and books may increase students’ participation and motivation in class (Lemke, 1990; Watts and Bentley, 1994; Taber and Watts, 1996; Zohar and Ginossar, 1998; Dorion, 2011; Geerdts, 2016). Using anthropomorphic visuals and scientific language in storytelling may be more beneficial for younger students to understand subjects. However, there are doubts about the usefulness of anthropomorphic discourse in advanced ages – older students can have difficulty understanding the subject and can have mind confusion due to anthropomorphic or metaphorical explanations (Tamir and Zohar, 1991; Friedler et al., 1993; Zohar and Ginossar, 1998; Kallery and Psillos, 2004; Nakiboğlu and Poyraz, 2006; Byrne et al., 2009; Dorion, 2011; Geerdts et al., 2016; Gjersoe and Wortham, 2019).

Second, in terms of sociability, anthropomorphic discourse may help people socialize and communicate effectively. The social environment and culture may naturally influence the usage of this discourse (Taber and Watts, 1996; Epley et al., 2007). Lonely children, in particular, spend more time with toys, cartoons, and pets and can therefore tend to use a more anthropomorphic discourse (Gjersoe and Wortham, 2019). A similar situation may be observed in the elderly (Geerdts, 2016).

Lastly, in terms of communication, students can easily express their emotions and effectively communicate through anthropomorphic discourse (Zohar and Ginossar, 1998).

The disadvantages of anthropomorphic discourse from the literature are as follows under five headings: choosing pseudoexplanations instead of scientific explanations, reducing confidence in scientific knowledge, filling knowledge gaps with incorrect generalizations and inferences, developing alive belief in matter, and misuse of models and visuals. First, pseudoexplanations, which are tautological, teleological and anthropomorphic discourses, are a superficial form of explanation that is characterized by students’ ‘Because’ answers to ‘Why’ type questions to the phenomenon explained in a broader conceptual framework. Students’ pseudoexplanations for the ‘why’ questions include their own beliefs, daily knowledge, repetitive cyclical responses, and limited predictions (Taber and Watts, 2000). Students who have difficulty distinguishing causal and mechanical explanations from anthropomorphic discourses can tend to use alternative concepts (Tamir, 1985).

Second, there is an inverse relationship between anthropomorphic discourse and confidence in scientific knowledge; therefore, the anthropomorphic discourse may more encourage anthropomorphic explanation by reducing students’ confidence in scientific knowledge (Al-Balushi, 2012). For example, if students learn chemical-bonding concepts by overconfidence in anthropomorphic discourses, students can have difficulty in causal and mechanical explanations about these concepts (Kallery and Psillos, 2004), and so they can make incorrect explanations about these concepts.

Third, students can fill their knowledge gaps with anthropomorphic discourses, which cause inaccurate generalizations and inferences (Wood, 2019). For example, if students learn the concept of covalent bonding with an anthropomorphic discourse such as ‘According to the octet rule, if atoms share electrons between each other, covalent bonds are formed’, students can conclude that the electron sharing between atoms is equal in all cases (Taber and Watts, 1996). Similarly, these discourses may cause students to make incorrect inferences about apolar covalent bonds, hydrogen bonds, and metallic bonds (Taber and Watts, 1996; Othman et al., 2008; Bergqvist et al., 2013). Therefore, students can have difficulty in understanding all chemistry concepts based on chemical bonds.

Fourth, anthropomorphic discourse may lead students to believe that atoms and molecules (like animals, bacteria, etc.) are sentient, thinking, and feeling creatures in the same ways that humans are. Some students can believe that ‘all atoms are alive, grow and multiply’ (Griffiths and Preston, 1992; Nakhleh, 1992; Harrison and Treagust, 1996; Özalp and Kahveci, 2015) and, therefore, confuse the atom concept with the cell concept (Roland, 2009). This confusion may discourage students’ understanding of chemical reactions that progress spontaneously, or not (Kelly et al., 2010).

Lastly, the anthropomorphic models and visuals in educational references such as textbooks and animation may create disadvantages. In these educational references, anthropomorphic discourses are encountered in examples such as the atom model, DNA chain model, and benzene ring model (Taber and Watts, 1996; Nakiboğlu and Poyraz, 2006; Tam, 2014). Anthropomorphic discourses and visuals portrayed both emotionally and formally in books, television and media may negatively impact students’ attitudes toward animals. Additionally, excessive use of anthropomorphic discourses and visuals in storytelling may discourage students’ learning (Geerdts, 2016).

Evaluation of the advantages and disadvantages toward the anthropomorphic discourse for science education reveals three different views. These are views rejecting its use, views accepting its use, and views conditionally accepting its use. For example, Taber (2001) considered the anthropomorphic discourse to be an obstacle to learning and hesitated to recommend the use of this discourse by teachers. However, Tamir and Zohar (1991) thought that many scientists and teachers use anthropomorphic discourse with justification and for good reasons. In this dilemma, should an anthropomorphic discourse be accepted in science education? Or should it be rejected? Of course, the answer to this question is not easy. However, in this dilemma, the decision to accept or reject may be related to the awareness of teachers and students of the advantages and disadvantages of anthropomorphic discourse (Taber and Watts, 1996). Although teachers try using scientific language in their class, previous experiences force them to use anthropomorphic discourse. Thus, teachers can unwittingly tend to use anthropomorphic discourse according to scientific language in class (Tamir and Zohar, 1991; Hills, 1995; Waytz et al., 2007; Tam, 2014). The hybrid use of anthropomorphic discourse and scientific language may benefit science lessons. Students can find it difficult to explain scientific concepts, instead they can resort to a metaphorical discourse. Here, teachers can allow students to use anthropomorphic discourse. If concepts have abstract content, teachers can also use this discourse for students to help them visualize or imagine concepts. Later, teachers can continue the lesson by focusing on scientific language with frequent repetitions, without allowing students to fully get used to anthropomorphic discourse (Taber, 2001). Teachers can use anthropomorphic discourse to motivate students to the lesson at the beginning of the subject. Additionally, anthropomorphic teaching materials may motivate students and help them understand the subject.

The role of the particulate nature of matter (PNM) in understanding chemistry

The particulate nature of matter is a concept that has become more important with the development of science and technology (Nakhleh, 1992). The basic subject of the PNM is the atom concept and atom models. In most countries, students first begin learning the atom concept in the PNM unit of the middle school science course (Ayas and Özmen, 2002; Yaseen and Akaygun, 2016). The atom concept plays a major role in chemistry and physics (Harrison and Treagust, 2002; Yaseen, 2012; Eryılmaz Muştu and Özkan, 2019; Hejnová and Králík, 2019). The PNM subject is necessary for the understanding of some basic concepts in daily life (Özmen and Kenan, 2007; Özmen, 2011), such as the structure of matter, states of matter, osmosis, diffusion, solution and solubility, chemical reactions (the relationship between heat, pressure, and volume), chemical bonds (Othman et al., 2008), gases (the relationship between pressure, temperature, and volume), heat and temperature, equilibrium, and electric current (Griffiths and Preston, 1992; Nakhleh, 1992; Adadan et al., 2009; Özmen, 2011). Atomic models, concepts of the electron, proton, and neutron, and the formation of substances by chemical bonds are included in these subjects (Hejnová and Králík, 2019). As long as the students understand the PNM, they can learn concepts such as a molecule, an ion, and chemical-bonding (Othman et al., 2008; Nahum et al., 2010; Yaseen, 2012).

Issues in understanding the PNM

Studies show that students from primary school to university levels have difficulty understanding the PNM concept (Othman et al., 2008; Adadan et al., 2009; Özmen et al., 2009; Badrian et al., 2011; Özmen, 2011; Aydeniz and Kotowski, 2012; Hejnová and Králík, 2019). The reason for this is that the PNM concept is abstract and difficult to visualize, and is taught in a language different from everyday life in schools (Özmen and Kenan, 2007; Stojanovska et al., 2012; Hejnová and Králík, 2019). Another learning obstacle to the concept is the modeling of the PNM in lectures and books with simple representations and analogies based on the historical process, and is the inability to give motion positions in the expressions in these models (Yaseen, 2012; Yaseen and Akaygun, 2016). When the PNM is presented with complicated models and representations in the course, then students can generalize the macroscopic property of the matter to its microscopic property (Taber, 2001). Explaining the PNM by using a different scientific language in physics and chemistry classes is an important problem in understanding the subject (Eryılmaz Muştu and Özkan, 2019). In summary, students of all ages can have a lack of knowledge on the PNM and can incompletely explain the microscopic nature of the PNM. The imperfect-teaching methods and insufficient explanations in scientific language may cause this lack of knowledge. If multiple representations such as the model and animation for the PNM's teaching process are flawly designed, the bridge between the macroscopic and microscopic levels of the PNM may be erroneously constructed. Additionally, the difference in scientific language and discourse in daily life may have a negative effect in connecting macroscopic and microscopic levels.

Anthropomorphic discourse on the PNM

The difficulty of understanding the PNM may concern macroscopic (observable world) and microscopic (invisible world) representations. The simultaneous use of these representations may confuse students who lack prior knowledge by putting a cognitive load on them. Thus, students cannot distinguish these representations and can have difficulty in understanding the PNM (Chandrasegaran et al., 2007; Stojanovska et al., 2012). For example, the lack of knowledge on the PNM, the structure of the atom, and atomic models may rely on the inability to distinguish these representations (Taber, 2001; Othman et al., 2008; Badrian et al., 2011). If students have difficulty making scientific explanations from the macroscopic to the microscopic level or distinguishing these representations, they can tend to use anthropomorphic discourse. For instance, chemical bonding is an abstract subject that requires a microscopic explanation due to its invisible nature, and therefore students can prefer anthropomorphic discourse using observations in the macroscopic world to understand this concept (Miller, 1992; Taber and Watts, 2000). In different examples related to reasoning about the microscopic structure of atoms based on macroscopic observations of matter, students can make anthropomorphic inferences such as ‘atoms are alive’ from these observations (Griffiths and Preston, 1992; Harrison and Treagust, 1996, Özalp and Kahveci, 2011). Students can prefer alternative explanations (i.e. tautological, teleological, confusing causality, etc.) along with anthropomorphic discourses to make effective scientific or causal explanations on the concepts of molecular bonds and atoms (Tamir, 1985; Taber and Watts, 2000). Thus, students’ preference to anthropomorphic discourse in this way may make it difficult to learn these concepts, which are at the center of the PNM. In this context, the current study focuses on examining the effect of anthropomorphic discourse on understanding the PNM.

Learning forms of anthropomorphic discourse

Students can learn the anthropomorphic discourse from school, TV, the Internet, books, magazines, and their family and friends. Anthropomorphic discourse may be transferred in three ways from these references: (1) with only anthropomorphic visuals, (2) with an anthropomorphic explanation added to the real image, and (3) with explanations added to scientific explanations (Geerdts, 2016). Thus, by choosing one of these ways, students can adopt anthropomorphic visuals and explanations that match their characters or seem the features of objects/agents in their daily life from these references. Students can anthropomorphize all human-like objects, both visually and in their behavior. For this, the compatibility of the anthropomorphized visual or explanation with the person may also be a major factor (Gjersoe and Wortham, 2019). The use of anthropomorphic discourse with visuals based on daily life may have a positive effect on understanding the concepts (Geerdts et al., 2016). Characters in animation or concept cartoons mostly contain anthropomorphic images and discourses, and most of their scenarios are related to daily life. These features of concept cartoons are suitable for use in a discussion environment with the argumentation-based learning method (Chin and Teou, 2009; Naylor and Keogh, 2013). Therefore, one of the preferred visual activities in teaching the PNM in the study is concept cartoons.

