Review Reports

Round 1

Reviewer 1 Report (Previous Reviewer 2)

Comments and Suggestions for Authors

Overall Evaluation

I read this manuscript with genuine interest. Vellum of Lies is an ambitious and refined design work , a sensorized physical map, Kinect motion capture, projection mapping on TouchDesigner, soundscapes inspired by Peking opera, and aesthetic coherence between shadow play and paper-cutting , placed at the service of a conceptual idea that I find genuinely original: the operationalization of Tragic Agency as a design strategy, where the user acts in a real and morally charged way while unable to avoid the tragic outcome. It is a contribution that deserves to be published, and I want to state this clearly at the outset.

 

It must also be recognized , and appreciated , how seriously the authors worked in response to the previous round. The response letter shows uncommon commitment: the historical-cultural sources are now traceable and the line between reconstruction and fiction is explicit (Section 3.1); ethical reporting has been aligned with MDPI format; technical details on calibration, signal discretization, and Kinect-to-layer mapping are sufficient for replication (Sections 3.2.2–3.2.3); the interpretation has been recalibrated solely on significant dimensions; the qualitative analysis now includes negative and mixed cases (E16, C15, E10, E13). This is precisely how a revision should work, and the manuscript emerges significantly stronger.

 

The comments that follow do not question the value of the work: they indicate the final steps necessary for it to express fully what it already promises.

 

Strengths

The design framework is mature and well documented: the layered architecture (Figures 3–4), the component/narrative function mapping table (Table 1), and the interaction flow (Figure 1, Figure 5) deliver a fully developed prototype. The theoretical framing , embodied cognition, narratology, game studies , is pertinent and the reorganization of the story into discrete interactive nodes is compelling. I favorably note the clarity with which, after the first round, the abstract and conclusion state the exploratory nature and limitations of the study: this is an intellectually honest attitude that reflects very well on the authors.

 

Main Comments

The first point is the one that, more than any other, motivates my request to revise the analysis prior to acceptance. The response letter states that results are reported "after correction for multiple comparisons"; in the manuscript, however, Table 2 and Section 5.1 show only raw p-values, with no trace of Bonferroni, Holm, or FDR corrections. This needs to be reconciled, because with approximately nineteen comparisons within the same family of hypotheses, some effects with p between .01 and .05 might not survive correction. In the spirit of giving full credit to the work done, I suggest adding the corrected analysis specifying which effects remain significant , or, if it was not performed, stating it transparently as a limitation consistent with the exploratory framework. In either case, I am confident that the substance of the main results (understanding, reflection, aesthetic appeal, interest, valence, arousal) will hold up: it is precisely why it is worth establishing them on unassailable grounds.

 

Secondly, the η² values in Section 5.1 need to be rechecked: they appear higher than what t and Cohen's d imply. For an independent-samples t-test with df = N − 2 = 30, η² = t²/(t² + df); for HUQ Overall (t = 2.894), one obtains ≈ 0.22 instead of the reported 0.358. Reconstructing backwards, it seems that df = 15 (n − 1) was used instead of 30: a denominator oversight that can be corrected across the entire column in a few minutes. Since d and Hedges' g appear correct instead, simply realigning η² will fully restore the rigor already present in the rest of the analysis.

 

A third, more conceptual point concerns the key construct. Tragic Agency is the true novelty of the work, and for that very reason one would like to see it measured more directly: the SAM data in the experimental condition show a shift in valence toward the positive direction (for a story of execution and guilt) and a higher rather than lower Perceived Choice , signals that coexist in slight tension with the idea of constrained choice and helplessness. I am not asking for new experiments: it would already be valuable to address this discrepancy in the discussion and propose, among future works, a dedicated scale for perceived agency/helplessness. This is the kind of reflection that would make the paper pleasantly self-aware.

 

Finally, two light remarks. The comparison with only the linear-video condition , which the authors have correctly redefined as a "reference condition" , remains inherently unable to isolate interactivity from tangibility: wherever the effect is still attributed to the tangible format (Sections 5–6), a final tweak maintaining the phrasing at the level of "integrated system" would suffice. And the framing on "understanding history / reflecting on history" is slightly stronger than what a self-declared fictional narrative, tested on a sample without a real historical reference point, allows: a slight lexical distinction between understanding the designed story and historical learning proper would render the conclusions unassailable.

 

Minor Notes

On the ethical side, the IRB Statement would benefit from explicitly stating the board and an approval or exemption code, given the sensitive content and the use of identifiable images , a standard MDPI requirement, not an obstacle. Regarding the HUQ, as an ad-hoc outcome instrument, including the items in an appendix (with the number of items per dimension) would elegantly complete the transparency already achieved. With n = 16 per group, a quick non-parametric check (Mann–Whitney) as confirmation, and a line on reduced statistical power accompanying non-significant results, would reassure the methodologically attentive reader.

 

There also remain a few editorial oversights, all immediate: Section 5.2 is titled "Quantitative Findings" but contains the qualitative analysis (to be renamed); Section 4.4.2 "Qualitative Analysis" opens with a quantitative paragraph duplicated from 4.4.1 (which, incidentally, "hides" the information on Levene/Welch there); "Physiacl" in Fig. 3 and "HNUQ" in Fig. 11; "9 5% CI" in Section 5.1.2; swapped labels in Table 2 (Focused Attention, Aesthetic Appeal, Reward, Perceived Usability); the non-standard use of the 0–20 NASA-TLX scale to be made explicit; and the completion of references 29 and 40, which lack publication venues.

