Pre-Service Teachers’ Declarative Knowledge of Wave-Particle Dualism of Electrons and Photons: Finding Lexicons by Using Network Analysis
Round 1
Reviewer 1 Report
The paper presents a novel approach to the analysis of written representations and shows the potential of the application of networks in science education. Using so-called lexicons, the authors capture declarative knowledge structures that are otherwise difficult to access. In this way, it contributes to the establishment of networks in science education research, which is still underrepresented.
The strength of the study lies in the immense amount of work on (and the depth of) the network theory background. In addition, the novelty of this approach is also a highlight of the work. However, both highlights cast long shadows, especially on the comprehensibility of the procedure as described in the paper. To guide the reader safely through this article, I would strongly recommend to make some changes. In particular, I would suggest that the introduction should provide a better justification (research gap and benefit) for the entire study, that the relevant network concepts and terms should perhaps be established in theory with examples, that the method should also be made understandable by means of an example from the study, that the boundaries should be clearly stated and, finally, that some implications should be formulated as to how the results could contribute to scientific learning.
For spell check: Please check the text for missing definite or indefinite articles before nouns. Moreover, you switch between american english (e.g., viZualisation) and british english (e.g., categoriSation).
1. Introduction:
In the introduction you should formulate an argument why this study is necessary. Are there problems in learning/explaining quantum physics? Why declarative knowledge? Is there a research gap? How could an approach that differs from previous attempts possibly help to improve explanations in science education?
Instead, you refer to a set of terms that you (partly) explain only in the theory part or that remain incomprehensible in the whole paper (e.g., L32: `ontological categories’, L35: `epistemological shift’, L68: `lexicons’, L87: `shared dimensions’).
In line 71-72 you already describe the sample, what is, by contrast, in detail missed the 3rd section.
Taken together, this makes the introduction difficult to read and to understand, why the reader should get through this paper.
2. Theory:
Lines 95-102 are redundant to what you write in the lines 119-133.
In lines 123-124 you claim that differences between lexical and semantic networks are not crucial. So it is not nessecary to mention it (twice!). On the other hand, in the discussion section (L509-513) you differ between semantic and lexical structure and say, that the semantic structure contains a deeper structure. Please resolve this contradiction.
The aim of the theory is to make clear the meaning and connection of the relevant terms of your paper. But even at the end of this theory part, it is hard for the reader to understand what a „lexicon“ is, how a lexical network looks like, what are the nodes and edges, why is connectivity practical? It would be very helpful to delete the part where you establish differences between approaches and instead make clear the meaning of the terms by giving an example related to the subject (wave-particle-dualism).
3. Research Design and Sample
3.1.: What is the sample size? Was it a lab work with a real measurement, or a simulation, or only given data? What exactly was the task? It is difficult to understand what exactly the students were supposed to do.
3.2.: Please give an example explanation. Else, the reader gets no clue what you are talking about. I.e., what are sub-contexts? What are „relevant aspects“? Therefor you can erase the second paragraph. This is standard research policy. You could put it into one sentence and don’t have to elaborate it in this level of detail.
3.3.: I would recommend placing the research questions between the theory and the research design. Else it is difficult to see, why you’re designing your research as you did. Additionally, since you have no outline of the research gap in the introduction, in remains unclear, why the research questions are relevant. At this point, they seem to fall from the sky. Finally, what do you mean with „what are the lexicons“? What are they ontologically? How are they structured? How do they differ from each other? Please be more concrete in formulating you RQ’s.
4. Methods
Key terms haven’t neither been introduced nor motivated beforehand (e.g., proximity network, communicability centrality, target concepts). What do they stand for?
L191 - 192: So far, benchmarking and comparing to experts research reports so far doesn’t fit to the research questions.
L239: On what criteria this classification is proceeded? This assignment has to be evaluated in terms of reliability, too. If you haven’t done it yet, you should also put it in the limitation section and find an argument, why you haven’t done it yet.