Concept cartoons are visual drawings in which human or animal characters talk about a phenomenon with daily life through speech bubbles (Keogh and Naylor, 2000). To have an effective discussion in the cartoon, at least two characters should be included (Stephenson and Warwick, 2002). Each character must express a different opinion. The opinion of one character in the cartoon must be correct, and the opinions of other characters must consist of nonscientific (incorrect) explanations based on daily experiences. Cartoons should attract students’ attention, surprise them, and encourage discussion by creating mental conflict in their minds (Keogh and Naylor, 2000; Kabapınar, 2005). While animal characters such as cats, dogs and birds were used in concept cartoons to see the anthropomorphic effect in the experimental group, human characters were used in the other group. Similarly, in concept cartoons, animal characters discussed a daily event with anthropomorphic discourses in the EG, while human characters discussed scientific discourses in the CG. One of these discourses in concept cartoons is correct and the others are incorrect (see Fig. 1). During the class, these cartoons were used by the argumentation-based learning method.


	Fig. 1 Example of an anthropomorphic concept cartoon.

Stories in books have undeniable and great roles. Stories in books, especially in children's storybooks, could help children learn new words in infancy. Alternatively, stories in textbooks could improve students’ attention and imagination (Egan, 1988), motivate the student to the lesson by encouraging learning (Weber, 1993; Rowcliffe, 2004), and make the lesson more interesting (Egan, 1988). The use of anthropomorphic visuals and discourse in the design of storybooks makes the books more interesting and cute. Anthropomorphic discourses and visuals may play great important roles in making stories in books interesting and making children like them. Anthropomorphic stories in books may attract students’ interest in the course, motivating them to expand and shape their original ideas (Banister and Ryan, 2001), and help them remember knowledge (Geerdts et al., 2016). However, students can acquire anthropomorphic beliefs in these stories in books and can interpret knowledge with incorrect generalizations according to examples from daily life (Geerdts, 2016; Geerdts et al., 2016). Consequently, students can learn anthropomorphism from stories consisting of anthropomorphic discourses and visuals in the media or books. Sample cartoons that convey conversations on understanding some concepts of the PNM with an anthropomorphic image of human emotions are given in Fig. 2. In the study, anthropomorphic human emotions are partially included in some posters, cartoons, animations and visualizations related to atom and molecule concepts. These figures were used in activities in experimental implementations.


	Fig. 2 Example of an anthropomorphic cartoon reflecting human emotions on the electron transfer of two-atom models (note: Na is a metal atom and Cl is a gaseous atom).

Visual and verbal representations in teaching

Teaching with multiple representations may both motivate students in lessons and may encourage them to learn concepts. Representations may help learning by completing each other's deficiencies. In other words, in the learning process, the strong side of one representation may counteract the disruptive effect of the deficiency side of the other representation. Mayer (2003) explained that the combined use of visual and verbal representations in class is more effective than using them separately. With multiple representations, students can envision an invisible microscopic phenomenon in their minds and improve their visualization skills (Adadan et al., 2009). According to the dual coding theory of Paivio (1986), external representations (e.g., the word ‘water’ or an image of ‘a glass of water’) may actuate either the verbal or non-verbal (visual) channels in the human brain (Adadan et al., 2009). Visual information and verbal information are operated differently, along distinct channels, which raise the recalling potential of a given item versus the stimulus being coded one way. Moreover, the animation and visualization with imagery enable remembering of verbal material and vice versa, so that both channels should support each other (Perrault et al., 2015). Research on visual representations related to the teaching of molecular models reveals that the use of static or computer-animated visual aids supported by verbal and written explanations facilitates learning of scientific concepts (Özmen and Kenan, 2007; Adadan et al., 2009; Kenan and Özmen, 2012).

Animations are a useful educational aid that helps students in constructing mental models of the PNM in chemistry. Several studies report that students understand concepts better with animation after seeing chemistry experiments in labs or videos (Kelly and Jones, 2007; Kelly, 2014). Similarly, Guzzetti (2000) and Kozma et al. (2000) suggested that animated supplementary materials should be included in lessons. Studies indicate that students make the most mistakes in the atom and molecule concepts in chemistry subjects, and incomplete and complex models or animation in the course cause these mistakes (Tasker, 2017). The animation may help create mental models of chemical concepts that are difficult to learn using different teaching methods. However, verbal and visual representations of animation should be well-planned (Chang et al., 2009; Aydeniz and Kotowski, 2012). General features of animation may affect students’ learning. Some studies report that many students had difficulty in learning the PNM with animation (Kelly and Jones, 2007; Kelly, 2014). The animation has important functions in conveying details in the PNM in chemistry, but it remains unclear as to how to transform ideal and useful models into animation. The ideal animation should be designed in accordance with the relevant content depending on the perceptual feature of the student (Kelly and Hansen, 2017). The animation should convey basic simple elements rather than complex structures and not impose cognitive load on students. Animation that flow too fast or slowly on the screen and whose colors are too bright or dull on the screen may interfere with learning pedagogically (Tasker, 2017). Today, much educational software visualizes the PNM, atomic models, and molecules at macroscopic and microscopic levels. These may help students learn about atomic structures and models at the microscopic level (Yezierski and Birk, 2006). In this context, teachers can use animated or static picture representations together with concept mapping, analogies, concrete models, simulations, laboratory activities, and conceptual change texts for effective teaching of abstract topics such as the PNM.

Purpose

All students from preschool to university can be faced with anthropomorphic discourse in everyday life around them (Watts and Bentley, 1994), and therefore, it may be completely difficult for them to eliminate these discourses (Zohar and Ginossar, 1998). The widespread use of anthropomorphic discourses and visuals in social media, books, and television may encourage anthropomorphic language. Additionally, anthropomorphic discourses and visuals may affect learning subjects and especially understanding scientific concepts of students. In the literature, the usage reasons for anthropomorphic discourse are explained by student interviews in qualitative research. However, the effects of anthropomorphic discourses and visuals on students’ understanding of scientific concepts and their use of anthropomorphic language are rarely examined in experimental research. Therefore, the study examines the effects of activities supported by visual anthropomorphic stories on the 7th-grade students’ understanding of the PNM and their use of anthropomorphic discourse.

Research questions

The study has two problem statements as follows:

1. Do science activities supported by visual anthropomorphic stories improve middle school students’ understanding of the PNM?

(a) Do these activities improve students’ PNM conceptions as measured by their total scores on the instrument?

(b) Do these activities improve students’ performance in visual questions as measured by their visual scores?

2. Do science activities supported by visual anthropomorphic stories improve middle school students’ anthropomorphic discourse related to the PNM?

Method

Content of science curriculum

The science education curriculum has been revised four times in the past two decades in Turkey, in 2005, 2013, 2017, and 2018. The weekly schedule of science lessons was three hours a week in primary schools and four hours a week in middle schools. The science curriculum has included ‘scientific process skills’, ‘life skills’, ‘engineering, and design skills’ along with the basic concepts of ‘earth and universe’, ‘living things and life’, ‘physical events’, and ‘matter and its nature’. The science course-teaching model was a model expanding hierarchically according to the principle of spiral versus linearity advancement. The contents of the ‘matter and nature’ subject were different at each grade: ‘let's get to know matter’ (3rd-grade), ‘properties of matter’ (4th-grade), ‘matter and change’ (5th-grade), ‘matter and heat’ (6th-grade), ‘pure matter and mixtures’ (7th-grade), and ‘matter and industry’ (8th-grade). The topic for the current study was the ‘Pure substance and mixtures’ unit of the ‘Matter and its nature’ subject in the 2018 science program for 7th-grade middle school students. This unit had 16 learning objectives presented in a 28 hour teaching schedule. The main concepts of the unit were atoms (nucleus, proton, neutron, electron), the properties of scientific knowledge, molecules, the elements, symbols of elements, compounds, compound formulae, homogeneous mixtures, solution (solute, solvent), heterogeneous mixtures, dissolution, solubility, evaporation, distillation, domestic liquid–solid waste, and recycling. The PNM concept was also the subject of the ‘Pure substances and mixtures’ unit of the 7th-grade of middle school.

Research design

In this study, a quasi-experimental research design with pretest and posttest control groups was used (Çepni, 2007; Büyüköztürk, 2008; Balcı, 2009). In this context, the effect of science activities supported by visual anthropomorphic stories on students* understanding of the PNM and their use of anthropomorphic discourse for the concept of the atom was investigated. Data were collected during the 8 week experimental implementation process on the ‘PNM’ section of the science course. The pretests were administered in the first week of the study (i.e. the first lesson) and the posttests were administered in the eighth week of the study (i.e. the last lesson). Experimental activities were conducted in the remaining weeks. Data were collected from the ‘two-tier diagnostic test for the PNM’ (T₂DT) and the ‘anthropomorphic discourse usage-level test for the atomic concept’ (ADULT). The quasi-experimental research design of the study is given in Table 1.

Table 1 Quasi-experimental design of the study

Group	Pretest	Treatment	Posttest
Experimental	T ₂ DT, ADULT	Science activities and implementations supported by visual anthropomorphic stories	T ₂ DT, ADULT
Control	T ₂ DT, ADULT	Science activities and implementations in the 2018–2019 Science Education Curriculum	T ₂ DT, ADULT

In the study, the class activities in the experimental group (EG) and control group (CG) were arranged according to the 5E class activity plan, which was a teaching model that includes skills and activities that activate students’ curiosity. There were five stages of this model. These were ‘engagement’, ‘exploration’, ‘explanation’, ‘elaboration’, and ‘evaluation’ (Özmen, 2004). In the EG, science activities and implementations supported by visual anthropomorphic stories were applied to the PNM. In the CG, science activities and implementations of the 2018–2019 Science Education Curriculum for the same subject were implemented. Visual anthropomorphic stories were never used in these activities for the CG. A single teacher conducted the classroom activities in the EG and CG. Therefore, the weekly class activity plans of the EG and CG differed. The six-week class activity plan for the EG and CG according to the stages of the 5E model is given in Table 2.

Table 2 Six-week class activity plan for the EG and CG

Learning objectives	Visual and performance activities used in some stages of the 5E model	Experimental group	Control group
Week 1. Data gathering	Completing the pretests	✓	✓
Week 2. Atomic structure, fundamental particles	Explanation: cartoon stories or comic books	Using weak anthropomorphic discourses followed by scientific explanations with animal figures	Using only scientific explanations with human figures or real images, no anthropomorphic discourse
Week 3. Pure substance, compound, molecule concepts	Exploration: concept cartoons
Week 4. Periodic table, names of some elements, use of symbols	Exploration: poster presentations
	Elaboration: short video animations
Week 5. Solutions, mixtures, their classification	Exploration: concept cartoons
Week 6. Solutions, solvent, solute, factors affecting the dissolution rate	Elaboration: story reading and dramas
Week 7. Separation methods for mixtures	Elaboration: story reading and dramas
Week 8. Data gathering	Completing the posttests	✓	✓

Summarizing Table 2, for the EG, in the ‘engagement’ stage of the 5E model, an interesting question was directed to the students to draw attention to the PNM and their readiness levels were determined. In the ‘exploration’ stage, activities were conducted including concept cartoons and posters supported by visual anthropomorphic stories on the PNM. In the ‘explanation’ stage, scientific concepts were explained with the help of cartoons, or posters supported by visual anthropomorphic stories. In the ‘elaboration’ stage, atom and element models were made from play dough and dramas were enacted based on visual anthropomorphic stories. In the ‘evaluation’ stage, playing cards consisting of atoms and elements were used. The experimental implementations were completed during the 6 week period. To quote an example scenario used in the class activities of the EG: ‘An electron comes to town and wants to go to the cheapest hotel. It goes for the cheapest available. If the cheapest first room is empty, it wants to go there, otherwise it goes to the next room and looks. If the second room is also full, then it goes to a slightly more expensive room. It sees that the room 2p is empty and settles there.’ (Taber and Watts, 1996). The story of this scenario was to teach ‘electrons and energy levels’. The 5E-teaching model given as an example for the EG was similarly structured according to the subject scope without using anthropomorphism for the CG.