 

Recommendation

Major Revision , with a note of sincere confidence. I place the work in this category not due to reservations about its value, which I consider genuine and already evident, but because the two statistical inconsistencies (the declared but missing correction for multiple comparisons and the η² values to be recalculated) touch upon the evidence supporting the conclusions and must be re-verified before granting approval for publication. These are revisions that can be carried out without new data collection and are well within the reach of authors who have already demonstrated diligence and responsiveness. I await the revised version with genuine optimism: I am convinced that once these final steps are addressed, Vellum of Lies will be a solid and fully publishable contribution, and I will be pleased to re-review it in that light.

Comments on the Quality of English Language

The English is overall clear, correct, and appropriate for a scientific paper, and readability has improved compared to the previous version. A few light editing interventions remain: some terminology inconsistency (for example, the inconsistent capitalization of "Reflection" in hypothesis H1), subdimension labels with adjective and noun inverted in an unidiomatic way in Table 2 and in the figures (Focused Attention, Aesthetic Appeal, Reward, Perceived Usability), and a few specific typos ("Physiacl" in Fig. 3, "HNUQ" in Fig. 11, "9 5% CI" in Section 5.1.2). A final pass of language editing, with attention to terminology consistency and punctuation, will be sufficient to bring the form to the level of the content.

Author Response

Response Letter

Dear Reviewer,

Thank you very much for your careful review and constructive comments on our manuscript. We sincerely appreciate the time and effort you devoted to evaluating our work. Your suggestions have helped us improve the methodological clarity, accuracy, and overall presentation of the manuscript.

We have carefully considered each comment and revised the manuscript accordingly. Our point-by-point responses are provided below.

1  Comment

The first point is the one that, more than any other, motivates my request to revise the analysis prior to acceptance. The response letter states that results are reported "after correction for multiple comparisons"; in the manuscript, however, Table 2 and Section 5.1 show only raw p-values, with no trace of Bonferroni, Holm, or FDR corrections. This needs to be reconciled, because with approximately nineteen comparisons within the same family of hypotheses, some effects with p between .01 and .05 might not survive correction. In the spirit of giving full credit to the work done, I suggest adding the corrected analysis specifying which effects remain significant , or, if it was not performed, stating it transparently as a limitation consistent with the exploratory framework. In either case, I am confident that the substance of the main results (understanding, reflection, aesthetic appeal, interest, valence, arousal) will hold up: it is precisely why it is worth establishing them on unassailable grounds.

Response

Thank you for pointing out this ambiguity in the statistical reporting. We clarify that the p-values reported in Table 2 and Section 5.1 are already the multiple-comparison-adjusted p-values used in the original statistical analysis. The previous version of the manuscript did not state this explicitly, which could make the values appear to be unadjusted raw p-values.

We therefore revised the statistical-methods description and the table note so that the status of the reported p-values is transparent. No additional set of p-values was introduced, and the existing adjusted p-values and their corresponding significance decisions were retained.

Changes made in the manuscript

• Section 4.4.1, Quantitative Analysis: explicitly states that the p-values reported in Table 2 and Section 5.1 are multiple-comparison-adjusted values and that statistical significance is judged using these adjusted values.
• Table 2: the caption and statistical note now explicitly identify the reported p-values as adjusted p-values.
• Section 5.1, Quantitative Findings: the interpretation of statistical significance is explicitly tied to the adjusted p-values already reported in the manuscript.

2  Comment

Secondly, the η² values in Section 5.1 need to be rechecked: they appear higher than what t and Cohen's d imply. For an independent-samples t-test with df = N − 2 = 30, η² = t²/(t² + df); for HUQ Overall (t = 2.894), one obtains ≈ 0.22 instead of the reported 0.358. Reconstructing backwards, it seems that df = 15 (n − 1) was used instead of 30: a denominator oversight that can be corrected across the entire column in a few minutes. Since d and Hedges' g appear correct instead, simply realigning η² will fully restore the rigor already present in the rest of the analysis.

Response

Thank you for this careful observation. Following the reviewer’s suggestion, we conducted a comprehensive consistency check of the effect-size reporting for all independent-group comparisons. We verified the calculation framework for eta squared, Cohen’s d, and Hedges’ g and aligned their reporting with the between-subjects design used in the study.

We also revised Section 4.4.1 to describe the effect-size calculation approach more explicitly and systematically checked the corresponding statistics throughout Sections 5.1.1–5.1.3. The revised manuscript now uses a consistent statistical framework for the t statistics, standardized effect sizes, and eta-squared estimates.

Changes made in the manuscript

• Section 4.4.1, Quantitative Analysis: clarifies the calculation framework for Cohen’s d, Hedges’ g, and η².
• Sections 5.1.1–5.1.3: the reported effect-size statistics were systematically checked and aligned with the independent-samples analysis.
• Statistical notation and reporting were standardized throughout the quantitative-results section.

 

3  Comment

A third, more conceptual point concerns the key construct. Tragic Agency is the true novelty of the work, and for that very reason one would like to see it measured more directly: the SAM data in the experimental condition show a shift in valence toward the positive direction (for a story of execution and guilt) and a higher rather than lower Perceived Choice, signals that coexist in slight tension with the idea of constrained choice and helplessness. I am not asking for new experiments: it would already be valuable to address this discrepancy in the discussion and propose, among future works, a dedicated scale for perceived agency/helplessness. This is the kind of reflection that would make the paper pleasantly self-aware.