5. Results:
Figure 2 (together with Figure 1, 3 and 4): Please clarify: for what purpose do you create "lexicons", "schematic figures of lexical networks" and "proximity networks"? And, it is difficult to comprehend how these are related.
In line 315 you state that "lexicons 1-7 cover many properties (the octagons are large)". But the lexicons 12 is as big as lexicon 1 and bigger than 4, the lexicon 10 is also bigger than lexicon 4. So why to emphasize the size of the lexicons 1-7, if others are bigger? And, please explain why you compare the size (in this manner).
Again to figure 3 and 4: The lexicon 9 (Müller and Wiesner) is in both cases (photons and electrons) quite small. Isn’t that counter intuitive? One would expect that experts use a bigger lexicon. Could you please clarify that?
In line 366-369 you say that L-lexicons are important. Could you give a reference or a reason for that? Furthermore, please clarify what do you mean with L-lexicons „surface in communication and exchange knowledge“?
5.3 Relevance and Reliability
Here you state something about relevance. You should urgently place this (and maybe another) argument for relevance in the introduction!
In the first paragraph, you do not advocate for the relevance of lexicons, but of analyzing verbal expressions as a key to the knowledge in general. Also in the next two paragraphs, an argument for the relevance of lexicons is still missing.
Line 483: I would challenge the claim that finding verbs, nouns and objects in clauses is not dependent to the subject who does it. It needs training and expertise in linguistics and in many cases objects are only indirectly referred to. For example, sometimes it is hard to decide, to which part of a statement a relative pronouns (whose, which, …) could be assigned to. You have to prove the quality of several steps in your analysis! If you haven’t done it yet, you should put it into the limitations section.
6. Discussion and Conclusions
In this section, I miss a brief overview over the limitations of the study. E.g., the procedure is restricted to research, until an automatized, software-based approach is available for teachers and lecturers. Moreover you should adress the questions of reliability as mentioned before.
And I miss a real conclusion. So far, you more summarize your results. But what can we learn about how to improve science learning in the classroom or in lectures based on your findings? Please include such implications.
Author Response
C1-0: I would suggest that the introduction should provide a better justification (research gap and benefit) for the entire study, that the relevant network concepts and terms should perhaps be established in theory with examples, that the method should also be made understandable by means of an example from the study, that the boundaries should be clearly stated and, finally, that some implications should be formulated as to how the results could contribute to scientific learning.
R1-0: Lexical analysis helps teacher educators to better identify the extent of pre-service teachers lexicons and thereby guides to pay more attention in teaching multifaceted terminology. See also response R1-1 to referee 1.
1. Introduction
C1-1a: In the introduction you should formulate an argument why this study is necessary. Are there problems in learning/explaining quantum physics? Why declarative knowledge? Is there a research gap? How could an approach that differs from previous attempts possibly help to improve explanations in science education?
R1-1a: The introduction up to line 63 briefly summarizes the current stage of research on how students understand wave-particle dualism. It is mostly focusing on mental models. This should be very clear in the original manuscript. We think that longer summary is unnecessary because at least two rather extensive (and overlapping) reviews are already available (refs. 1, 2 also to some degree, ref. 3). Repetition of summaries of existing summaries for those interested of them beyond the scope what is relevant here seems to us unnecessary. From line 64 onwards up to line 91 we discuss what is our viewpoint, and which, as we think, provides somewhat new viewpoints (and this seems not be denied by either of the reviewers). Since the expectation is that every manuscript should contain some novelty or originality, we think that it is unnecessary to point out “research gaps”.
C1-1b: Instead, you refer to a set of terms that you (partly) explain only in the theory part or that remain incomprehensible in the whole paper (e.g., L32: `ontological categories’, L35: `epistemological shift’, L68: `lexicons’, L87: `shared dimensions’).
R1-1b: We think that “ontological category” and “epistemological shift” and terms alike referring to conceptual change and mental models should be very common in the context of our study. For example, most of the references use them without defining them (since they have well enough established meanings). However, we have briefly defined concepts used in the context they first appear. Concerning the lexicons, we think that they are already defined to the necessary details for introduction in lines 67-70. More detailed description is given in 4.3 in lines 255-269, and for shared dimensions in lines 92-96. In addition, shared lexicons in more general sense are discussed in 120-121 as they are related to Kuhn’s views of the role of lexicons.