Participants

This study was conducted in a state middle school in Izmir. The participants of the study were 7th-grade middle school students. The language of instruction at this school is Turkish. For this reason, all activities and data collection tools were applied in Turkish during the experimental treatment and data collection process. Experimental and control groups were randomly selected from the classes created by the middle school administration from the student registration lists. In the study there were 33 students (20 female, 13 male) in the EG and 30 students (14 female, 16 male) in the CG, making 63 overall.

The Ethics Committee Commission (five experts) of Dokuz Eylül University on 19/12/2018 and the Research and Ethics Permission Commission (five experts) of Izmir Directorate of Turkish National Education allowed the research and implementation to be conducted in any of the five middle schools. The official decision was numbered 12018877-604.01.02-E.25085393 on 27/12/2018. The most appropriate school was selected for the implementation by discussing with the administrators of five schools the number of students in the class (the maximum number of students) and weekly course schedules (suitable for the working hours of the researcher). After deciding on two classes from the selected school, an informative meeting for the parents of the students was organized with the help of the classroom teachers. Studies started after the parents’ permissions were obtained.

Data collection tools

Two-tier diagnostic test for the PNM (T₂DT). Avcı et al. (2018) developed the 33-item T₂DT according to the Science Education Curriculum updated in 2013, and calculated the Cronbach alpha reliability coefficient as 0.91. The content of the 33-item T₂DT included topics of ‘electron arrangement and chemical properties’, ‘chemical bonds’, ‘compounds and formulae’, and ‘mixtures’ in the ‘Structure and properties of matter’ unit of the 7th-grade science curriculum. The 33-item T₂DT was a two-tier diagnostic test. The first tier of items in the test consisted of the question text and choices; the second tier consisted of reasoning choices related to the question text and choices.

Twelve experts (two lecturers each from physics education and chemistry education, two graduate students from chemistry education, and six graduate students from science education) reevaluated the content, face, and construct validity of the 33-item T₂DT according to the science curriculum updated in 2018. The experts examined the test items in the 33-item T₂DT as ‘suitable’, ‘not suitable, should be corrected’, and ‘not suitable’ according to the suitability for middle school students and the updated science curriculum. They wrote their additional views, if any, in the blank space left next to the test item. The item content validity rate (CVR_i) for each T₂DT and the test content validity index (CVI) for all items of the 33-item T₂DT were calculated using Lawshe (1975) formulae:

N_u: number of experts approving the test item, N: number of all experts, i: item code, n: item number.

Experts stated that the content of the middle school science curriculum changed and some items in the 33-item T₂DT were not suitable for this content after their reviews. Some items (items numbered 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 13, 14, 15, 16, 17, 18 and 33 in the 33-item T₂DT) were eliminated in considering expert opinions, CVR_i and CVI. Thus, the 33-item T₂DT was reduced to the 15-item T₂DT. The values of CVR_i and CVI for 15 items selected from the 33-item T₂DT are given in Table 3. The numbers in the first column of Table 3 were the test numbers of the 33-item T₂DT for the selected 15 items. These numbers had irregular order because of the chosen items from the 33-item T₂DT. To avoid any confusion in the reduction from the 33-item T₂DT to the 15-item T₂DT, the selected 15 items are renumbered in the second column in Table 3.

Table 3 CVR _i and CVI values for the 15-item T₂DT

SN	RN	N _u	CVR _i	SN	RN	N _u	CVR _i
a The visual items in the test. b The verbal items in the test. SN: Selected-item numbers from 33-item T₂DT. RN: Renumbered-item numbers for 15-item T₂DT.
12^a	1	10	1.00	26^b	9	9	0.80
19^b	2	10	1.00	27^b	10	9	0.80
20^a	3	9	0.80	28^b	11	9	0.80
21^b	4	9	0.80	29^b	12	9	0.80
22^a	5	9	0.80	30^b	13	9	0.80
23^a	6	9	0.80	31^b	14	10	1.00
24^a	7	9	0.80	32^a	15	9	0.80
25^b	8	9	0.80	—	—	—	—
Total for 15-item T₂DT						138	12.60
Total for 6-visual items^a						55	5.00
Total for 9-verbal items^b						83	7.60

The 15-item T₂DT test included 9-verbal questions and 6 items including a visual. The visual anthropomorphic stories could influence the students’ responses to the visual and verbal items in the 15 item-T₂DT. Therefore, the test was treated as having two parts and validity analyses for both parts were conducted. In this context, the CVR_i and CVI values were calculated by grouping the 6-item visual questions and the 9-item verbal questions in the test. Ayre and Scally (2014) determined the threshold value of the content validity of the 10-person expert as 0.800. In Table 3, considering this threshold value determined by Ayre and Scally (2014), the CVR_i and CVI values were in the appropriate range. For the 15-item T₂DT, the lowest calculated CVR_i was 0.80, and the highest calculated CVR_i was 1.0. The calculated CVI of the 15-item T₂DT was higher than the CVI found in the index list. Additionally, similar results were found for 6-visual and 9-verbal items in the 15-item T₂DT. Therefore, the 15-item T₂DT with the 6-visual and 9-verbal items was used as a valid measurement tool.

For item 20 selected from the 33-item T₂DT (item 3 renumbered for the 15-item T₂DT):

For the 15-item T₂DT with CVR_i >0.800:

For 6-visual items in the 15-item T₂DT with CVR_i >0.800:

For 9-verbal items in the 15-item T₂DT with CVR_i >0.800:

Additionally, false positive and false negative values were also calculated for content validity. Hestenes and Halloun (1995) proposed the definitions of the false positive (FP) and the false negative (FN) as evidence of external validity in tiered diagnostic tests. Hestenes and Halloun (1995) defined the FP as the correct response to the test item with a confident attitude based on an incorrect reason, while they defined the FN as an incorrect response to the test item with a confident attitude based on the correct reason. The false negative for external validity in tiered diagnostic tests should be less than 10 per cent (Gürçay and Gülbaş, 2015). However, it was difficult to reduce the FP in tiered diagnostic tests. Students (who lack knowledge) could have a chance to guess the correct answer in multiple-choice tests and were likely to choose the correct option from the distractors of the test item (Peşman and Eryılmaz 2010). In the scoring of Table 4, the false positive ratios of the EG for the pretest and posttest of the 15-item T₂DT were calculated as 25.7% and 15.2%, respectively. Similarly, the pretest and posttest FN rates for the CG were 24.7% and 12.2%, respectively. The decrease in FP rates in the posttest could indicate that the chance factor decreased. The decrease in the FP could be evidence for the validity of the measurement tool given the views of Peşman and Eryılmaz (2010). Similarly, in the scoring of Table 4, the FN rates of the EG for the pretest and posttest of the 15-item T₂DT were calculated as 14.7% and 4.8%, respectively. Similarly, the FN rates of the CG for the pretest and posttest of the 15-item T₂DT were 11.6% and 1.1%, respectively. False negative rates could indicate that the 15-item T₂DT was a valid tool according to the views of Hestenes and Halloun (1995).

Table 4 The T₂DT classification of response types

First tier	Second tier	Score classification
Correct	Correct	True positive (TP)
Correct	Incorrect	False positive (FP)
Incorrect	Correct	False negative (FN)
Incorrect	Incorrect	True negative (TN)

For the construct validity of the 15-item T₂DT, exploratory factor analysis (EFA) was performed on the data. To reach a sufficient sample size in the EFA, the posttest data in the EG and CG were combined, and the analysis was conducted by the scoring of the TP classification in Hestenes and Halloun's (1995) taxonomy. The exploratory factor analysis was run by the varimax rotation method based on the correlation matrix of principal component analysis. The results of the Kaiser–Meyer–Olkin (KMO) and Bartlett's sphericity tests were examined to determine the suitability of the data with the EFA. Accordingly, the KMO value was 0.765 and the chi-square value of Bartlett's sphericity test was calculated as 242.22 for the T₂DT (p < 0.05) (Pallant, 2007). These results showed that the test was suitable for factor analysis. According to the EFA, the items in the 15-item T₂DT were grouped under three factors with an eigenvalue greater than 1, and these factors explained 48.55% of the total test variance. This rate of variance was above the acceptable amount of 41% (Kline, 1994). Accordingly, items of the first factor included items 3, 7, 9, 10, 11, 12, and 15; the second factor included items 1, 2, 4, 5, 6, and 8; and the third factor included items 13 and 14. The factor load values changed between (0.448) & (0.755), (0.401) & (0.666), and (0.680) & (0.876), respectively. The item contents in the factors were ‘Compound’, ‘Solubility and Solutions’, and ‘Mixtures’, respectively. The Cronbach's alpha reliability coefficients of these factors in the 15-item T₂DT were 0.80, 0.62, and 0.63, respectively. Moreover, the Cronbach alpha reliability coefficient of the reduced 15-item T₂DT was 0.82. Consequently, 18 test items that did not fit the 2018 science curriculum were removed from the 33-item test developed by Avcı et al. (2018), and the T₂DT was rearranged to include 15 items. The sample question items for the 15-item T₂DT are in Appendix 1.

Anthropomorphic discourse usage-level test (ADULT) for the atomic concept

The anthropomorphic discourse usage-level test was a 25-item scale with six choices and was developed by Türkoguz (2018) and the Cronbach alpha reliability coefficient was 0.90. Three of the six choices concerned ‘anthropomorphic discourse’; the other three were about ‘scientific discourse.’ On this scale, students were asked to list the first three choices out of six that were suitable for them. The students numbered the choices suitable for them as 1, 2, 3, and so on, starting from level 1. Sample scoring of the test is shown in Table 5. Sample question items about the ADULT are in Appendix 2. For some items of the ADULT in Appendix 2, bold type font options represented anthropomorphic discourse, and non-bold type font style options represented scientific discourse.

Table 5 The scoring of anthropomorphic and scientific discourses according to the preferred responses

First choice	Second choice	Third choice	Score
Scientific discourse: Sci-Dsc; anthropomorphic discourse: Anth-Dsc.
Sci-Dsc	Sci-Dsc	Sci-Dsc	0
Sci-Dsc	Sci-Dsc	Anth-Dsc	1
Sci-Dsc	Anth-Dsc	Sci-Dsc	1
Anth-Dsc	Sci-Dsc	Sci-Dsc	1
Sci-Dsc	Anth-Dsc	Anth-Dsc	2
Anth-Dsc	Sci-Dsc	Anth-Dsc	2
Anth-Dsc	Anth-Dsc	Sci-Dsc	2
Anth-Dsc	Anth-Dsc	Anth-Dsc	3

According to the ADULT’s scoring in Table 5, its minimum score was 0 (zero) and its maximum score was 75. If the score of the test reached 75, according to this test, the discourse of the student was anthropomorphic, and if it reached 0 (zero), the discourse of the student was scientific.