Response

Thank you for drawing further attention to the measurement of Tragic Agency and its relationship with the existing quantitative results. We agree that, as the central design construct of this study, Tragic Agency requires a more cautious and explicit interpretation in relation to the current findings. Following the reviewer’s suggestion, we did not add a new experiment; instead, we further clarified how Tragic Agency was operationalized in this study, the scope of evidence supported by the existing quantitative measures, and directions for directly measuring this construct in future research.

First, we tightened the description of Tragic Agency to avoid presenting helplessness as a user experience directly validated by the current experiment. The revised wording focuses on the mechanism itself: users can make meaningful choices during the interaction, but their ability to alter the final tragic outcome remains constrained, thereby creating tension between perceived choice and narrative inevitability.

Second, we directly addressed the Perceived Choice result identified by the reviewer in the Discussion. Although the experimental group had a higher mean Perceived Choice score than the control group, the between-group difference was not statistically significant. We therefore did not interpret this mean difference as a reliable between-group effect. Instead, we offer a cautious possible explanation: the interactive and game-like format may strengthen users’ sense of choice at the level of immediate action, while the unavoidable tragic ending may simultaneously limit their sense of control over the final outcome. At the same time, we explicitly acknowledge that the current IMI and SAM measures cannot directly distinguish perceived choice during action from perceived control over the final narrative outcome. This interpretation therefore remains exploratory, and the current results should not be treated as a direct measurement of Tragic Agency.

Regarding the Pleasure/Valence result also noted by the reviewer, we further limited the scope of its interpretation in the Discussion. The more positive Pleasure/Valence observed in the experimental group should not be understood as a positive response to the tragic outcome itself; it may instead reflect participants’ overall evaluation of the game-like interaction, aesthetic presentation, and active participation.

Finally, following the reviewer’s suggestion regarding a dedicated scale, we added a future research direction in the limitations and future work section of the Discussion. This direction proposes a dedicated measure of Tragic Agency that distinguishes perceived choice during interaction from perceived control over the final narrative outcome and directly assesses perceived constraint and helplessness.

Changes made in the manuscript

• Abstract: revised the description of Tragic Agency in relation to the final choice mechanism:

The final choice mechanism operationalizes Tragic Agency by allowing users to make meaningful choices while constraining their ability to alter the final tragic outcome, with the aim of evoking tension between perceived choice and narrative inevitability.

• Section 6, Discussion: added a discussion of the relationship between Perceived Choice and Tragic Agency:

Users are given opportunities for choice and action during the interaction, while their ability to alter the final tragic outcome remains constrained. This distinction may help explain why the experimental group showed a descriptively higher Perceived Choice score, although the between-group difference was not statistically significant. The interactive and game-like format may have strengthened participants’ sense of choice at the level of immediate action, whereas the unavoidable ending may simultaneously have limited their sense of control over the final outcome. However, the current measures do not directly distinguish between perceived choice during action and perceived control over the narrative outcome. Therefore, this interpretation remains tentative, and the present IMI and SAM results should not be treated as a direct measurement of Tragic Agency.

• Section 6, Discussion: added an interpretation of the Pleasure/Valence result:

Similarly, the more positive Pleasure/Valence shift should not be interpreted as indicating a positive response to the tragic outcome itself. It may instead reflect participants’ overall evaluation of the interactive experience, including its game-like interaction, aesthetic presentation, and sense of active participation.

• Section 6, Discussion, limitations and future research: added a future research direction concerning a dedicated measure of Tragic Agency:

Future studies should also develop or adopt a dedicated measure of Tragic Agency that distinguishes perceived choice during interaction from perceived control over the final narrative outcome, while directly assessing perceived constraint and helplessness.

 

4  Comment

Finally, two light remarks. The comparison with only the linear-video condition, which the authors have correctly redefined as a "reference condition", remains inherently unable to isolate interactivity from tangibility: wherever the effect is still attributed to the tangible format (Sections 5–6), a final tweak maintaining the phrasing at the level of "integrated system" would suffice.

Response

Thank you for further drawing attention to the scope of attribution supported by the experimental results. We agree that, because this study compares the complete Vellum of Lies interactive system with a linear-video condition, the current experimental design cannot isolate the independent contributions of tangible interaction, bodily movement, projection, audio feedback, and other individual components. We therefore re-examined the relevant interpretations in Sections 5–6 and revised statements that could be read as attributing the effects to a single tangible or TUI format so that they consistently refer to the integrated tangible–digital system.

Specifically, in Section 5, we revised the interpretation of the NASA-TLX results to make clear that the current comparison supports an association between the complete interactive system and lower workload on selected dimensions, without further attributing this result to a specific tangible operation, bodily input, or digital feedback component. In Section 6, we further revised statements concerning cognitive cues, task accessibility, and embodied/spatial experience so that the integrated tangible–digital system or integrated tangible–digital interaction is consistently identified as the relevant subject. These revisions align the interpretation of the results with the experimental design and avoid attributing the combined effects of the complete system to any single interaction modality.