C1-1c: In line 71-72 you already describe the sample, what is, by contrast, in detail missed the 3rd section. Taken together, this makes the introduction difficult to read and to understand, why the reader should get through this paper.
R1-1c: We have now added the more detailed description of the sample and the task (lines 166-173) in the third section. However, since reviewer 2 asked for short description of the phenomena already in the introduction, we have left (and augmented) the description of the sample in the introduction, too (see response R21-1 comments by reviewer 1).
2. Theory:
C1-2a: Lines 95-102 are redundant to what you write in the lines 119-133.
R1-2a: Thanks for noting that. It is true there is unnecessary repetition. We have now removed the first paragraph.
C1-2b: In lines 123-124 you claim that differences between lexical and semantic networks are not crucial. So it is not nessecary to mention it (twice!). On the other hand, in the discussion section (L509-513) you differ between semantic and lexical structure and say, that the semantic structure contains a deeper structure. Please resolve this contradiction.
R1-2b: That is right; it is enough to discuss it once, as now is done (see lines 132-137). However, it i important to realize that there different views on that question (if interested, Evans provides a good account).
C1-2c: The aim of the theory is to make clear the meaning and connection of the relevant terms of your paper. But even at the end of this theory part, it is hard for the reader to understand what a „lexicon“ is, how a lexical network looks like, what are the nodes and edges, why is connectivity practical? It would be very helpful to delete the part where you establish differences between approaches and instead make clear the meaning of the terms by giving an example related to the subject (wave-particle-dualism).
R1-2c: The theory section 2 here provides the general underpinning and motivation to construct the lexicons. The specific type of lexicon constructed her was already explained in section 4.2. However, we have now underlined better some details of that construction, what the nodes represent and how links are established, see lines 255-260 and 273-278. It is also necessary to carefully pay attention on the Figure 1 which should answer many of the questions (about role of nodes and links) raised by reviewer 1.
3. Research Design and Sample
C1-3a: 3.1: What is the sample size? Was it a lab work with a real measurement, or a simulation, or only given data? What exactly was the task? It is difficult to understand what exactly the students were supposed to do.
R1-3a: 3.1: We have added the information of the sample size (N=14) as well as the more detailed task description in the subsection 3.1 (see lines 166-173). The students were asked to explain the quantum phenomena in these two well-known experiments not to do the experiments themselves.
C1-3b: 3.2: Please give an example explanation. Else, the reader gets no clue what you are talking about. I.e., what are sub-contexts? What are „relevant aspects“? Therefor you can erase the second paragraph. This is standard research policy. You could put it into one sentence and don’t have to elaborate it in this level of detail.
R1-3b: 3.2: We have added one example of an explanation and its sub-contexts in the Appendix A. According to our university’s research policy (National Advisory Board on Research Integrity) and GDPR directive, we must have the mentioned description of research process and data handling.
C1-3c: 3.3: I would recommend placing the research questions between the theory and the research design. Else it is difficult to see, why you’re designing your research as you did. Additionally, since you have no outline of the research gap in the introduction, in remains unclear, why the research questions are relevant. At this point, they seem to fall from the sky.
R1-3c: 3.3: The research questions are in the place they usually appear in other articles. We hope that the additions, which are made in the beginning of the article, would help the readers to understand better the idea of this article.
C1-3d: Finally, what do you mean with „what are the lexicons“? What are they ontologically? How are they structured? How do they differ from each other? Please be more concrete in formulating you RQ’s.
R1-3d: We reformulated the research question to avoid misunderstanding. It is rephrased now as “What kinds of lexicons students use in their writings about wave-particle dualism”. The original one was apparently ill-formulated.
4. Methods
C1-4a: Key terms haven’t neither been introduced nor motivated beforehand (e.g., proximity network, communicability centrality, target concepts). What do they stand for?