The same experts who examined the T₂DT evaluated the content, face, and construct validity of the ADULT according to the method of Lawshe (1975). In this context, a seminar was given to experts on anthropomorphic discourses, showing examples from physics, biology, and chemistry (especially related to atom and molecular bonding). Experts evaluated the test items on the ADULT as ‘suitable’, ‘not suitable, should be corrected’, and ‘not suitable’ according to the suitability for middle school students and anthropomorphic discourse, and they wrote their additional views, if any, in the blank space left next to the test item. According to Lawshe (1975), the CVI of the ADULT was 0.91. For the 25-item ADULT, the lowest calculated CVR_i was 0.80, and the highest calculated CVR_i was 1.0. According to the index list created by Ayre and Scally (2014), the threshold value of the content validity of the 10-person expert was 0.800. The calculated CVI of the ADULT was higher than the CVI found in the index list, and therefore, the ADULT was accepted as a valid measurement tool.

For the construct validity of the ADULT, exploratory factor analysis (EFA) was performed on the data. To reach a sufficient sample size in the EFA, the posttest data in the EG and CG were combined, and the analysis was conducted by the scoring in Table 5. The exploratory factor analysis was run by the varimax rotation method based on the correlation matrix of principal component analysis. The results of the Kaiser–Meyer–Olkin (KMO) and Bartlett's sphericity tests were examined to determine the suitability of the data with the EFA. Accordingly, the KMO value was 0.793 and the chi-square value of Bartlett's sphericity test was calculated as 862.69 for the ADULT (p < 0.05) (Pallant, 2007). These results showed that the test was suitable for factor analysis. According to the EFA, the items on the ADULT were grouped under four factors with an eigenvalue greater than 1, and these factors explained 57.77% of the total test variance. The variance rates explained by these factors were 44.77%, 4.55%, 4.30%, and 4.15%, respectively. However, since the rate of variance explained by the first factor was high and the other factors remained below 5%, the scale structure of the ADULT was accepted as a single factor. Therefore, the variance explained by the single-factor ADULT was 44.77%. This rate of variance was above the acceptable amount of 41% (Kline, 1994). The single factor structure of the ADULT included 25 items. The lowest factor load of the items in the single-factor ADULT was 0.517, and the highest factor load was 0.789. The test items in the single-factor ADULT measured the variation between anthropomorphic discourse and scientific discourse. Moreover, the Cronbach alpha reliability coefficient of the single-factor ADULT was 0.94 for this study. The sample question items for the single-factor ADULT are in Appendix 2.

Data analysis and normality of tests

Hestenes and Halloun (1995) defined the FP and FN for two-tiered diagnostic tests. We created the T₂DT classification of response types in Table 4 for test scoring of responses according to these definitions. Alternatively, we modified Table 4 according to the T₂DT's test-pairing taxonomy in Table 6 to better see the score changes between the pretest and posttest of the T₂DT. The new scoring classification for Table 4 is given in Table 6.

Table 6 T ₂ DT's test-pairing taxonomy showing the score changes between the pretest and posttest

Pretest	Posttest	Test-pairing taxonomy
TP	TP	Scientific knowledge (full-understanding of concepts after the pretest)
TN	TP	Positive tendency (good-understanding of concepts after the pretest)
FP or FN	TP
TP	TN	Negative tendency (poor understanding of concepts after the pretest)
TP	FP or FN
FP or FN	TN	Guess or lack of knowledge (non-understanding of concepts after the pretest)
FP or FN	FP or FN
TN	FP or FN
TN	TN

The normal distribution results of the Shapiro–Wilk (S–W) test for the pairing taxonomy of the T₂DT between the pretest and the posttest in the EG and CG are given in Table 7. In normality analysis of the T₂DT's data, if the p-value of the S–W test was greater than 0.05, we assumed that the data had a normal distribution.

Table 7 S–W normal distribution values according to the T₂DT's test-pairing taxonomy

T ₂ DT 's test-pairing taxonomy	Group	n	S–W	p-Value
a If the p-value is greater than 0.05, the data have a normal distribution.
Scientific knowledge	EG	33	0.853	0.000
Scientific knowledge	CG	30	0.861	0.001
Positive tendency	EG	33	0.971	0.516^a
Positive tendency	CG	30	0.958	0.279^a
Negative tendency	EG	33	0.595	0.000
Negative tendency	CG	30	0.713	0.000
Guess	EG	33	0.948	0.116^a
Guess	CG	30	0.935	0.068^a

In the analysis results in Table 7, the T₂DT's ‘positive tendency’ and ‘guess’ taxonomies were normally distributed in the EG and CG (p < 0.05). However, the other ‘scientific knowledge’ and ‘negative tendency’ taxonomies were not normally distributed (p > 0.05) (Büyüköztürk, 2008; Razali and Wah, 2011). For this reason, the S–W test was used because the sample size of the EG and CG was less than 50 people, and the data in the ‘scientific knowledge’ and ‘negative tendency’ taxonomies were not normally distributed. Additionally, the discrete values of the data according to the test-pairing taxonomy and the grouping of the T₂DT test items under sub-topics led us to use the Mann–Whitney U (MWU) test in comparisons between the EG and CG.

Inspired by the test-pairing taxonomy approach on the T₂DT, we created a test-pairing taxonomy to better see the score changes in the pretest and posttest of the ADULT in the EG and CG (see Table 8).

Table 8 ADULT 's test-pairing taxonomy showing the score changes between the pretest and posttest

Score difference between the pretest and posttest of the ADULT	Test-pairing taxonomy
Difference < 0	Scientific discourse direction
Difference = 0	Static discourse
Difference > 0	Anthropomorphic discourse direction

The normal distribution of the test-pairing taxonomies of the ADULT was analysed in a way similar to the test-pairing taxonomy of the T₂DT. In Table 9, the analysis results showed that the ADULT's ‘static discourse’ and ‘anthropomorphic discourse direction’ taxonomies in the CG and the ADULT's ‘scientific discourse direction’ taxonomy in the EG were not normally distributed (p < 0.05). According to Table 9, we observed that the ‘scientific discourse direction’ and ‘static discourse’ taxonomies for the EG and the ‘anthropomorphic discourse direction’ for the CG were normally distributed (p > 0.05) (Büyüköztürk, 2008; Razali and Wah, 2011). In Table 9, the results showed that, except for these three taxonomies, the other taxonomies in both the EG and CG were not normally distributed. Therefore, these normality analysis results indicated that the MWU test was suitable for comparing ADULT data with the EG and CG.

Table 9 S–W normal distribution values according to the ADULT's test-pairing taxonomy

ADULT 's test-pairing taxonomy	Group	n	S–W	p-Value
a If the p-value is greater than 0.05, the data have a normal distribution.
Scientific discourse direction	EG	33	0.902	0.006
Scientific discourse direction	CG	30	0.969	0.502^a
Static discourse	EG	33	0.969	0.445^a
Static discourse	CG	30	0.904	0.010
Anthropomorphic discourse direction	EG	33	0.978	0.736^a
Anthropomorphic discourse direction	CG	30	0.898	0.008

In comparisons between the EG and CG, if the p-value of the MWU test was less than 0.05, we assumed that the test result was statistically significant and there was a significant difference between the two groups. Alternatively, to support the statistical significance of these test results, we calculated effect sizes for the MWU test by dividing the absolute (positive) standardized test statistic Z by the square root of the number of pairs as the following formula: (where Z: absolute (positive) standardized value for the U-value of the MWU test, r: correlation coefficient where r assumes the value ranging from −1.00 to 1.00, n₁: sample 1 (sample of the EG), and n₂: sample 2 (sample of the CG)) (Tomczak and Tomczak, 2014).

In interpreting comparisons between the EG and CG, we classified effect sizes as 0.1 (small effect), 0.3 (medium effect), and 0.5 and above (large effect) regarding Cohen's (1992) taxonomy.

Results

The first problem was expressed as ‘Do science activities supported by visual anthropomorphic stories improve middle school students’ understanding of the PNM?’ This problem was examined with three sub-problems according to the total, visual and verbal scores on the T₂DT. Three test scores were calculated to understand how the students responded to the visual and verbal questions in addition to their responses to all the questions on the T₂DT. First, the total scores were calculated using all 15 items on the test. Then, the verbal scores were calculated using nine items (questions without a visual) on the test. Finally, the visual scores were calculated using six items (questions including a picture) on the test. In solving the total scores on the T₂DT for the first problem, the T₂DT's test-pairing taxonomy was analyzed between the EG and CG through the MWU test (see Table 10).

Table 10 MWU test results between the EG and CG for the total scores of the T₂DT's test-pairing taxonomy

Score taxonomies	Group	n	Total f	Mean ranks	MWU test (Z)	p-value	Effect size r
a If the p-value of the MWU test is less than 0.05, the test result is statistically significant and there is a significant difference between the two groups. Alternatively, the small, medium and large effect sizes (r) for MWU test are denoted using letters ‘S’, ‘M’ and ‘L’ for, respectively.
Scientific knowledge	EG	33	61	31.45	477.0 (−0.25)	0.800	0.032
Scientific knowledge	CG	30	62	32.60	477.0 (−0.25)	0.800	0.032
Positive tendency	EG	33	230	34.83	401.5 (−1.30)	0.195	0.164 (S)
Positive tendency	CG	30	183	28.88	401.5 (−1.30)	0.195	0.164 (S)
Negative tendency	EG	33	13	27.89	359.5 (−2.15)	0.032^a	0.271 (S)
Negative tendency	CG	30	39	36.52	359.5 (−2.15)	0.032^a	0.271 (S)
Guess	EG	33	191	33.09	459.0 (−0.50)	0.619	0.063
Guess	CG	30	166	30.80	459.0 (−0.50)	0.619	0.063

The comparative MWU test results between the EG and CG for the total scores of the T₂DT's test-pairing taxonomy are given in Table 10. First, there was no significant difference between the EG and CG for the total scores of the ‘scientific knowledge’, ‘positive tendency’ and ‘guess’ taxonomies. Second, the MWU test showed a statistically significant small effect in favor of the CG compared to the EG for the total scores of the ‘negative tendency’ taxonomy (U = 359.5, Z = −2.15, p < 0.05, r = 0.271). Additionally, the mean ranks of the EG and CG for the total scores of this taxonomy were 27.89 and 36.52, respectively. The box and whisker plot for the total scores of the T₂DT's test-pairing taxonomy is given in Fig. 3.


	Fig. 3 The box and whisker plot for the total scores of the T₂DT's test-pairing taxonomy.

Since this test was suitable for observing the change in students according to the question type, it was analyzed separately in two parts according to visual and verbal question types on the T₂DT. In the first part, the visual scores included 6 items on theT₂DT. In solving of the 6-item visual scores on the T₂DT for the first problem, the T₂DT's test-pairing taxonomy was analyzed between the EG and CG through the MWU test. The MWU test results showed that there was no significant difference between the EG and CG for the 6-item visual scores of all taxonomies of the T₂DT, so we did not need to present numerical values of the analysis results in the table. In the second part, the verbal scores include 9 items on the T₂DT. Similarly, in solving the 9-item verbal scores on the T₂DT for the first problem, the T₂DT's test-pairing taxonomy was analyzed between the EG and CG through the MWU test. The MWU test results showed that a statistically significant small effect between the EG and CG for the 9-item verbal scores of the T₂DT's ‘negative tendency’ taxonomy, so we gave numerical values of the analysis results in Table 11.