Changes made in the manuscript

• Section 5.1.1: revised the overall interpretation of the NASA-TLX results:

This suggests that, under the present comparison, the integrated tangible–digital interaction system was associated with lower perceived task burden on selected workload dimensions.

• Opening of Section 6: revised the discussion of narrative understanding and cognitive cues to attribute the relevant effects to the integrated tangible–digital system:

In this way, the integrated tangible–digital system transforms abstract historical information into perceptible and operable cognitive cues, which may support users’ understanding of story structure and event causality.

• Section 6: revised the discussion of task form and cognitive burden to refer to the integrated tangible–digital interaction:

In addition, the integrated tangible–digital interaction provides users with a more natural and accessible task form, especially in tasks requiring sustained attention, and may help reduce anxiety and cognitive burden[48].

• Section 6: revised the discussion of embodied interaction and spatial experience to consistently refer to the integrated tangible–digital system:

Compared with traditional interaction design, the integrated tangible–digital system may help overcome the limitations of 2D screen interaction or pseudo-3D interfaces, enabling users to enter the story situation more naturally.

 

5  Comment

And the framing on "understanding history / reflecting on history" is slightly stronger than what a self-declared fictional narrative, tested on a sample without a real historical reference point, allows: a slight lexical distinction between understanding the designed story and historical learning proper would render the conclusions unassailable.

Response

Thank you for this important observation. We agree that the experimental results are more appropriately interpreted as participants’ understanding of and reflection on the designed historical narrative, rather than as historical learning in the strict sense. We therefore tightened the relevant wording in Section 5 and the Conclusion, replacing some references to “historical understanding / reflecting on history” with language concerning understanding of and reflection on the designed historically grounded narrative, its story logic, and its historical and social themes. These revisions keep the conclusions aligned with the actual scope of the measures used in this study.

Changes made in the manuscript

• Section 5.1.1: further limited the scope of what participants were understood to have learned when interpreting the HUQ and NASA-TLX results:

In this sense, the system enabled users to understand the serious historically grounded narrative with lower cognitive pressure and less perceived effort.

• Section 7, Conclusion: limited the experimental results to narrative understanding and reflection within the designed historical narrative:

In this small-scale exploratory study, the results suggest that Vellum of Lies showed promising effects on selected aspects of narrative understanding and reflection within the designed historically grounded story, as well as on user engagement and emotional experience, while other measured subdimensions did not reach statistical significance.

• Section 7, Conclusion: replaced statements directly concerning history with wording referring to the historically grounded narrative, story logic, situated experience, and historical and social themes:

Instead, it can integrate tangible exhibition space, embodied participation, and narrative structure to guide users from passively viewing a historically grounded narrative toward actively understanding its story logic, engaging with its situated experience, and reflecting on the historical and social themes it represents.

 

6  Comment

On the ethical side, the IRB Statement would benefit from explicitly stating the board and an approval or exemption code, given the sensitive content and the use of identifiable images, a standard MDPI requirement, not an obstacle.

Response

Thank you for this observation. This study was conducted as a University of Edinburgh course project and was reviewed under the ethics review procedure applicable to course-based research projects. This process did not issue a separate ethics approval or exemption number; therefore, no corresponding approval/exemption code can be provided. To ensure that the ethics statement remains accurate, we retained wording consistent with the actual review process and did not add a code that was not issued.

 

7  Comment

Regarding the HUQ, as an ad-hoc outcome instrument, including the items in an appendix (with the number of items per dimension) would elegantly complete the transparency already achieved.

Response

Thank you for this helpful suggestion. To improve the transparency of the HUQ instrument, we added the complete HUQ content to the revised manuscript and included a reference to Appendix A in the main text. Appendix A lists all HUQ items, response options, dimensions, and intended answers, thereby providing a clearer account of the questionnaire’s structure and scoring basis.

Changes made in the manuscript

• Section 4.3.1, Quantitative Data Collection: added a reference to the complete HUQ appendix:

The complete HUQ items, response options, and intended answers are provided in Appendix A.

• End of the manuscript: added Appendix A presenting the complete set of HUQ items:

Appendix A. Narrative Understanding Questionnaire (HUQ)

The appendix includes all HUQ items, the response options for each item, the corresponding dimension, and the intended answers.

 

8  Comment

With n = 16 per group, a quick non-parametric check (Mann–Whitney) as confirmation, and a line on reduced statistical power accompanying non-significant results, would reassure the methodologically attentive reader.

Response

We appreciate the reviewer’s recommendation regarding this additional analysis. Thank you for this helpful suggestion. We added a two-sided Mann–Whitney U analysis as a non-parametric sensitivity check for the same quantitative outcomes. For the SAM measures, the sensitivity analysis follows the primary analysis and uses participant-level post-minus-pre change scores.

The Mann–Whitney analysis is presented as a robustness check rather than as a replacement for the primary independent-samples analysis. This addition allows the revised manuscript to show whether the interpretation of the between-group findings is sensitive to the choice of parametric versus non-parametric testing.

Given that each experimental condition included only 16 participants, we further clarified in the revised manuscript how the small sample size affects the interpretation of non-significant results. Specifically, we added a statement that the limited sample size may have reduced the statistical power to detect small or moderate between-group effects and that results not reaching statistical significance should therefore be interpreted cautiously.