R1-4a: This claim is not correct. The key terms are motivated, introduced and explained in the original manuscript in 4.2, where construction of lexicons is explained. However, we understand that the way they are introduced can be made more approachable and we have now re-structured that part of the manuscript.
C1-4b: L191 - 192: So far, benchmarking and comparing to experts research reports so far doesn’t fit to the research questions.
R1-4b: We realize that “benchmarking” was ill-chosen expression here. Rather, it just for comparisons. The point is that authors of reports A and B state their goals explicitly, and we thus know their stance on the wave-particle dualism. The stances are different, that is clear, and that is expected to be reflected on the vocabulary they use. This, as is seen, is the case as shown by our analysis. It should be understood that we do not take stance on the scientific correctness/relevance of the texts A and B, nor on how the level and scope of the expertise of authors is revealed by the content of texts. It is only vocabulary they use, and what it allows to communicate. The lexicons for A and B, as is seen from the manuscript, are discussed on par with students’ lexicons and no special role is granted for them. Therefore, we do not wish to include this as part of research questions.
C1-4c: L239: On what criteria this classification is proceeded? This assignment has to be evaluated in terms of reliability, too. If you haven’t done it yet, you should also put it in the limitation section and find an argument, why you haven’t done it yet.
R1-4c: This is explained in section 4.3 in lines 263-265 and 267-269. Limitations related to this are discussed in section 5.3 (see response R1-5g and R1-6b).
5. Results
C1-5a: Figure 2 (together with Figure 1, 3 and 4): Please clarify: for what purpose do you create "lexicons", "schematic figures of lexical networks" and "proximity networks"? And, it is difficult to comprehend how these are related.
R1-5a: The purpose of lexicons is now explained in better details in revised section 4. As we think, the purpose of figure 1 should be clear and thoroughly explained in lines 255-260. It is difficult for us to know without responses that are more detailed, what makes it difficult to understand and how to clarify it. The role of proximity network for illustrating the connectedness of words is now explained in more detail in section 4. How these are related, is explained in text flowing from line 255 (lexicons), to line 271 (communicability) to operationalized form of lexicons (from line 292) and to proximity network (line 301). We think that the text is now coherent and flows logically, although it takes probably some effort to follow. It might be that decision to relocate details in Appendices may slow down reading, but it has the advantage of not to burden the reader with details which are not immediately necessary (as we think). We hope that the revisions has made it easier to follow the flow of text.
C1-5b: In line 315 you state that "lexicons 1-7 cover many properties (the octagons are large)". But the lexicons 12 is as big as lexicon 1 and bigger than 4, the lexicon 10 is also bigger than lexicon 4. So why to emphasize the size of the lexicons 1-7, if others are bigger? And, please explain why you compare the size (in this manner).
R1-5b: We have now elaborated this better in 324-329. It is not only the size, but what aspects are covered (usually, if many dimensions are covered, the polygons are also large, apparently) and whether or not the lexicons for both electrons and photons are large (not only one of these, like in case of 10). This, we think, was already evident from discussion following from the lines 329 onwards.
C1-5c: Again to figure 3 and 4: The lexicon 9 (Müller and Wiesner) is in both cases (photons and electrons) quite small. Isn’t that counter intuitive? One would expect that experts use a bigger lexicon. Could you please clarify that?
R1-5c: This is true and we well noted. However, we would like to refrain discussing this further, since it would lead discussions to direction which are not necessarily useful here. In our opinion, although the text B by Müller and Wiesner is well known and often cited, from point of view of physics content it is not well argumented and contains - if not misleading ideas- at least gross oversimplifications. Unfortunately, in science education often the expertise in didactics does not come with equally solid expertise in content. But, we hope that we can leave this outside the discussions in the present study.
C1-5d: In line 366-369 you say that L-lexicons are important. Could you give a reference or a reason for that? Furthermore, please clarify what do you mean with L-lexicons „surface in communication and exchange knowledge“?