Table 11 MWU test results for the 9-item verbal scores of the T₂DT's test-pairing taxonomy

Score taxonomies	Group	n	Total f	Mean ranks	MWU test (Z)	p-value	Effect size r
a If the p-value of the MWU test is less than 0.05, the test result is statistically significant and there is a significant difference between the two groups. Alternatively, the small, medium and large effect sizes (r) for MWU test are denoted using letters ‘S’, ‘M’ and ‘L’ for, respectively.
Scientific knowledge	EG	33	31	29.55	414.0 (−1.17)	0.241	0.147 (S)
Scientific knowledge	CG	30	37	34.70	414.0 (−1.17)	0.241	0.147 (S)
Positive tendency	EG	33	153	35.86	367.5 (−1.77)	0.077	0.223 (S)
Positive tendency	CG	30	109	27.75	367.5 (−1.77)	0.077	0.223 (S)
Negative tendency	EG	33	7	27.83	357.5 (−2.34)	0.019^a	0.295 (S)
Negative tendency	CG	30	24	36.58	357.5 (−2.34)	0.019^a	0.295 (S)
Guess	EG	33	106	32.26	486.5 (−0.12)	0.906	0.015
Guess	CG	30	100	31.72	486.5 (−0.12)	0.906	0.015

The comparative MWU test results between the EG and CG for the 9-item verbal scores of the T₂DT's test-pairing taxonomy are in Table 11. First, there was no significant difference between the EG and CG for the 9-item verbal scores of the ‘scientific knowledge’, ‘positive tendency’ and ‘guess’ taxonomies. Second, the MWU test showed that a statistically significant small effect in favor of the CG compared to the EG for the 9-item verbal scores of the ‘negative tendency’ taxonomy (U = 357.5, Z = −2.34, p < 0.05, r = 0.295). Additionally, the mean ranks of the EG and CG for the 9-item verbal scores of this taxonomy were 27.83 and 36.58, respectively.

The second problem was expressed as ‘Do science activities supported by visual anthropomorphic stories improve middle school students’ anthropomorphic discourse related to the PNM?’ In solving the total scores on the 25-item ADULT for the second problem, the ADULT's test-pairing taxonomy was analyzed between the EG and CG through the MWU test (see Table 12).

Table 12 MWU test results for the 25-item total scores of the ADULT's test-pairing taxonomy

Score taxonomies	Group	n	Total f	Mean ranks	MWU test (Z)	p-value	Effect size r
Scientific discourse: Sci-Dsc; anthropomorphic discourse: Anth-Dsc.a If the p-value of the MWU test is less than 0.05, the test result is statistically significant and there is a significant difference between the two groups. Alternatively, the small, medium and large effect sizes (r) for MWU test are denoted using letters ‘S’, ‘M’ and ‘L’ for, respectively.
Sci-Dsc direction	EG	33	73	19.48	82.0 (−6.39)	0.000^a	0.805 (L)
Sci-Dsc direction	CG	30	306	45.77	82.0 (−6.39)	0.000^a	0.805 (L)
Static discourse	EG	33	229	25.52	281.0 (−2.96)	0.003^a	0.373 (M)
Static discourse	CG	30	291	39.13	281.0 (−2.96)	0.003^a	0.373 (M)
Anth-Dsc direction	EG	33	523	46.03	32.0 (−5.71)	0.000^a	0.719 (L)
Anth-Dsc direction	CG	30	153	16.57	32.0 (−5.71)	0.000^a	0.719 (L)

The comparative MWU test results between the EG and CG for the 25-item total scores of the ADULT's test-pairing taxonomy are given in Table 12. First, the MWU test result on the ‘scientific discourse direction’ showed a statistically significant large effect in favor of the CG compared to the EG (U = 82.0, Z = −6.39, p < 0.05, r = 0.805). Additionally, the mean ranks of the EG and CG for the 25-item total scores of this taxonomy were 19.48 and 45.77, respectively. In this context, the teaching of scientific concepts with the Turkish National Science Education Curriculum in 2018–2019 could encourage the scientific discourse of the CG. Additionally, students in the EG had traces of scientific discourse. This was because the usage of scientific explanations after anthropomorphic discourse activities in the EG could cause these traces. Second, the MWU test result on the 25-item total scores of the ‘static discourse’ taxonomy showed that a statistically significant medium effect in favor of the CG compared to the EG (U = 281.0, Z = −2.96, p < 0.05, r = 0.373). Additionally, the mean ranks of the EG and CG for the 25-item total scores of this taxonomy were 25.52 and 39.13, respectively. In this context, students preferred to use both scientific and anthropomorphic discourses together, but slightly more in the CG. Lastly, the MWU test result for the 25-item total scores of the ‘anthropomorphic discourse direction’ taxonomy showed that a statistically significant large effect in favor of the EG compared to the CG (U = 32.0, Z = −5.71, p < 0.05, r = 0.719). Additionally, the mean ranks of the EG and CG for the 25-item total scores of this taxonomy were 46.03 and 16.57, respectively. In this context, the teaching of scientific concepts with visual anthropomorphic stories could encourage the anthropomorphic discourse of the EG. However, although the CG did not use anthropomorphic discourse activities, students had traces of anthropomorphic discourse. The daily anthropomorphic discourses of their family, television, books and the Internet could cause these traces. Consequently, these results explained that the implementations of science activities supported by visual anthropomorphic stories largely affected students’ anthropomorphic discourse for the PNM. The box and whisker plot for the total scores of the ADULT's test-pairing taxonomy is given in Fig. 4.


	Fig. 4 The box and whisker plot for the 25-item total scores of the ADULT's test-pairing taxonomy.

Discussion and conclusion

This study investigated the effects of visual anthropomorphic stories on students’ understanding of the PNM and their level of using anthropomorphic discourse in 7th-grade science class.

Analysis results related to the first problem of the study showed that there was no significant difference between the EG and CG for the total scores of the T₂DT's ‘scientific knowledge’, ‘positive tendency’, and ‘guess’ taxonomies. The total scores of the T₂DT's ‘negative tendency’ taxonomy were lower than the total scores of the other taxonomies in the EG and CG and there was a statistically significant difference in favor of the CG. Considering the total scores of the T₂DT's ‘negative tendency’, the implementations in the CG caused students to poorly understand concepts about the PNM compared to science activities supported by visual anthropomorphic stories in the EG. Since the results of the general evaluation of the T₂DT between the EG and CG did not provide clear information, we must group the test items of the T₂DT in terms of verbal and visual question types.

There was no significant difference between the EG and CG for the 6-item visual scores of the T₂DT's all taxonomies. Alternatively, we focused on comparison results between the EG and CG regarding the T₂DT's 9-item verbal score for all taxonomies of test-pairing. First, there was no significant difference between the EG and CG for each 9-item verbal score of the T₂DT's ‘scientific knowledge’, ‘positive tendency’, and ‘guess’ taxonomies. However, the 9-item verbal scores of the T₂DT's ‘negative tendency’ taxonomy varied significantly between the EG and CG. There was a statistically significant difference with a small effect in favor of the CG compared to the EG for the 9-item verbal scores of the T₂DT's ‘negative tendency’ taxonomy. Considering the verbal scores of the T₂DT's ‘negative tendency’, the implementations in the CG caused students to poorly understanding of concepts of the PNM compared to science activities supported by visual anthropomorphic stories in the EG. The results showed that the anthropomorphic discourse helped students in the EG a little bit more to well understand concepts of the PNM for verbal items as opposed to visual items. Although the verbal items were more abstract and invisible than the visual items, using anthropomorphic discourse could make them seem less abstract and more concrete, and easier to relate to or understand. According to Gardner (1993), verbal-linguistic intelligence was the most widely used type of intelligence by humans, and its uses by humans with other types of intelligence were at different levels. Considering Gündüz and Özcan (2016), activities such as conversation, discussion, reading, and storytelling influenced verbal-linguistic intelligence alongside visual-spatial, bodily-kinesthetic, and natural intelligence. For example, Banister and Ryan (2001) found that the use of anthropomorphism in storytelling helped primary school children understand the water cycle. In another example, Weber (1993) stated that storytelling with science scenarios made it easy to memorize, and therefore, storytelling in scientific concepts helped children link between cause and effect. Additionally, storytelling could motivate students to learn. For example, Morais et al. (2019) discovered that storytelling was an effective method for motivating young people to learn chemistry. Geerdts et al. (2016) found that students who read visual anthropomorphic storybooks remembered biology concepts better. The results of the study showed that the teaching supported with visual anthropomorphic stories made the EG encouraged to respond to verbal test items than in the CG (see Table 11). Considering the effect of storytelling on verbal-linguistic intelligence, stories supported by visual anthropomorphic discourses were likely to develop this skill of students.

Overall, the T₂DT's results for the total, visual and verbal scores for the ‘negative tendency’ taxonomy showed a small effect in favor of the CG compared to the EG, but this small effect was statistically insignificant for visual scores. Namely, the total and verbal scores of the T₂DT's ‘negative tendency’ taxonomy were a little less likely to appear in the EG. In other words, the implementation in the CG was more likely to move students away from the scientifically valid answer or to cause a poor understanding of concepts of the PNM, compared to implementation in the EG.

Analysis results related to the second problem of the study showed that visual anthropomorphic stories could improve students’ anthropomorphic discourse. In other words, the analysis of the ‘Anthropomorphic Discourse Use Level Test (ADULT)’ showed that there was a large statistically significant effect in favor of the EG compared to the CG for the total scores of the ADULT's ‘anthropomorphic discourse direction’ taxonomy, and additionally, these activities advanced anthropomorphic discourse more than scientific language. Lemke's (1990) view on ‘anthropomorphic discourse reduces the use of scientific language’ was consistent with this study. Further, this result followed stating that the anthropomorphic discourse had a reminiscent feature. This was because students used anthropomorphic discourses as memory tools for understanding complex biological processes and chemical reactions (Watts and Bentley 1994; Zohar and Ginossar, 1998; Al-Balushi 2013; Geerdts et al., 2016). For example, Zohar and Ginossar (1998) stated that anthropomorphic discourses in stories or storytelling could help students remember the facts. Likewise, Geerdts et al. (2016) clarified that the rate of spontaneous anthropomorphic discourse use was 10% in children who read a storybook with factual discourse and images, and 81.8% in children who read a storybook with anthropomorphic discourses and images. Similarly, Al-Balushi (2013) observed that there was an inverse relationship between students’ spatial intelligence and anthropomorphic discourse, and students with low visualization skills preferred anthropomorphic discourse more. In this context, students could use anthropomorphic discourse when they want to visualize or remember the concepts of atoms and molecules in their minds, and therefore, it was likely that there would be an increase in students’ anthropomorphic discourse. In summary, the study confirmed that the students in the EG compared to the CG had high scores of anthropomorphic discourse and low scores of scientific explanation after the implementation process. That was, the science activities supported by visual anthropomorphic stories further developed students' anthropomorphic discourse in the EG, but did not further develop their scientific explanations.