Changes made in the manuscript

• Section 5.1.1: added the following statement after the summary of the non-significant HUQ and NASA-TLX results:

Given the small sample size (n = 16 per condition), the study may have had limited statistical power to detect small or moderate between-group effects; therefore, these non-significant findings should be interpreted cautiously.

• Section 4.4.1, Quantitative Analysis: adds the Mann–Whitney U test as a non-parametric sensitivity analysis and clarifies its supplementary role.
• Section 5.1: adds a dedicated sensitivity-analysis paragraph/subsection reporting the Mann–Whitney check alongside the primary analysis.
• The independent-samples analysis with the already adjusted p-values remains the primary basis for statistical inference.

 

9  Comment

There also remain a few editorial oversights, all immediate: Section 5.2 is titled "Quantitative Findings" but contains the qualitative analysis (to be renamed); Section 4.4.2 "Qualitative Analysis" opens with a quantitative paragraph duplicated from 4.4.1 (which, incidentally, "hides" the information on Levene/Welch there);

Response

Thank you for identifying these issues. We re-examined the titles and organization of the quantitative and qualitative analysis sections and corrected the relevant wording. First, we corrected the title of Section 5.2 so that it accurately reflects the semi-structured interview findings presented in that section. Second, we reorganized Section 4.4.2 by removing the misplaced statistical-analysis content duplicated from the quantitative section and by expanding the description of the qualitative procedure actually used, including thematic analysis, open coding, comparison of coding by two researchers, development of a shared codebook, resolution of coding disagreements, and retention of negative, mixed, and ambiguous feedback. These revisions provide a clearer distinction between the quantitative and qualitative analyses in both the Methods and Results sections.

Changes made in the manuscript

• Section 5.2: corrected the section title to:

5.2 Qualitative Findings

• Section 4.4.2: removed the misplaced quantitative statistical-analysis content and reorganized the section to describe the qualitative analysis method:

4.4.2 Qualitative Analysis

For the qualitative data, thematic analysis was conducted on the semi-structured interview transcripts. Based on the research objectives and the three experimental hypotheses, cognitive understanding and reflection, immersive experience, and emotional engagement were initially established as broad analytical directions. Two researchers independently and repeatedly reviewed the interview transcripts and conducted open coding of participants’ expressions concerning story understanding, task flow, spatial perception, bodily participation, emotional changes, and real-world associations. Rather than assigning all responses directly to the predefined analytical directions, specific meaning units were first identified from the transcripts and subsequently grouped according to similarities and relationships among the emerging codes. The two researchers then compared their coding results, consolidated overlapping codes, and developed a shared codebook containing theme names, definitions, and representative excerpts. Coding disagreements were resolved by returning to the original transcripts and reaching consensus through discussion. Negative, mixed, and ambiguous responses were retained during the analysis to ensure that the resulting themes reflected the diversity of participants’ experiences.

 

10  Comment

A few light editing interventions remain: some terminology inconsistency (for example, the inconsistent capitalization of "Reflection" in hypothesis H1), subdimension labels with adjective and noun inverted in an unidiomatic way in Table 2 and in the figures (Focused Attention, Aesthetic Appeal, Reward, Perceived Usability), and a few specific typos ("Physiacl" in Fig. 3, "HNUQ" in Fig. 11, "9 5% CI" in Section 5.1.2). The non-standard use of the 0–20 NASA-TLX scale to be made explicit; and the completion of references 29 and 40, which lack publication venues.

Response

Thank you for this careful review. In response, we corrected the manuscript item by item for terminology consistency, capitalization, spelling in figures, the NASA-TLX rating range, and reference information. Specifically, we standardized the capitalization of “reflection” in H1; standardized the UES-SF subdimension labels; corrected spelling errors in Fig. 3 and Fig. 11; checked the formatting of “95% CI”; explicitly stated that this study used the 0–20 NASA-TLX rating range rather than the commonly used 0–100 range; and added the publication venues and complete publication details for References 29 and 40.

Changes made in the manuscript

• Section 4, H1: standardized the capitalization of “reflection”:

H1: The integrated tangible–digital interaction system can effectively support users’ cognitive understanding and reflection of serious historical narratives.

• Section 4.3.1, Table 2, Section 5.1.2, and the relevant figures: standardized the UES-SF subdimension labels:

Focused Attention
Aesthetic Appeal
Reward
Perceived Usability

• 3: corrected the spelling in the figure:

“Physiacl” → “Physical”

• 11: corrected the abbreviation in the figure:

“HNUQ” → “HUQ”

• Section 4.3.1: further clarified the NASA-TLX rating range used in this study:

This study used a 0–20 rating scale rather than the more commonly used 0–100 scale, with higher scores indicating greater workload in the corresponding dimension [43].

• References 29 and 40: added and verified the publication venues and related publication details:

Reference 29：

Zhang, X. Oedipus at the Crossroads: Tragic Agency in Contemporary Videogames. Abstract Proceedings of DiGRA 2025: Games at the Crossroads 2025, doi:10.26503/dl.v2025i3.2508.

Reference 40：

Chen, Z. Resonance between Humans and Deities: The Operation of Rain-Praying Rituals in Guizhou during the Ming and Qing Dynasties. CHR 2025, 75, 162–171, doi:10.54254/2753-7064/2025.HT27170.

We sincerely thank the reviewer again for the thoughtful and constructive feedback. We believe that these revisions have substantially improved the clarity, rigor, and overall quality of the manuscript. We hope that our responses and the revised manuscript have adequately addressed all the concerns raised.

