Skip to Content
MDPIMDPI

Advanced

by
  • Massimo Arnone 1,
  • Angelo Leogrande 2,* and
  • Fabio Anobile 2
  • et al.

Reviewer 1: Anonymous Reviewer 2: Anonymous

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This study utilizes regional panel data from the Italian ISTAT-BES database for the period 2004-2022, and comprehensively employs econometric models such as fixed effects, random effects, and dynamic panel models, as well as machine learning methods including regularized linear regression and random forest, along with cluster analysis, to systematically examine the multidimensional determinants of Perceived Risk of Crime (PRC). Overall, the study adopts a comprehensive research perspective, incorporating multiple dimensions such as objective crime exposure, subjective emotions, social trust, institutional trust, subjective well-being, and environmental quality into a unified analytical framework. It demonstrates innovation in the comprehensive application of methodologies and the in-depth exploration of policy implications, and also possesses strong practical significance. Overall, this is a well-executed paper with a topic that holds considerable social value. However, the following issues in the paper deserve the authors' attention and appropriate responses:

1.Some details are not clearly explained. For example, the calculation and extraction process of the dependent and independent variables in this study, and how individual-level variables are unified with regional-level data. Are they all taken as averages? Italy has 20 regions; how many survey samples are there for each region? How is it ensured that after aggregating individual data to the regional level, it can scientifically represent the overall situation of the region, given that there is significant heterogeneity within the regions? Furthermore, regarding the pickpocketing rate, does it come from official statistics or from residents' perceptions? These are fundamentally different and should be clearly stated. These and other details need to be clarified.

2.In the discussion section, the limitations of this study should be more thoroughly discussed. This includes the ecological fallacy problem brought about by aggregating individual data to the regional level. I believe there is no need to avoid this issue; it is necessary to explicitly raise it.

3.The literature review covers a wide range of aspects, but some parts are not highly relevant to this study. Instead, the focus should be on content more directly related to this research, such as the gap between crime exposure and perception, the role of social capital and institutional trust, and the influence of subjective well-being and future expectations.

4. The tables in the article urgently need improvement. The tables are not standardized, and their styles are inconsistent. It would be much better to adopt a three-line table format. Table 4, in particular, contains too much content, and it seems unnecessary to put all this information into a single table.

Author Response

Point to Point Answers to Reviewer 1

Q1. This study utilizes regional panel data from the Italian ISTAT-BES database for the period 2004-2022, and comprehensively employs econometric models such as fixed effects, random effects, and dynamic panel models, as well as machine learning methods including regularized linear regression and random forest, along with cluster analysis, to systematically examine the multidimensional determinants of Perceived Risk of Crime (PRC). Overall, the study adopts a comprehensive research perspective, incorporating multiple dimensions such as objective crime exposure, subjective emotions, social trust, institutional trust, subjective well-being, and environmental quality into a unified analytical framework. It demonstrates innovation in the comprehensive application of methodologies and the in-depth exploration of policy implications, and also possesses strong practical significance. Overall, this is a well-executed paper with a topic that holds considerable social value.

A1. We sincerely thank the reviewer for the positive assessment of our manuscript and for recognizing the relevance of the topic, the multidimensional analytical framework, and the combined use of econometric, machine-learning, and cluster-based methods. We are pleased that the reviewer appreciated the integration of objective crime exposure, subjective perceptions, trust, well-being, and environmental dimensions in the analysis of perceived risk of crime. In the revised manuscript, we have carefully checked the presentation of the methodological framework and policy implications to further improve clarity and coherence.

However, the following issues in the paper deserve the authors' attention and appropriate responses:

Q2. Some details are not clearly explained. For example, the calculation and extraction process of the dependent and independent variables in this study, and how individual-level variables are unified with regional-level data. Are they all taken as averages? Italy has 20 regions; how many survey samples are there for each region? How is it ensured that after aggregating individual data to the regional level, it can scientifically represent the overall situation of the region, given that there is significant heterogeneity within the regions? Furthermore, regarding the pickpocketing rate, does it come from official statistics or from residents' perceptions? These are fundamentally different and should be clearly stated. These and other details need to be clarified.