The low-total score of the ADULT's ‘scientific discourse direction’ taxonomy in the EG compared to the CG indicated that it led to little progress of the scientific discourse in the EG since scientific explanations were also included to support the anthropomorphic discourse in the EG. The low-total score of the ADULT's ‘anthropomorphic discourse direction’ taxonomy in the CG compared to the EG pointed to little progress in the anthropomorphic discourse in the EG although an anthropomorphic discourse was never included in the activities of the CG. There were many reasons for this. First, students could not forget anthropomorphic discourses from adolescence to adulthood (Watts and Bentley, 1994), and therefore, it could be completely difficult for them to eliminate these discourses (Zohar and Ginossar, 1998). Second, in daily life, parents could use anthropomorphic discourse to communicate better with their children and reach the mental level of the child (Blown and Bryce, 2017). Third, the mass media, especially children's media, was filled with human-like characters of animals and objects that wear clothing, speak, drive cars, and experience human emotions (Li et al., 2019). Lastly, anthropomorphic characters were frequently encountered in storybooks for children, TV programs and advertisements (Paul, 1996; Marriott, 2002). In the past 50 years, the mass media, in general, and TV, in particular, caused significant cultural changes. The mass media replaced the task of educating children both at school and at home. The mass media had its own rules for presenting scientific topics. The media preferred to use simple language (anthropomorphism) and popular metaphors in their programs to attract the audience to the screen. Especially, the media used anthropomorphic discourses extensively in nature and science movies (and books for similar reasons). In this case, children were likely to encounter unlimited anthropomorphic discourses in popular science movies and books through these mass media (Zohar and Ginossar, 1998). The simultaneous use of anthropomorphic discourses and images in the mass media could anthropomorphize children's daily language more easily (Li et al., 2019). Therefore, the evocative feature of anthropomorphic characters could affect children's preferences for reading books and watching movies (McCrindle and Odendaal, 1994; DeLoache et al., 2010), and their daily language (Geerdts et al., 2016; McGellin et al., 2021). Consequently, although the anthropomorphic discourse was not used in the CG, these low scores on the anthropomorphic discourse direction for the CG showed that students could learn these discourses from their parents, TV, books, and the Internet. From their experiences outside school, students could find an anthropomorphic discourse more attractive than scientific language and easily assimilate this discourse. Students’ adoption of anthropomorphic discourse in this way could negatively affect their choice of using scientific language. The results showed that students preferred an anthropomorphic discourse to scientific language. Accordingly, teachers should pay more attention to the use of scientific language in classroom activities according to the language levels of the students.

In the relevant literature, both the advantages and disadvantages of using anthropomorphic discourse were examined and the results were shared in terms of its contribution to the field. The study highlighted important results in terms of learning the PNM and using scientific language. Overall, the science activities supported by visual anthropomorphic stories helped students to well understand the concepts of the PNM. However, it pointed out that the visual anthropomorphic stories encouraged students’ anthropomorphic discourse, but were not useful in enhancing scientific explanations.

Possible study limitations

This study had some potential limitations. The first limitation was that the 33-item T₂DT, for which ethics and school permissions were obtained officially before the experimental implementations, was reduced to 15 items due to the re-update of the Turkish Science Curriculum at the beginning of data collection. Therefore, the results of the study were limited to the data of the 15-item T₂DT.

The third limitation was that the ADULT's items had anthropomorphic options for atom and molecule concepts on the PNM. Students could learn extra anthropomorphic discourses from their daily lives outside course activities during experimental implementations, and link these discourses with the ADULT's items with a low probability. For example, although there was no anthropomorphic discourse in CG implementations, the increase in the ADULT's anthropomorphic discourse aspect score showed this external noise data effect. In this way, ADULT results were probably somewhat affected by uncontrollable external noise data. In future studies, these could be corrected by adding similar test items related to other science subjects to the ADULT's test items except for the atom concept.

The last limitation was related to the design of the study due to the small sample. The study could have been designed by different combinations of anthropomorphic visuals and discourses in experimental studies organized according to more than two experimental groups and a control group. However, the study examined the joint effect of anthropomorphic discourses and visuals in only one experimental group. The effect level of these alternative experimental combinations on results could not be determined. Perhaps anthropomorphic images could have contributed more than anthropomorphic discourses, or vice versa.

Implications for practice and research

This study has some implications for research and practice. The first three paragraphs explain the implications for the research, and the last four paragraphs reveal the implications for practice on the instruction.

First, the visual anthropomorphic science stories on the PNM can advance students’ anthropomorphism, as well as lead students to believe that atoms and molecules are sentient, thinking, and feeling creatures in the same ways that humans are. Since these activities included images of animals from daily life, these animistic beliefs/feelings might have occurred in the students. However, since the effects of anthropomorphic discourses on students’ conceptual understanding and anthropomorphic discourses on the PNM were investigated in this study, these effects on their beliefs about atomic or molecule susceptibility/emotions could not be observed. Therefore, the formation and structure of animistic beliefs/feelings in students along with anthropomorphic discourse on the PNM should be examined in future research. Additionally, it should be investigated how these animistic beliefs and feelings on the PNM cause misconceptions in students.

Second, this study was conducted with students in the 13–14 age group. Some studies have indicated that the anthropomorphic discourse is common in children and this discourse decreases over time depending on age (Tamir and Zohar, 1991; Friedler et al., 1993; Zohar and Ginossar, 1998; Kallery and Psillos, 2004; Nakiboğlu and Poyraz, 2006; Byrne et al., 2009; Dorion, 2011; Geerdts et al., 2016; Gjersoe and Wortham, 2019). The results of this study, on the other hand, explained that the anthropomorphic discourse of the 13–14 age group students in the EG increased. However, was this discourse learned through experimental instruction permanent or temporary?, did it decrease or increase with age? and did students use it metaphorically? This study could not answer these questions. Future studies should seek answers to these questions.

Third, anthropomorphic discourses can help learn and remember words based on the type and content of visuals (Levin and Lesgold, 1978). Visual representations and animations can increase or decrease students’ attention, or relax them. In short, these can influence students’ attention levels, depending on the intended use. In this study, the effect of visual anthropomorphic discourses on students’ attention levels cannot be controlled; therefore, future studies may need data showing these levels in these activities. For example, Kelly and Hansen (2017) used an eye-tracking tool for determining where students’ attention is focused on the computer screen while watching chemistry-related animation. Similarly, Hansen (2014) examined students’ attention through an eye-tracking tool while performing symbolic and submicroscopic visualizations in some chemistry simulations. Consequently, students’ attention levels can be monitored with remote observations or an eye tracker tool (follows eye movements on the monitor) or NeuroSky (measure the brain's attention level) in future studies.

Fourth, the study can encourage teachers to prepare science activities supported by visual anthropomorphic stories; its results can act as a guide for teachers, and even inspire cartoon producers and animation programmers. The study is an important experimental study in terms of the contribution of anthropomorphic discourse to the conceptual understanding of the PNM and its easy learning compared to scientific language. Students can easily understand the PNM and develop their own scientific language if science activities supported by visual anthropomorphic stories are used together with fun and interesting representations simultaneously with scientific explanations at the engagement stage of the course. The key to conceptually easy understanding of the PNM with science activities supported by visual anthropomorphic stories is hidden in the balanced and simultaneous use of anthropomorphic discourse and scientific language. Instead of weakening the anthropomorphic discourse, which can be learned informally from an early age and cannot be easily abandoned, under intense scientific language pressure in the classroom, allowing simultaneous use of both can help understand the PNM. Maybe these two partnerships can increase students’ interest in the lesson and contribute to their understanding of the PNM.

Fifth, in science courses, teachers can use anthropomorphic visual and verbal teaching materials such as concept cartoons, books, banners, posters and animation at the beginning of the subject to engage students in the PNM. Although anthropomorphic visual and verbal teaching materials were used at every stage of the 5E model in this study, the engagement stage of the model was a more appropriate stage for anthropomorphic discourse. However, these teaching materials should be suitable for the student's age and include popular characters and examples from daily life. Additionally, these materials should be used for macroscopic concepts; otherwise, anthropomorphic visual materials prepared for microscopic concepts may complicate the understanding of the PNM.

Sixth, teachers can use the anthropomorphic discourses and visuals on the PNM as an evaluation activity in the classroom. For example, in the classroom activities related to the PNM, students can write a story with scientific explanations from an anthropomorphic visual, or complete an anthropomorphic story with scientific explanations, or transform anthropomorphic discourses in a story into scientific discourses. Moreover, while teaching the PNM, teachers can leave gaps in the speech bubbles of concept cartoons consisting of anthropomorphic characters in activity pages (maybe textbooks) and ask students to fill in those gaps. In this way, teachers can follow whether students pay attention to scientific discourse on the PNM.

Lastly, there are some tricks that teachers should pay attention to in the use of anthropomorphic discourses on the PNM in science lessons. Teachers should use anthropomorphic discourses on the PNM more carefully in microscopic concepts. If anthropomorphic discourses and visuals are used in microscopic concepts on the PNM, scientific explanations should follow. Otherwise, students can infer different meanings from anthropomorphic visuals and discourses, which are not supported by scientific explanations on the PNM. The study showed that the anthropomorphic discourse is learned more than scientific explanations in a short 8 week practice period in which anthropomorphic science stories on the PNM are frequently used in the EG. Therefore, teachers should adjust the frequency of use of anthropomorphic discourses on the PNM and should not use it too often.

Author contributions

The authors contributed equally to all parts of the article.

Conflicts of interest

There are no conflicts to declare.

Appendix 1. Sample question items about the T₂DT

Sample item 1 renumbered for the 15-item T ₂ DT (item 12 selected from the 33-item T ₂ DT)

Which one/which of the above shapes are in a molecular structure?

(A) 1, 2, 3, (B) 1, 2, 3, 5 (C) 2, 3, 4 (D) 4 (E) 5

I. The transfer of electrons from the same type of atoms forms molecular compounds.

II. The sharing of electrons from the same or different types of atoms forms molecular compounds.

III. The transfer of electrons from different types of atoms forms molecular compounds.

IV. The molecule is the smallest part of matter that retains its properties.

Sample item 4 renumbered for the 15-item T ₂ DT (item 21 selected from the 33-item T ₂ DT)

Pure water (H₂O) we use in our houses, if electrolyzed, decomposes into flammable hydrogen gas and burning oxygen gas. Accordingly, water is

(A) a compound (B) a solution (C) a mixture

I. It consists of different atoms coming together in certain proportions by making chemical bonds.

II. It may be separated into its constituent elements by physical means.

III. It shows the chemical properties of its constituent substances.

IV. It is formed by combining hydrogen and oxygen in the desired proportions.

Sample item 15 renumbered for the 15-item T ₂ DT (item 32 selected from the 33-item T ₂ DT)

The above three experimental setups are a simple electrical circuit consisting of a glass container, battery, switch, lamp holder, and lamp.

The first experimental setup contains pure water, the second salt water, and the third sugar water.

Which of the experimental setups light up the lamp when the switch is turned on?

(A) II (B) III (C) I and III (D) II and III (E) I, II, and III

I. Water molecules decompose sugar into its molecules and surround the sugar molecules.

II. Water molecules decompose the salt into its ions and surround the salt ions.

III. Water molecules decompose the sugar into its ions and surround sugar ions.

IV. Water molecules decompose salt and sugar into their ions and surround their ions.

V. Water molecules conduct electricity, whether there is an ion or not.

Appendix 2. Sample question items about the ADULT

Sample item 1 for the ADULT.	Sample item 2 for the ADULT.
Atoms ……………. lower energy by making more bonds.	Atoms …………… a more ordered structure. Because they complete their valence electrons.
() want to have	() want to have
() are arranged to have	() like to have
() like to have	() love to have
() love to have	() are stable in
() are organized to have	() must have
() are structured to have	() are arranged by
Sample item 3 for the ADULT.	Sample item 4 for the ADULT.
The bond between atoms in compounds is short because atoms always ……. attract electrons to bond.	A sodium atom in table salt ………. one of its electrons for chlorine.
() drag to	() lends
() want to	() gives out
() like to	() sends
() desire to	() shares
() pull to	() donates
() get to	() transfers

Acknowledgements

I want to thank Prof. Dr Mehmet Şahin, Member of Faculty of Education, Dokuz Eylül University, who checked the scientific language and content of the article. I want to thank the administrators of Dokuz Eylül University Institute of Educational Sciences for their support for the thesis on which this article was produced.