Sincerely,
The Authors

 

 

Author Response File: Author Response.pdf

Reviewer 2 Report (Previous Reviewer 1)

Comments and Suggestions for Authors

The article has improved after the revisions. From my point of view it could be published

Author Response

Thank you for your positive evaluation and recommendation for publication. We sincerely appreciate your time and thoughtful review of our manuscript.

Round 2

Reviewer 1 Report (Previous Reviewer 2)

Comments and Suggestions for Authors

1. General Evaluation

The authors have undertaken a thorough and careful revision of the manuscript and have addressed the major comments and concerns raised during the first round of review. The revised version demonstrates substantially improved methodological transparency, conceptual precision, and clarity of presentation.

 

The theoretical framing of Tragic Agency has been significantly refined, particularly through a clearer distinction between tragic agency, false choice, and scripted inevitability, and through its more precise positioning within the literature on interactive narrative. The historical and cultural context has also been clarified, with an appropriate distinction between historical references and artistic fictionalization. Importantly, the authors have moderated several interpretative claims and now align their conclusions more closely with the dimensions that reached statistical significance in this exploratory study.

 

Overall, the manuscript has improved considerably in rigor, clarity, and academic contribution. Only a small number of technical and reporting issues remain and can be addressed through minor revisions.

 

2. Main Strengths and Points Addressed in the Revision

 

Methodological and Statistical Reporting: The authors have clarified that the reported p-values in Table 2 and Section 5.1 are adjusted for multiple comparisons, revised the corresponding statistical notes, and added a non-parametric sensitivity analysis using the Mann–Whitney U test. These additions improve the transparency and robustness of the quantitative analysis.

Effect Sizes and Calculations: The calculations and reporting of effect sizes (η², Cohen’s d, and Hedges’ g) and degrees of freedom have been reviewed and corrected to ensure consistency with the between-subjects design.

Balanced Qualitative Analysis: Sections 4.4.2 and 5.2 now provide a clearer and more transparent description of the thematic analysis, including independent coding, codebook development, and the resolution of coding discrepancies. The inclusion of mixed and critical participant feedback also provides a more balanced account of the qualitative evidence.

Nuanced Interpretation of Constructs: The revised discussion of Tragic Agency, Perceived Choice, and Pleasure/Valence more accurately captures the tension between immediate player action and broader narrative constraint while avoiding unsupported causal interpretations. The comparison between the TUI system and the linear video reference condition is also now appropriately framed as a comparison between integrated experiential conditions rather than isolated components.

Transparency and Reproducibility: The inclusion of the complete Narrative Understanding Questionnaire (HUQ) in Appendix A, together with additional information regarding sensor calibration and the software pipeline, strengthens the reproducibility and methodological transparency of the study.

 

3. Remaining Points for Minor Revision

 

Before publication, I recommend that the authors address the following minor points:

 

Explicit Identification of the Multiple-Comparison Correction Method:

The manuscript now clearly states that the p-values reported in Table 2 were adjusted for multiple comparisons. However, the precise correction procedure should also be explicitly identified in the Table 2 caption or note (e.g., Holm–Bonferroni, Benjamini–Hochberg, or another procedure, as applicable). This would allow readers to understand exactly how the adjusted values were obtained without having to infer the statistical procedure.

Consistency of UES-SF Terminology and Final Proofreading of Figures:

Several typographical issues identified in the previous version have been corrected, including “Physiacl” in Figure 3 and “HNUQ” in Figure 11, and the clarification concerning the NASA-TLX scale. I nevertheless recommend one final systematic check of Figures 3–11, including captions, axis labels, legends, and construct names, to ensure that terminology and capitalization are fully consistent throughout the manuscript. In particular, UES-SF subdimensions such as Focused Attention, Aesthetic Appeal, and Perceived Usability should be reported consistently with the terminology used in Table 2 and in the main text.

Institutional Review Board / Ethics Statement:

The response letter provides a clear explanation of the University of Edinburgh course-based ethics review procedure. The authors should ensure that the final Institutional Review Board Statement accurately reflects this institutional framework and clearly explains the status of the ethical review, including the absence of a dedicated numerical approval code if this is indeed how the relevant course-based ethics procedure operates. The wording should remain fully consistent with the institution’s actual ethics procedure and the journal’s reporting requirements.

 

4. Concluding Recommendation

 

Recommendation: Accept after Minor Revisions

 

The authors have responded carefully and substantively to the concerns raised during the previous round of review. The principal methodological, conceptual, and interpretative issues have been satisfactorily addressed, and the manuscript is now considerably stronger.

 

The remaining requests are limited to reporting precision and final presentation: explicitly identifying the multiple-comparison correction procedure, completing a final consistency check of figure labels and terminology, and ensuring that the Institutional Review Board Statement accurately reflects the applicable institutional ethics framework.

 

Subject to these minor amendments, I consider the manuscript suitable for publication in Multimodal Technologies and Interaction.