A2. We thank the reviewer for this valuable comment. We agree that the construction of the variables and the relationship between individual-level information and regional-level indicators required further clarification. In the revised manuscript, we have substantially expanded Section 3 (“Data Sources, Variables, and Descriptive Analysis”). We now explicitly state that the study does not employ individual-level microdata directly. Instead, it relies on official regional-level indicators published within the ISTAT-BES framework. Therefore, the unit of analysis is the region-year, and all variables are regional indicators already processed and released by ISTAT. Depending on the specific indicator, values are expressed as percentages, rates, or average values derived from survey-based and administrative sources. We have also clarified that the aggregation from individual-level information to regional-level indicators is performed by ISTAT according to its official statistical methodologies. Since the analysis relies on official regional indicators rather than raw microdata, the underlying survey sample sizes are determined by ISTAT’s original survey design and weighting procedures and are not recalculated within the present study. To address the reviewer’s concern regarding regional heterogeneity, we now explicitly acknowledge that regional aggregation may conceal within-region heterogeneity. However, the objective of the study is to identify territorial patterns and regional determinants of perceived crime risk rather than individual-level behavioral mechanisms. Accordingly, the results should be interpreted as regional-level associations. We have also clarified the operational definition of the dependent variable. Perceived Risk of Crime (PRC) is defined as the percentage of families reporting that the area in which they live is perceived as being at high or moderate risk of criminality. The indicator is published by ISTAT at the regional level and is derived from the nationally representative “Aspetti della Vita Quotidiana” survey. Finally, we have clarified the nature of the Pickpocketing (PICK) variable. Unlike the perception-based indicators included in the analysis, PICK is treated as an indicator of objective exposure to predatory crime. Its inclusion allows us to distinguish between objective crime exposure and subjective perceptions of insecurity, a distinction that is central to the analytical framework of the study. These clarifications have been incorporated into the revised manuscript in order to improve transparency regarding data construction, aggregation procedures, and variable definitions.

 

Q3. In the discussion section, the limitations of this study should be more thoroughly discussed. This includes the ecological fallacy problem brought about by aggregating individual data to the regional level. I believe there is no need to avoid this issue; it is necessary to explicitly raise it.

A3. We have added a dedicated “Limitations” section in which we explicitly discuss the ecological fallacy issue and clarify that the results should be interpreted as regional-level associations rather than individual-level mechanisms. We have also strengthened the interpretation of the empirical results by discussing the positive coefficient of Life Satisfaction (LISA) and the different signs of Social Participation (SPAR) across model specifications.

Q4. The literature review covers a wide range of aspects, but some parts are not highly relevant to this study. Instead, the focus should be on content more directly related to this research, such as the gap between crime exposure and perception, the role of social capital and institutional trust, and the influence of subjective well-being and future expectations.

A4.  Thank you very much to the reviewer for this constructive comment. Indeed, we do agree that the first version of the literature review touched upon some areas that have little connection to the objectives of this paper, and, thus, the review could have been focused on exploring the literature relevant to the determinants of crime risk perception. In the new version, the literature review is thoroughly reorganized. Specifically, irrelevant discussions were shortened, and the organization of the review corresponds to the main aspects addressed in the empirical part of the paper. The revised literature review focuses on such issues as: (i) objective crime exposure and subjective perceptions of crime risk; (ii) social capital and trust; (iii) environmental factors and perceived safety; (iv) affective aspect and fear of crime; and (v) well-being and future expectations. Moreover, we revised Table 1 and made it more specific in terms of the connections between the major research themes and variables used in the empirical part. Thus, the updated table reflects links between the reviewed research strands and such explanatory variables as objective crime exposure (PICK), fear of crime (FOC), social participation (SPAR), generalized trust (GTR), trust in judicial system (TJS), life satisfaction (LISA), negative future outlook (NFO), and dissatisfaction with local landscape (LDIS).

Q5. The tables in the article urgently need improvement. The tables are not standardized, and their styles are inconsistent. It would be much better to adopt a three-line table format. Table 4, in particular, contains too much content, and it seems unnecessary to put all this information into a single table.