R1-5d: The importance of L-lexicons derives from the simple notion that if one does not master the vocabulary or the vocabulary is very limited, communication is difficult. Of course, we only know that some students have chosen to express them by using limited vocabulary and do not know if they could have mastered a broader vocabulary. But even so, it is clearly inadequate to discuss the topic and communicate the central ideas. Just imagine how you could make sense of quantum nature of electrons or photons without using words and terms for writing or speaking about quanta or quantization. The important role of lexicons in scientific communication and even in formation of disciplinary groups is extensively discussed by Kuhn (references 16-17). But this we have explained in section 2. It is a bit difficult for us to see why this has remained obscure. However, we have now reminded of this in lines 417-426 in section 5.
5.3 Relevance and Reliability
C1-5e: Here you state something about relevance. You should urgently place this (and maybe another) argument for relevance in the introduction!
R1-5e: We have now restructured this passage (which originally was indeed rather poorly formulated), now in the Introduction, last paragraph lines 102-109, and discussed the relevance from point of view research in 5.3 (lines 517-520), and finally in conclusions from line 578 onwards (see the highlighted additions).
C1-5f: In the first paragraph, you do not advocate for the relevance of lexicons, but of analyzing verbal expressions as a key to the knowledge in general. Also in the next two paragraphs, an argument for the relevance of lexicons is still missing.
R1-5f: The revision which addresses the comment C2-5e should now also answer this question. In several places, we have now underlined that lexicons (in that they contain the key words and terms) are important in:
- a) revealing information of the conceptual models (although indirectly, but what else we could have to start with), as cognitive linguistics like Evans and Langacker have pointed out.
- b) for communication our knowledge by writing and speaking (undoubtedly the major forms of communication)
- c) for teaching and learning, because of a) and b), because in teaching teacher expresses his/her knowledge and ideas based on his/her knowledge by speaking and writing.
It is difficult for us to see, what more is needed to justify the importance of lexicons? Therefore, we disagree with the claim that the argument for the relevance of the lexicons is missing or that we have somehow failed to point out their relevance.
C1-5g: Line 483: I would challenge the claim that finding verbs, nouns and objects in clauses is not dependent to the subject who does it. It needs training and expertise in linguistics and in many cases objects are only indirectly referred to. For example, sometimes it is hard to decide, to which part of a statement a relative pronouns (whose, which, …) could be assigned to. You have to prove the quality of several steps in your analysis! If you haven’t done it yet, you should put it into the limitations section.
R1-5g: This concern is somewhat off the point. As we say, the analysis is based on grammatical analysis of syntax. For this, each major language like English has well-established rules (the grammar) and no disagreement arise about classification for nouns, verbs, adjectives, subjects, objects etc...The concerns of the reviewer seems to be related to relative pronouns and perhaps some more subtle issues (gerunds, perhaps), which are not always easy cases. However, (as we have explained in section 4.1.) only the nouns and verbs are of interest. Therefore, many things are decidedly left out for simplicity, and in order to reduce the complexity and need for interpretation. Still, as we have shown, even such limited analysis is able to differentiate different lexicons.
6. Discussion and Conclusions
C1-6a: In this section, I miss a brief overview over the limitations of the study. E.g., the procedure is restricted to research, until an automatized, software-based approach is available for teachers and lecturers.
R1-6a: That is true, the method is still restricted for research and it would be too much to expect that teachers can use it. It has been suggested (e.g. by Clariana, refs 19-20) that a simplified text analysis could be of practical use, but we have not yet seen such methods used more widely. Suitable software for text in English are available, like NLP in Python, but it takes some expertise to use it. Our sample is not written in English, and the original language has a different structure than English, German or related languages. Therefore, in our case, automatized analysis even more difficult (though possible) by using automated or semi-automated analysis. That is why, and because the sample size is relatively small, we have done analysis manually and only the networks are analyzed formally, by using matrix-analysis. Nevertheless, software exists, and is freely available in different forms. We would like to leave it for practitioners and teacher educators themselves to decide whether or not they feel competent to use such software. Moreover, in no place in our manuscript we claim that school teachers should carry out such analyses, so we think that objections on that basis are somewhat off-the-point. The reviewer must be aware that very few studies in science education research actually intend to develop research methods the teachers could use themselves so we think that this objection is not properly motivated here and thus ignore it.