References

Adadan E., Irving K. E. and Trundle K. C., (2009), Impacts of multi-representational instruction on high school students’ conceptual understandings of the particulate nature of matter, Int. J. Sci. Educ., 31(13), 1743–1775 DOI:10.1080/09500690802178628.
Al-Balushi S. M., (2012), The relationship between learners’ distrust of scientific models, their spatial ability, and the vividness of their mental images, Int. J. Sci. Math. Educ., 11(3), 707–732 DOI:10.1007/s10763-012-9360-1.
Al-Balushi S. M., (2013), The nature of anthropomorphic mental images created by low and high spatial ability students for different astronomical and microscopic scientific topics, Int. J. Sci. Soc., 4(4), 51–63 DOI:10.18848/1836-6236/cgp/v04i04/59286.
Avcı F., Acar Şeşen B. and Kırbaşlar F. G., (2018), Maddenin Yapısı ve Özelikleri Ünitesine Yönelik İki Aşamalı Teşhis Testinin Geliştirilmesi [The development of two-tier concept test for the structure and properties of matter unit]. Kastamonu Eğitim Fakültesi Dergisi [Kastamonu Education Journal], 26(4), 1007–1019 DOI:10.24106/kefdergi.434239. (In Turkish).
Ayas A. and Özmen H., (2002), A study of students’ level of understanding of the particulate nature of matter at secondary school level, Bogazici Univ. J. Educ., 19(2), 45–60.
Aydeniz M. and Kotowski E. L., (2012), What do middle and high school students know about the particulate nature of matter after ınstruction? implications for practice, Sch. Sci. Math., 112(2), 59–65 DOI:10.1111/j.1949-8594.2011.00120.x.
Ayre C. and Scally A. J., (2014), Critical values for Lawshe's content validity ratio: revisiting the original methods of calculation, Meas. Eval. Couns. Dev., 47 (1), 79–86 DOI:10.1177/0748175613513808.
Badrian A., Abdinejad T. and Naseriazar A., (2011), A cross-age study of Iranian students’ various conceptions about the particulate nature of matter, J. Turkish Sci. Educ., 8(2), 49–63.
Balcı A., (2009), Sosyal bilimlerde araştırma yöntem, teknik ve ilkeler (8.baskı) [Research methods, techniques and principles in social sciences, 8th edn, Ankara: Pegem Yayıncılık [Pegem Publication] (In Turkish).
Banister F. and Ryan C., (2001), Developing science concepts through story-telling, Sch. Sci. Rev., 83(302), 75–84.
Bergqvist A., Drechsler M., De Jong O. and Rundgren S. C., (2013), Representations of chemical bonding models in school textbooks – help or hindrance for understanding? Chem. Educ. Res. Pract., 14, 589–606 10.1039/c3rp20159g.
Blown E. J. and Bryce T. G. K., (2017), Switching between everyday and scientific language, Res. Sci. Educ., 47, 621–653 DOI:10.1007/s11165-016-9520-3.
Büyüköztürk Ş., (2008), Sosyal bilimler için veri analizi el kitabı (9.baskı) [Handbook of data analysis for social sciences, 9th edn, Ankara: PegemA Yayıncılık [Pegem Publication] (In Turkish).
Byrne J., Grace M. and Hanley P., (2009), Children's anthropomorphic and anthropocentric ideas about micro-organisms, J. Biol. Educ., 44(1), 37–43 DOI:10.1080/00219266.2009.9656190.
Çepni S., (2007), Araştırma ve proje çalışmalarına giriş (8.baskı) [Introduction to research and project studies, 8th edn, Trabzon: Celepler Matbaacılık [Celepler Press] (In Turkish).
Chandrasegaran A. L., Treagust D. F. and Mocerino M., (2007), The development of a two-tier multiple-choice diagnostic instrument for evaluating secondary school students’ ability to describe and explain chemical reactions using multiple levels of representation, Chem. Educ. Res. Pract., 8, 293–307 10.1039/B7RP90006F.
Chang H. Y., Quintana C. and Krajcik J. S., (2009), The impact of designing and evaluating molecular animations on how well middle school students understand the particulate nature of matter, Sci. Educ., 94(1), 73–94 DOI:10.1002/sce.20352.
Chin C. and Teou L. Y., (2009), Using concept cartoons in formative assessment: Scaffolding students’ argumentation, Int. J. Sci. Educ., 31 (10), 1307–1332 DOI:10.1080/09500690801953179.
Cohen J., (1992), A power primer, Psychol. Bull., 112(1), 155–159 DOI:10.1037/0033-2909.112.1.155.
Connell P. M., (2013), The role of baseline physical similarity to humans in consumer responses to anthropomorphic animal images, Psychol. Mark., 30(6), 461–469 DOI:10.1002/mar.20619.
DeLoache J. S., Bloom-Pickard M. and LoBue V., (2010), Babies and bears: Human infants’ interest in non-human animals, in McCardle P., McCune, Griffen J. A. and Malhomes V. (ed.), How animals affect us: Examining the influence of human-animal interaction on child development and human health, Washington, DC: National Institutes of Health, pp. 85–95.
Dorion K., (2011), A learner's tactic: How secondary students’ anthropomorphic language may support learning of abstract science concepts, Electron. J. Sci. Educ., 12(2), 1–22.
Egan K., (1988), Primary understanding, 1st edn, Chicago: University of Chicago Press DOI:10.4324/9780203813577.
Epley N., Waytz A. and Cacioppo J. T., (2007), On seeing human: A three-factor theory of anthropomorphism, Psychol. Rev., 114(4), 864–886 DOI:10.1037/0033-295X.114.4.864.
Epley N., Waytz A., Akalis S. and Cacioppo J. T., (2008), When we need a human: Motivational determinants of anthropomorphism, Soc. Cogn., 26(2), 143–155 DOI:10.1521/soco.2008.26.2.143.
Eryılmaz Muştu Ö. and Özkan E. B., (2019), Determining the pre-service teachers’ perceptions of atom and atomic structure through word association test, Asia Pac. Forum Sci. Learn. Teach., 20(1), 1–29.
Friedler Y., Zohar A. and Tamir P., (1993), The effect of age and of learning on the ability to distinguish between anthropomorphic and teleological explanations, Int. J. Sci. Educ., 15(4), 439–443 DOI:10.1080/0950069930150407.
Gardner H., (1993), Frames of mind: The theory of multiple intelligences. New York, NY: Basic Books.
Geerdts M. S., (2016), (Un)Real animals: Anthropomorphism and early learning about animals, Child Dev. Perspect., 10(1), 10–14. DOI:10.1111/cdep.12153.
Geerdts M. S., Van de Walle G. A. and LoBue V., (2016), Learning about real animals from anthropomorphic media, Imagination, Cogn. Pers., 36(1), 5–26 DOI:10.1177/0276236615611798.
Gjersoe N. and Wortham R. H., (2019, April), What behaviours lead children to anthropomorphise robots? Paper presented at the meeting AISB 2019 Symposium on Movement that Shapes Behaviour, Falmouth, UK: Falmouth University, pp. 24–27.
Griffiths A. K. and Preston K. R., (1992), Grade-12 students’ misconceptions relating to fundamental characteristics of atoms and molecules, J. Res. Sci. Teach., 29(6), 611–628 DOI:10.1002/tea.3660290609.
Gündüz N. and Özcan D. (2016) The Development of multiple intelligence with storytelling, Int. J. Educ. Sci., 15(1–2), 242–251 DOI:10.1080/09751122.2016.11890533.
Gürçay D. and Gülbaş E., (2015), Development of three-tier heat, temperature and internal energy diagnostic test, Res. Sci. Technol. Educ., 33(2), 197–217 DOI:10.1080/02635143.2015.1018154.
Guzzetti B. J., (2000), Learning counter-intuitive science concepts: What have we learned from over a decade of research? Reading Writing Q., 16(2), 89–98 DOI:10.1080/105735600277971.
Hansen S. J. R., (2014), Multimodal Study of Visual Problem Solving in Chemistry with Multiple Representations, (Unpublished dissertation), New York, NY, USA: Columbia University DOI:10.7916/D83B5X9H.
Harrison A. G. and Treagust D. F., (1996), Secondary students’ mental models of atoms and molecules: Implications for teaching chemistry, Sci. Educ., 80(5), 509–534 DOI:10.1002/(SICI)1098-237X(199609)80:5<509::AID-SCE2>3.0.CO;2-F.
Harrison A. G. and Treagust D. F., (2002), The particulate nature of matter: challenges in understanding the submicroscopic world, in Gilbert J. K., De Jong O., Justi R., Treagust D. F., Van Driel J. H. (ed.), Chemical Education: Towards Research-based Practice. Science and Technology Education Library, Dordrecht: Springer, vol. 17, pp. 189–212 DOI:10.1007/0-306-47977-X_9.
Hejnová E. and Králík J., (2019, October), Images of atoms in physics textbooks for lower secondary schools vs. misconceptions of pupils about atoms. Paper presented at AIP Conference Proceedings, vol. 2152, (pp. 030006-(1-6), Terchova: Slovakia DOI:10.1063/1.5124750.
Hestenes D. and Halloun I., (1995), Interpreting the force concept inventory: A response to March 1995 critique by Huffman and Heller, Phys. Teach., 33, 502–506 DOI:10.1119/1.2344278.
Hills A. M., (1995), Empathy and belief in the mental experience of animals, Anthrozoös, 8, 132–142 DOI:10.2752/089279395787156347.
Kabapınar F., (2005), Effectiveness of teaching via concept cartoons from the point of view of constructivist approach, Educ. Sci.: Theory Pract. (KUYEB), 5(1), 135–146.
Kallery M. and Psillos D., (2004), Anthropomorphism and animism in early years science: Why teachers use them, how they conceptualise them and what are their views on their use, Res. Sci. Educ., 34(3), 291–311 DOI:10.1023/B:RISE.0000044613.64634.03.
Kelly R. M., (2014), Using Variation theory with metacognitive monitoring to develop insights into how students learn from molecular visualizations, J. Chem. Educ., 91, 1152–1161 DOI:10.1021/ed500182g.
Kelly R. M. and Hansen S. J. R., (2017), Exploring the design and use of molecular animations that conflict for understanding chemical reactions, Quim. Nova, 40(4), 476–481 DOI:10.21577/0100-4042.20170043.
Kelly R. M. and Jones L. L., (2007), Exploring how different features of animations of sodium chloride dissolution affect students’ explanations, J. Sci. Educ. Technol., 16(5), 413–429 DOI:10.1007/s10956-007-9065-3.
Kelly R. M., Barrera J. H. and Mohamed S. C., (2010), An analysis of undergraduate general chemistry students’ misconceptions of the submicroscopic level of precipitation reactions. J. Chem. Educ., 87(1), 113–118 DOI:10.1021/ed800011a.
Kenan O. and Özmen H., (2012), “Maddenin tanecikli yapısı” ünitesine yönelik zenginleştirilmiş bilgisayar destekli bir öğretim materyalinin tanıtımı [Introduction of an enriched computer based teaching material on the particulate nature of matter]. e-J. New World Sci. Acad., 7(1), 269–280. (In Turkish).
Keogh B. and Naylor S., (2000), Teaching and Learning in Science Using concept cartoons: Why Dennis wants to stay in at playtime, Aust. Primary Jr. Sci. J., 16(3), 10–14.