A5. We appreciate the reviewer’s helpful recommendation. In the updated paper, we considerably improved our presentation of the empirical tables, especially Table 4. The original Table 4 combined coefficient estimates with model statistics and test statistics in one table. In order to make the tables clearer, we present the essential coefficient estimates in the body of the text and put the model statistics and specification tests in Appendix A. Such modifications simplify the main empirical table while retaining all information for interested readers. Moreover, we made changes to make the table layout more concise where possible; final journal submission will adhere to the required format.

 

Author Response File: Author Response.docx

Reviewer 2 Report

Comments and Suggestions for Authors

Social Sciences Article Review

“Integrated Evidence on the Determinants of Perceived Crime Risk and Implications for Preventive Community Safety Policies”

The authors of this article examine the objective and subjective predictors of Perceived Risk of Crime (PRC). Using data from the ISTAT-BES system, results from panel data methods (including fixed, random, and dynamic-panel models), weighted least squares regression, and machine learning methods indicated that pickpocketing crimes and fear of crime are significant predictors of a perceived risk of crime. While the overall theme of the study can be important in having practical implications, I have some concerns with the current version of the paper that should be addressed if the paper is to be published.

  1. Unfortunately, the manuscript was difficult to read due to issues with sentence structure, grammar, multiple 1-2 sentence paragraphs (paragraphs should be a minimum of 3 sentences), too many sentences that read very similar to each other in the same paragraph, the overuse of certain terms (“in order to”, “in this investigation”), the use of acronyms without identifying what they mean, etc. The manuscript would greatly benefit from copyediting to make the text easier to read/follow.
  2. Relatedly, and unfortunately, the manuscript reads as if it is trying to do “too much” within one paper, which makes the paper difficult to follow. For example, there are currently 14 tables and 4 figures for the manuscript, along with numerous types of analytic techniques that result in the manuscript being more confusing to read rather than adding clarity. The manuscript would benefit from being streamlined with a few number of analyses, tables, figures, etc. It is possible that this manuscript could be divided into multiple papers.

  3. Relatedly (again), due to the above two concerns, the manuscript was difficult to follow. Even after reading multiple times, there is a disjunction between what the manuscript is trying to do, and an inability to present the findings in a clear manner, and providing suggestions for policy/practice that are not broad and sweeping statements with specific suggestions for implementation. Unfortunately, I feel that readers will be unable to follow this manuscript and be as “impactful” as the authors are arguing the manuscript will be.

  4. The authors utilize the ISTAT-BES system to examine their research questions. However, this “system” is not identified or described in the Abstract.

  5. In the Abstract, the last sentence is 8 lines long. This is one example of the manuscript being confusing and difficult to follow.

  6. Throughout the manuscript, “filler words” are constantly used that make the manuscript “wordy” and difficult to follow. Examples include “In this sense” (to begin the entire manuscript), the use of “This” to begin a sentence (what, or who is “this”?)

  7. Citations are needed after the sentence “A significant body of literature has demonstrated the gap between objective…”

  8. The first three pages of the manuscript is one paragraph. There are no transitions from the different topics of thought, and this section reads as if “everything but the kitchen sink” has been put into this section.

  9. In the Introduction section, please number the research questions to make them stand out from the rest of the section.

  10. In the Introduction section, please move the text related to the “three main contributions” of the study into a new “Current Study” section before the Methodology. This text can be revised into three sentences for the Introduction section, but the bulk of the text about contributions is more appropriate in a separate “Current Study” section.

  11. In the Introduction section, please remove the paragraph about how “the article continues as follows”. Readers know the layout of academic articles, they do not need to be told the ordering of traditional manuscript sections.

  12. Unfortunately, the Literature Review reads as “grocery listing” the findings after a "theme" is presented. For example, a paragraph begins with a theme, and then various studies are described in the same fashion, versus synthesizing the literature.

  13. The formatting of the tables are not clear nor concise. There is text such as “See Table 1” without any description of what the tables are, and the text does not refer to the statistics in the tables (and why so many statistics are included) in each of the 14 tables.

  14. In Section 3 (Data Sources, Variables, and Descriptive Analyses), please describe how each of the variables are measured. For example, how is the dependent variable (Perceived Risk of Crime) exactly measured? Through a survey item? For example, Table 2 says that the “fear of crime” variable “captures the emotional response to crime, measuring the extent to which individuals feel afraid of becoming victims”, but it is not clear the exact survey item, or how this item is worded, to help measure this variable. Therefore, due to this lack of clarity in the variables, there is lack of clarity in the analyses, and then a lack of clarity in the results and conclusions presented.