C1-6b: Moreover you should adress the questions of reliability as mentioned before
R1-6b: For reliability issue, some of them were already answered in C1-5g. Other issues related to reliability are now discussed more thoroughly in the last paragraph of section 5.3, in lines 549-561.
C1-6c: And I miss a real conclusion. So far, you more summarize your results. But what can we learn about how to improve science learning in the classroom or in lectures based on your findings? Please include such implications.
R1-6c: We have added more reflection on conclusions. In the last paragraph of Introduction, we have now outlined briefly the main implications for teacher education. In more details, the implications are discussed in conclusion section starting from line 587 (see the highlighted additions). Of course, the practical implications at present stage are more notions on what aspects the results guide attention, but on basis of our experiences in real teaching as well as research on science education, the teachers as well as researcher of science education should pay more attention on their use of scientific language and proper, normative use of terminology. The study is a step to direction to provide tools for that. However, there are different views among the science educators and researchers what is the needed depth, specialization and accuracy of scientific language to be used in classroom or teacher education. In order to avoid going too deeply in these discussions, we prefer rather to provide evidence than to make recommendations based on that evidence. Therefore, we hope that the implications we have now added is found adequate for the purposes of the manuscript.
Reviewer 2 Report
This manuscript deals with the lexicons involving electrons and protons that pre-service teachers use to explain quantum-physics phenomena. On the one hand, the authors use lexicons as a proxy for mental models; on the other hand, the overlapping of lexicons (or the degree to which they're shared among students) reveal to which extent they're useful for sharing and negotiating meanings. The conclusions of the study support the idea that only a few pre-service teachers use a rich lexicon allowing them to participate in multifaceted discussions, and that many of them lack complex vocabulary about stochasticity or probability.
The study is well conceptualized, the methods are explained in sufficient detail and the results are clearly exposed.
There are parts of the introduction, particularly those referring to the nature of mental models (e.g. as to whether they are parsimonious) that might be unnecessary. On the contrary, a description of the involved phenomenon would enhance comprehension.
The second research question (R2) is unnecessary, as being shared is one of the criteria for being considered for further analysis and not an objective in itself.
As compared to the lengthy and exhaustive description of the analyses that have been conducted, the implications for preservice teachers' training, which would be of more interest to non-specialized readers, is short. I would encourage the authors to be more explicit about the actual affordances of these lexicons and the implications both for teacher practice and training.
Author Response
C2-1. There are parts of the introduction, particularly those referring to the nature of mental models (e.g. as to whether they are parsimonious) that might be unnecessary. On the contrary, a description of the involved phenomenon would enhance comprehension.
R2-1: We have removed unnecessary parts referring to mental models and added short description (in introduction lines 66-70 and in sub-section 3.1 lines 166-173) of the quantum phenomena which vocabularies are studied here.
C2-2: The second research question (R2) is unnecessary, as being shared is one of the criteria for being considered for further analysis and not an objective in itself.
R2-2: The research question R2 is connected to L-lexicons. It is true that in a sense it is continuation of R1 but L-lexicons are cannot be easily recognized just from answers to R1. Instead, combinatorial analysis with results in Figs 5-6 is needed. Therefore, we would like to retain R2 for clarity.
C2-3. As compared to the lengthy and exhaustive description of the analyses that have been conducted, the implications for preservice teachers' training, which would be of more interest to non-specialized readers, is short. I would encourage the authors to be more explicit about the actual affordances of these lexicons and the implications both for teacher practice and training.
R2-3: The implications of this study for science education and science teacher education is that it guides attention on how differently pre-service teachers (and, thus, supposedly also in-service teachers) use scientific vocabulary and terms. In many cases, the lexicons that consist of the terms are so restricted, that by using them many essential and important points of the topics they discuss cannot be adequately addressed. In the revision, in the discussion and conclusions section (starting from line 587, see highlighted additions), we have now added more reflections of the possible implications for science education and teacher training.