Comments on the Quality of English Language

The English is overall clear, correct, and appropriate for a scientific paper, and readability has improved compared to the previous version. A few light editing interventions remain: some terminology inconsistency (for example, the inconsistent capitalization of "Reflection" in hypothesis H1), subdimension labels with adjective and noun inverted in an unidiomatic way in Table 2 and in the figures (Focused Attention, Aesthetic Appeal, Reward, Perceived Usability), and a few specific typos ("Physiacl" in Fig. 3, "HNUQ" in Fig. 11, "9 5% CI" in Section 5.1.2). A final pass of language editing, with attention to terminology consistency and punctuation, will be sufficient to bring the form to the level of the content.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

This manuscript is a resubmission of an earlier submission. The following is a list of the peer review reports and author responses from that submission.

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

It is an interesting aicle, however, the results lacks statistical depth and overstates some conclusions.

 

- Report effect sizes. Currently the paper reports only Mean, SD, t, p.

Without effect sizes it is impossible to judge whether statistically significant findings are actually meaningful. Add Cohen's d, Hedges' g, η²

- Confidence intervals are missing. Instead of only reporting p-values, include Mean difference or 95% confidence interval

- Results are descriptive instead of analytical. Most subsections simply repeat the table.

- Assumptions of t-tests are incomplete. The paper mentions Shapiro-Wilk normality, but not homogeneity of variance or Levene's test

- Sample size discussion. N = 16 per group is relatively small.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

Comments and Suggestions for Authors

The manuscript “Vellum of Lies: Designing and Evaluating a Multimodal Tangible Interaction System for Embodied Serious Historical Narratives” presents an original and promising contribution to the fields of tangible user interfaces, embodied interaction, multimodal storytelling, serious games, and cultural heritage communication. The paper introduces an immersive system that combines physical props, sensors, Kinect-based motion recognition, projection mapping, audio feedback, and final choice-making in order to support users’ understanding, engagement, and emotional response in a serious historical narrative. The research questions are clearly stated and focus on cognitive understanding and reflection, engagement and spatial presence, and emotional response, empathy, and real-world association (lines 97–126).

The main strength of the manuscript lies in the design concept. Vellum of Lies does not treat interaction as a simple interface layer, but attempts to make bodily action, object manipulation, and constrained choice part of the narrative mechanism itself. The system architecture is clearly described through Arduino, Kinect, TouchDesigner, projection, lighting, and audio feedback (lines 273–290), and the visual documentation helps readers understand the installation and the interaction flow. The use of shadow-puppet-inspired visuals and Peking Opera-inspired sound design is also coherent with the cultural atmosphere of the work (lines 370–384). This gives the paper a strong design identity.

However, the theoretical framework requires significant strengthening. The concept of “Tragic Agency” is central to the manuscript, but it is not sufficiently grounded in the literature on interactive narrative, constrained agency, moral choice, serious games, and historical empathy. The authors present Tragic Agency as a mechanism in which users seem to choose but cannot prevent the tragic ending (lines 113–115 and 385–396). This is a powerful idea, but the manuscript should explain more clearly whether it is a new concept, an adaptation of an existing one, or a specific design strategy. At present, the distinction between Tragic Agency, false choice, moral dilemma, limited branching, and scripted inevitability remains underdeveloped.

The historical basis of the narrative also needs clarification. The authors state that the work is inspired by real historical events and addresses feudal oppression, collective violence, and victimization (lines 245–260). Yet the manuscript does not sufficiently explain which historical sources, cultural references, or archival materials informed the story. Since the paper deals with serious historical narrative, it is important to distinguish between historical reconstruction, fictionalization, symbolic representation, and artistic interpretation. This would also strengthen the ethical credibility of the work, especially because the narrative includes execution, guilt, helplessness, and collective violence.

The methodology is suitable for an exploratory prototype study, but not yet strong enough to support the broader claims made in the paper. The study includes 32 participants, divided into 16 participants in the experimental group and 16 in the control group, recruited through personal networks and snowball sampling (lines 401–428). This is a small and convenient sample, composed of students and university staff with frequent exposure to digital media. The authors should therefore avoid generalizing the findings too broadly. They should also clarify whether participants were randomly assigned to conditions.

The control condition also raises concerns. The experimental group experienced the full multimodal installation, while the control group watched a linear video with similar narrative and visual content (lines 441–467). The authors correctly state that the study evaluates the integrated effect of the whole system rather than isolating the contribution of each modality (lines 461–463). However, the discussion sometimes attributes the results specifically to tangible interaction, embodied movement, spatial design, or Tragic Agency. These causal claims should be moderated, because the current design cannot determine which component produced the observed effects. Moreover, the endings were not experienced in the same way by the two groups: experimental participants triggered only one ending and were shown the other afterward, while control participants watched both endings sequentially (lines 468–475). This difference may affect comprehension, emotion, and perceived agency.

The measurement and statistical reporting require major revision. The use of NASA-TLX, UES-SF, IMI, and SAM is appropriate, but the self-developed Narrative Understanding Questionnaire needs much more documentation. The authors state that HUQ was designed around the narrative content and divided into Overall, Detail, and Reflection dimensions (lines 483–492), but the items, validation procedure, reliability, and scoring logic are not provided. The statistical analysis reports Shapiro–Wilk tests and independent-samples t-tests (lines 529–540), but does not include effect sizes, confidence intervals, power considerations, or correction for multiple comparisons. Given the small sample and the number of tested dimensions, these omissions weaken the evidential strength of the results.