  15. Relatedly, how are the data collected? What does “380 regional observations” mean? Does this mean that 380 individuals were surveyed? How was the survey conducted? How were these participants chosen? What are their characteristics? There is no information about the characteristics of the “regional observations”.

  16. There is a great amount of detail concerning the mode and median of the variables, when the mean, range, and standard deviations would suffice.

  17. There are multiple one-sentence paragraphs throughout the Results section.

  18. Unfortunately, there are so many different types of analyses and presentation of the Results that the reader gets lost at the overall purpose of the research questions and how the analyses specifically relate to each question. Please consider streamlining the analyses and directly connect each analysis to a research question to make it easier for the reader to follow.

  19. Throughout the Results section, the authors discuss how the findings support or do not support prior research (with accompanying citations). This type of information would be more suitable in the Discussion section (not intertwined within the results). For example, the second to last paragraph on page 12 would be more appropriate in the Discussion section (along with the prior paragraphs before it with reference to past research), and the last paragraph on page 12 would be an appropriate conclusion paragraph to the entire manuscript. The same situation occurs on page 18 for the paragraph that begins with “Overall, the findings of the cluster analysis…”

  20. On page 20, the text says “Table X” where it looks like the authors forgot to insert what table number they are referencing.

  21. It may be helpful for the figures to focus on patterns versus color, as not all articles may be able to print/be shown online in color.

  22. Throughout the Results section, there is a great deal of time spent on focusing on defending the correct type or best fitting model versus focusing on the specific findings of the analyses. This information detracts from the Results section and makes this section difficult to follow.

  23. In the Discussion section, the authors make broad and sweeping statements about policy and program implications that are already occurring in practice. Examples such as “primary, secondary, and prevention programs” are already occurring in practice. Greater specifics are needed that can actually be implemented. For example, how can crime prevent strategies focus on social conditions that shape citizens’ perceptions of insecurity? What reform efforts can improve the efficacy of the judicial system, reduce the delay of legal proceedings, increase the visibility of fair and effective law enforcement, etc.? What community programs can be implemented to promote collective responsibility for safety? Currently, the Discussion section uses general terminology (and programs/policies that are already occurring, such as community policing) as suggestions for implications based on the study’s findings. Unfortunately, this section reads rather “fluffy” versus engaging in critical thinking.

  24. Currently, only two study limitations are briefly described in the manuscript. In the Discussion section, please identify additional study limitations and elaborate on how these limitations can impact the results.

  25. In the Discussion section, please describe suggestions for future research based on the current study’s findings. Currently, no suggestions are provided.

  26. Please provide a concluding paragraph to the entire manuscript, versus ending the manuscript with a basic sentence that has been repeated over and over again throughout the manuscript. Currently, the manuscript abruptly ends.

  27. In addition, please review the entire manuscript to ensure that statements are not overgeneralized or overstating conclusions that cannot be made based on the study’s findings.

  28. In the Discussion section, for this study to make a clear contribution that is translatable for both practitioners and policymakers, it would be helpful if the authors provided some of the main “takeaway” points of the entire study in the Discussion section. Overall, this section is an opportunity for the authors to think critically about the findings and provide a detailed discussion, and more depth is needed across this section.

Comments on the Quality of English Language

Unfortunately, the manuscript was difficult to read due to issues with sentence structure, grammar, multiple 1-2 sentence paragraphs (paragraphs should be a minimum of 3 sentences), too many sentences that read very similar to each other in the same paragraph, the overuse of certain terms (“in order to”, “in this investigation”), the use of acronyms without identifying what they mean, etc. The manuscript would greatly benefit from copyediting to make the text easier to read/follow.