Round 2
Reviewer 1 Report
Please see my replies written in italics. I'm fine with everything else. The rework so far improved the paper significantly.
L70: missing word at the end.
C1-1a: In the introduction you should formulate an argument why this study is necessary. Are there problems in learning/explaining quantum physics? Why declarative knowledge? Is there a research gap? How could an approach that differs from previous attempts possibly help to improve explanations in science education?
R1-1a: The introduction up to line 63 briefly summarizes the current stage of research on how students understand wave-particle dualism. It is mostly focusing on mental models. This should be very clear in the original manuscript. We think that longer summary is unnecessary because at least two rather extensive (and overlapping) reviews are already available (refs. 1, 2 also to some degree, ref. 3). Repetition of summaries of existing summaries for those interested of them beyond the scope what is relevant here seems to us unnecessary. From line 64 onwards up to line 91 we discuss what is our viewpoint, and which, as we think, provides somewhat new viewpoints (and this seems not be denied by either of the reviewers). Since the expectation is that every manuscript should contain some novelty or originality, we think that it is unnecessary to point out “research gaps”.
I didn’t asked for a longer summary, but for an argument. Regarding this, you only say that so far, mental models on quantum physics are context dependent. Ok, but is this a problem? And where is the need to investigate this further, and how? For example: Context dependence in explanations is problematic when it comes to …. Thus, we need to know how context dependence in detail looks like. Which parts of knowledge are stable among learners and which are changing? Informations like this would enable researchers and curriculum developers to design more appropriate interventions. Lexicons and a network approach allow for a high resolution analysis of explanations, and thus are suited for this purpose.
C1-4a: Key terms haven’t neither been introduced nor motivated beforehand (e.g., proximity network, communicability centrality, target concepts). What do they stand for?
R1-4a: This claim is not correct. The key terms are motivated, introduced and explained in the original manuscript in 4.2, where construction of lexicons is explained. However, we understand that the way they are introduced can be made more approachable and we have now re-structured that part of the manuscript.
Proximity (network) appears in the first paragraph of section 4 without being explained there or before. The next time it appears in the results section. But anyway, it is ok. But, do you really need „closeness centrality“? It only appears twice throughout the paper.
Author Response
Comment 1: L70: missing word at the end.
Reply 1: Thanks for noting this. The word “knowledge” was missing.
Comment 2: I didn’t asked for a longer summary, but for an argument. Regarding this, you only say that so far, mental models on quantum physics are context dependent. Ok, but is this a problem? And where is the need to investigate this further, and how? For example: Context dependence in explanations is problematic when it comes to …. Thus, we need to know how context dependence in detail looks like. Which parts of knowledge are stable among learners and which are changing? Informations like this would enable researchers and curriculum developers to design more appropriate interventions. Lexicons and a network approach allow for a high resolution analysis of explanations, and thus are suited for this purpose.
Reply 2:The analysis on lexicons reveals the strong context dependence, which is expected that strong metal models exist. We want to point out this context dependency. Those who are interested in more general context independent mental models may continue the work further.
Comment 3: Proximity (network) appears in the first paragraph of section 4 without being explained there or before. The next time it appears in the results section. But anyway, it is ok. But, do you really need „closeness centrality“? It only appears twice throughout the paper.
Reply 3: This is good notion. It is indeed of only minor importance. However, for exactness, we would mention it. We have changed the wording so that closeness is now understood be only a detail.
the text: “The size of each node, on the other hand, represents the closeness centrality [25] of the node, and is roughly proportional to the total communicability of the node in the network.” is replaced with: “The size of the nodes in the lexical proximity network is roughly proportional to the total communicability of the node shown (more exactly, it is proportional to closeness centrality (for definition, see [25]) in the proximity network, but it has high correlation with total communicability).”
In Figure: “The sizes of the nodes correspond to the closeness centrality of the node.” is replaced with “The sizes of the nodes correspond to roughly their total centrality in the network (see main text).”