Kline P., (1994), An easy guide to factor analysis, 1st edn, London: Routledge DOI:10.4324/9781315788135.
Kozma R., Chin E., Russell J. and Marx N., (2000), The roles of representations and tools in the chemistry laboratory and their implications for chemistry learning, J. Learn. Sci., 9(2), 105–143 DOI:10.1207/s15327809jls0902_1.
Lawshe C. H., (1975), A quantitative approach to content validity, Pers. Psychol., 28(4), 563–575 DOI:10.1111/j.1744-6570.1975.tb01393.x.
Lemke J. L., (1990), Talking science: Language, learning, and values, 1st edn, Norwood, New Jersey: Ablex Publishing Corporation.
Levin R. and Lesgold A. M., (1978), On pictures in prose, Educ. Commun. Technol., 26, 233–243 DOI:10.1007/bf02766607.
Li H., Eisen S. and Lillard A. S., (2019), Anthropomorphic media exposure and preschoolers’ anthropomorphic thinking in China, J. Child. Media, 13, 149–162 DOI:10.1080/17482798.2019.1570960.
Marriott S., (2002), Red in tooth and claw? Images of nature in modern picture books, Child. Lit. Educ., 33(3), 175–183.
Mayer R. E., (2003), The promise of multimedia learning: Using the same instructional, Learn. Instru., 13(2), 125–139 DOI:10.1016/s0959-4752(02)00016-6.
McCrindle C. M. and Odendaal J. S., (1994), Animals in books used for preschool children, Anthrozoös, 7(2), 135–146 DOI:10.2752/089279394787001998.
McGellin R. T., Grand A. and Sullivan M., (2021), Stop avoiding the inevitable: The effects of anthropomorphism in science writing for non-experts, Public Understanding Sci., 30(5), 621–640 DOI:10.1177/0963662521991732.
Miller L. L., (1992), Molecular anthropomorphism: a creative writing exercise, J. Chem. Educ., 69(2), 141–142 DOI:10.1021/ed069p141.
Morais C., Araújo J. L. and Saúde I., (2019), Awakening to chemistry through storytelling and practical activities: middle school students interacting with pre-school children, Chem. Educ. Res. Pract., 20, 302–315 10.1039/c8rp00096d.
Nahum T. L., Mamlok-Naaman R., Hofstein A. and Taber K. S., (2010), Teaching and learning the concept of chemical bonding, Stud. Sci. Educ., 46(2), 179–207 DOI:10.1080/03057267.2010.504548.
Nakhleh M. B., (1992), Why some students don’t learn chemistry: Chemical misconceptions, J. Chem. Educ., 69(3), 191–196 DOI:10.1021/ed069p191.
Nakiboğlu C. and Poyraz H. E., (2006), Üniversite kimya öğrencilerinin atom ve kimyasal bağlar konularını açıklamada “insana özgü dil” ve “canlılığı” kullanmalarının incelenmesi [Investigation of “anthropomorphism” and “animism” in university chemistry students’ explanations related to atom and chemical bonding]. Kasatamonu Eğitim Dergisi [Kastamonu Education Journal], 14(1), 83–90. (In Turkish).
Naylor S. and Keogh B., (2013), Concept cartoons: what have we learnt? J. Turkish Sci. Educ., 10(1), 3–11.
Othman J., Treagust D. F. and Chandrasegaran A. L., (2008), An investigation into the relationship between students’ conceptions of the particulate nature of matter and their understanding of chemical bonding, Int. J. Sci. Educ., 30(11), 1531–1550 DOI:10.1080/09500690701459897.
Özalp D. and Kahveci A., (2011), Maddenin tanecikli yapısı ile ilgili iki aşamalı tanılayıcı soruların ontoloji temelinde geliştirilmesi [Development of two tier diagnostic items based on ontology in the topic of the particulate nature of matter]. Milli Eğitim Dergisi [National Education Journal], 191, 135–155. (In Turkish).
Özalp D. and Kahveci A., (2015), Diagnostic assessment of student misconceptions about the particulate nature of matter from ontological perspective, Chem. Educ. Res. Pract., 16, 619–639 10.1039/c5rp00096c.
Özmen H., (2004), Fen öğretiminde öğrenme teorileri ve teknoloji destekli yapılandırmacı (constructivist) öğrenme [Learning Theories and Technology Supported Constructivist Learning in Science Teaching]. Turkish Online J. Educ. Technol., 3(1), 0–14. Retrieved from http://www.tojet.net/articles/v3i1/3114.pdf. (In Turkish).
Özmen H., (2011), Turkish primary students’ conceptions about the particulate nature of matter, Int. J. Environ. Sci. Educ., 6(1), 99–121.
Özmen H. and Kenan O., (2007), Determination of the Turkish primary students’ views about the particulate nature of matter, Asia Pac. Forum Sci. Learn. Teach., 8(1), 1–15.
Özmen H., Demircioğlu G. and Coll R. K., (2009), A comparative study of the effects of a concept mapping enhanced laboratory experience on Turkish high school students’ understanding of acid-base chemistry, Int. J. Sci. Math. Educ., 7(1), 1–24 DOI:10.1007/s10763-007-9087-6.
Paivio A., (1986), Mental representations: A dual coding approach, 1st edn, New York: Oxford University Press.
Pallant J., (2007), SPSS survival manual: A step by step guide to data analysis using SPSS, 3rd edn, Buckingham: Open University Press.
Paul E. S., (1996), The representation of animals on children's television, Anthrozoös, 9(4), 169–181 DOI:10.2752/089279396787001400.
Perrault S. T., Lecolinet E., Bourse Y. P., Zhao S. and Guiard Y., (2015, April), Physical Loci: Leveraging spatial, object and semantic memory for command selection, Paper presented at the 33rd Annual ACM Conference on Human Factors in Computing Systems, Seoul: Korea, pp. 299–308 DOI:10.1145/2702123.2702126.
Peşman H. and Eryılmaz A., (2010), Development of a three-tier test to assess misconceptions about simple electric circuits, J. Educ. Res., 103(3), 208–222. DOI:10.1080/00220670903383002.
Razali N. M. and Wah Y. B., (2011), Power comparisons of Shapiro–Wilk, Kolmogorov–Smirnov Lilliefors and Anderson-Darling tests, J. Stat. Model. Anal., 2(1), 21–33. Retrieved from http://www.de.ufpb.br/~ulisses/disciplinas/normality_tests_comparison.pdf.
Roland E. A. E., (2009), An exploratory study of high school students: Conceptions of atomic and cellular structure and conceptions of atomic and cellular structure and relationships between atoms and cells, (Unpublished dissertation), USA: University of Kentucky, Retrieved from https://uknowledge.uky.edu/gradschool_diss/701/.
Rowcliffe S., (2004), Storytelling in science, Sch. Sci. Rev., 86(314), 121–126.
Senemoğlu N., (2004), Gelişim, öğrenme ve öğretim: kuramdan uygulamaya [Development, learning and teaching: from theory to practice], 9th edn, Ankara: Gazi Kitapevi [Gazi Bookstore]. (In Turkish).
Stephenson P. and Warwick P., (2002), Using concept cartoons to support progression in students’ understanding of light, Phys. Educ., 37(2), 135–141 DOI:10.1088/0031-9120/37/2/306.
Stojanovska M. I., Soptrajanov B. T. and Petrusevski V. M., (2012), Addressing misconceptions about the particulate nature of matter among secondary-school and high-school students in the Republic of Macedonia, Creat. Educ., 03(05), 619–631 DOI:10.4236/ce.2012.35091.
Taber K. S., (2001), Building the structural concepts of chemistry: Some considerations from educational research, Chem. Educ. Res. Pract., 2, 123–158 10.1039/b1rp90014e.
Taber K. S. and Watts M., (1996), The secret life of the chemical bond: Students’ anthropomorphic and animistic references to bonding, Int. J. Sci. Educ., 18(5), 557–568 DOI:10.1080/0950069960180505.
Taber K. S. and Watts M., (2000), Learners’ explanations for chemical phenomena, Chem. Educ. Res. Pract., 1(3), 329–353.
Talanquer V., (2007), Explanations and teleology in chemistry education, Int. J. Sci. Educ., 29(7), 853–870 DOI:10.1080/09500690601087632.
Tam K. P., (2014), Anthropomorphism of nature and efficacy in coping with the environmental crisis, Soc. Cogn., 32(3), 276–296 DOI:10.1521/soco.2014.32.3.276.
Tamir P., (1985), Causality and teleology in high school biology, Res. Sci. Technol. Educ., 3(1), 19–28 DOI:10.1080/0263514850030103.
Tamir P. and Zohar A., (1991), Anthropomorphism and teleology in reasoning about biological phenomena, Sci. Educ., 75(1), 57–67 DOI:10.1002/sce.3730750106.
Tasker R., (2017), Molecular-level visualisation for educational purposes, Chem. Modell., 13, 261–282 10.1039/9781782626862-00261.
Tomczak M. and Tomczak E., (2014), The need to report effect size estimates revisited. An overview of some recommended measures of effect size, Trends Sport Sci., 1(21), 19–25.
Türkoguz S., (2018), Quantitative comparison responses of pre-service teachers in social and science education according to anthropomorphism and animism about atomic concept, West. Anatolia J. Educ. Sci., 9(1), 80–101. Retrieved from https://dergipark.org.tr/download/article-file/511467.
Watts M. and Bentley D., (1994), Humanizing and feminising school science: reviving anthropomorphic and animistic thinking in constructivist science education, Int. J. Sci. Educ., 16(1), 83–97 DOI:10.1080/0950069940160106.
Waytz A., Cacioppo J. T. and Epley N., (2007), Effectance as a determinantof anthropomorphism, Unpublished raw data, USA: University of Chicago.
Weber S., (1993), The narrative anecdote in teacher education, J. Educ. Teach., 19(1), 71–82 DOI:10.1080/0260747930190107.
Wood M., (2019), The potential for anthropomorphism in communicating science: Inspiration from Japan, Cult. Sci., 2(1), 23–34 DOI:10.1177/209660831900200103.
Yaseen Z., (2012), A comparison between elementary school students’ mental models and visualizations in textbooks for the concept of atom, Paper presented at AARE APERA International Conference, Sydney: Australian Association for Research in Education (AARE), pp. 1–19.
Yaseen Z. and Akaygun S., (2016), Lise Öğrencilerinin Atom ile İlgili Zihinsel Modellerinin Ders Kitaplarındaki Görseller ile Karşılaştırılması [A comparison of high school students’ mental models on atom and textbook visualizations]. Mehmet Akif Ersoy Üniversitesi Eğitim Fakültesi Dergisi [Mehmet Akif Ersoy University Journal Education Faculty], 1(40), 469–490 DOI:10.21764/efd.39536. (In Turkish).
Yezierski E. J. and Birk J. P., (2006), Misconceptions about the particulate nature of matter. Using animations to close the gender gap, J. Chem. Educ., 83(6), 954–960 DOI:10.1021/ed083p954.
Zohar A. and Ginossar S., (1998), Lifting the taboo regarding teleology and anthropomorphism in biology education - Heretical suggestions, Sci. Educ., 82(6), 679–697 DOI:10.1002/(sici)1098-237x(199811)82:6<679::aid-sce3>3.0.co;2-e.

Footnote

† This study is based on a master thesis conducted in Dokuz Eylül University Institute of Educational Sciences.