The interpretation of the findings should be more cautious. Some results are encouraging: the experimental group shows significant improvements in overall narrative understanding and reflection, but not in detail understanding (lines 565–577). NASA-TLX shows lower overall workload, mental demand, and effort, but not all workload dimensions are significant (lines 580–605). UES-SF shows significant differences in overall engagement and aesthetic appeal, but not in focused attention, reward, or perceived usability (lines 639–650). IMI is significant only for Interest/Enjoyment, while SAM is significant for Pleasure/Valence and Arousal but not Dominance (lines 654–685). Therefore, the abstract and conclusion should avoid suggesting that the system broadly improves all dimensions of understanding, engagement, agency, empathy, and emotional response. The findings support a more limited claim: the prototype shows promising effects on selected dimensions in a small exploratory study.

The qualitative analysis is useful, but it also needs greater methodological transparency. The manuscript states that thematic analysis was conducted through repeated reading, open coding, and grouping of semantic units (lines 541–550). However, it does not clarify how many researchers coded the material, whether coding was independent, whether a codebook was used, how disagreements were resolved, or whether negative and ambiguous cases were considered. The interview quotations are interesting and support the quantitative results, especially regarding embodiment, pressure, empathy, and real-life association (lines 741–814), but the analysis appears somewhat confirmatory. Including more critical or mixed responses would make the qualitative component more convincing.

The Discussion contains valuable reflections, especially when it interprets tangible interaction as cognitive scaffolding and the final button as a reflective device that creates tension between manual choice and tragic inevitability (lines 897–905). Nevertheless, some claims should be softened. The experience lasted approximately eight minutes (lines 449–454), so expressions such as “long-term narrative participation” or broad claims about deep emotional development should be avoided. The discussion of emotional progression and real-world reflection (lines 949–969) should be presented as an exploratory interpretation rather than a definitive conclusion.

The manuscript also needs clearer limitations and stronger ethical reporting. The authors mention future improvements concerning interaction stability, projection clarity, sensor recognition, and field studies (lines 969–992), but they should explicitly discuss the small sample, convenience recruitment, novelty effect, lack of delayed follow-up, cultural specificity, and inability to isolate individual modalities. The ethical section should also be completed in MDPI format. The manuscript states that informed consent was obtained and that the study complied with University of Edinburgh requirements (lines 426–428), but the authors should provide a full ethics statement, clarify consent for the use of participant images, and describe debriefing procedures, given the emotionally sensitive content of the narrative. The final declarations on consent and conflicts of interest should also be checked for completeness (lines 1015–1021).

Overall, this is a promising and original manuscript, with a strong prototype and a valuable research direction. Its main contribution lies in showing how tangible and embodied interaction may become part of serious historical meaning-making, rather than simply serving as an interface technique. However, the article requires major revision before publication. The authors should strengthen the theoretical grounding of Tragic Agency, clarify the historical basis of the narrative, improve methodological transparency, expand the statistical reporting, moderate causal claims, and complete the ethical and editorial sections. I therefore recommend reconsideration after major revision.

Comments on the Quality of English Language

The manuscript is generally well-written, and the narrative flow is clear and understandable. However, it would benefit from a thorough English proofreading and language polishing before final publication. Specifically, I suggest refining the academic phrasing in the theoretical section to ensure that specialized terms, particularly around the core concept of "Tragic Agency", are used with maximum precision. A minor check on sentence structures in the methodology section will also help improve the overall readability and professional tone of the paper.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 3 Report

Comments and Suggestions for Authors

The paper describes an immersive tangible user interface system called Vellum of Lies designed for anti-feudal historical narratives by integrating physical props, sensors, motion capture, and projection mapping.
The authors demonstrate that their embodied interaction approach and Tragic Agency mechanism significantly enhance users' historical understanding, emotional engagement, and reflective thinking compared to passive video watching.  
Though multimodal immersive tangible user interface is an important and popular topic, the current version does not meet the academic publication. Significant revisions regarding experimental depth and comparative validation are necessary.
I recommend considering the following comments to improve the research work.

1.
The core concept of combining tangible user interfaces, projection mapping, and physical props for interactive storytelling is already a heavily researched area in HCI and museum exhibition design.
To prove that this work isn't just a replication of existing museum prototypes, the authors need to expand their literature review and clearly articulate what makes their approach inherently different from prior systems.
Also, the distinct underlying novelty of the Tragic Agency must be explicitly defined.

2.
The manuscript skips most of the granular implementation details necessary for reproducibility. 
The descriptions of sensor thresholds, data calibration protocols, and how the Kinect input translates to real-time visual changes are described in only generic terms.
The authors need to describe the exact details of constraints, signal processing logic, and software pipelines so that other researchers can clearly understand and replicate the technical framework.

3.
Evaluating the system solely against a passive video-watching condition feels like there is no fair objective benchmarks and comparative baselines.
It is completely predictable that an interactive, multimodal system with physical props would outperform a flat video in terms of engagement and emotional response.
To fairly validate the superiority of the proposed method, the study must include a fair, quantitative comparison against an existing interactive baseline such as a standard screen-based digital game or a basic tablet interface using the same narrative.

4.
The user study was conducted in a tightly controlled laboratory environment with a small, homogeneous group of only 32 participants, all of whom were university students or staff.
To prove the robustness of the system, the evaluation needs to be extended to different environmental conditions and broader target audiences, accounting for variables like varying lighting, physical space layouts, and diverse user backgrounds.

Author Response

Please see the attachment.

Author Response File: Author Response.docx