Author Response

Point to Point Answers to Reviewer 2

 

Q0.  “Integrated Evidence on the Determinants of Perceived Crime Risk and Implications for Preventive Community Safety Policies” The authors of this article examine the objective and subjective predictors of Perceived Risk of Crime (PRC). Using data from the ISTAT-BES system, results from panel data methods (including fixed, random, and dynamic-panel models), weighted least squares regression, and machine learning methods indicated that pickpocketing crimes and fear of crime are significant predictors of a perceived risk of crime. While the overall theme of the study can be important in having practical implications, I have some concerns with the current version of the paper that should be addressed if the paper is to be published.

A0. We thank the Reviewer for the careful and thorough evaluation of our manuscript and for the constructive and detailed feedback provided. We are grateful that the Reviewer recognizes the importance of the study's theme and its potential practical implications. We have addressed all of the concerns raised in the review in full. The manuscript has been substantially revised in response to each specific comment (Q15 through Q28), with changes affecting the data description, the structure and presentation of the results, the Discussion, the Policy Implications, the Limitations, and the Conclusions sections. A point-by-point response to each comment is provided below. We believe that the revised manuscript is considerably stronger than the original version and that it now addresses all of the methodological, structural, and substantive concerns raised by the Reviewer. We hope that the revised manuscript will be considered suitable for publication in Social Sciences.

 

Q1. Unfortunately, the manuscript was difficult to read due to issues with sentence structure, grammar, multiple 1-2 sentence paragraphs (paragraphs should be a minimum of 3 sentences), too many sentences that read very similar to each other in the same paragraph, the overuse of certain terms (“in order to”, “in this investigation”), the use of acronyms without identifying what they mean, etc. The manuscript would greatly benefit from copyediting to make the text easier to read/follow.

 

A1. Thank you for this valuable comment. We carefully revised the manuscript to improve its overall readability and clarity. In particular, we performed a thorough language and style revision throughout the paper, focusing on sentence structure, grammar, paragraph organization, and redundancy. Several sections, including the Abstract, Introduction, Data and Methods, and Empirical Results, were substantially rewritten to improve readability and logical flow. Long and repetitive sentences were simplified, and paragraphs were reorganized to ensure greater coherence and continuity. We also reduced the repeated use of expressions such as “in order to,” “in this investigation,” and similar formulations. In addition, all acronyms are now defined at their first occurrence (e.g., ISTAT, BES, PRC, SPAR, GTR, TJS, PICK, FOC, LISA, NFO, and LDIS), and terminology has been standardized throughout the manuscript. We also revised several sections to eliminate repetitive statements and improve the clarity of the discussion of results. Overall, these revisions were undertaken to make the manuscript easier to read and follow while preserving the original scientific content and interpretation of the findings.

 

 

 

Q2. Relatedly, and unfortunately, the manuscript reads as if it is trying to do “too much” within one paper, which makes the paper difficult to follow. For example, there are currently 14 tables and 4 figures for the manuscript, along with numerous types of analytic techniques that result in the manuscript being more confusing to read rather than adding clarity. The manuscript would benefit from being streamlined with a few number of analyses, tables, figures, etc. It is possible that this manuscript could be divided into multiple papers.

 

A2. Thank you for this important comment. We revised the manuscript to make the empirical strategy more focused and easier to follow. The revised paper now clarifies that panel econometric models constitute the main empirical analysis, while machine learning is used only as a robustness check and clustering is used as an exploratory tool for identifying regional profiles. We also streamlined the results sections by reducing repetitive explanations, shortening the discussion of secondary analyses, and clarifying the role of each method within a single sequential research design. Several supplementary results and diagnostic materials were moved to the Appendix, while the main text now focuses on the core econometric findings, the main clustering profiles, and the robustness evidence from machine learning. These revisions were made to avoid the impression that the manuscript contains several independent studies. The revised version presents the analyses as complementary components of one empirical framework addressing the determinants of perceived crime risk across Italian regions.

 

 

 

Q3. Relatedly (again), due to the above two concerns, the manuscript was difficult to follow. Even after reading multiple times, there is a disjunction between what the manuscript is trying to do, and an inability to present the findings in a clear manner, and providing suggestions for policy/practice that are not broad and sweeping statements with specific suggestions for implementation. Unfortunately, I feel that readers will be unable to follow this manuscript and be as “impactful” as the authors are arguing the manuscript will be.

 

A3. Thank you for this important comment. We revised the manuscript to improve the connection between the research objectives, the empirical findings, and the policy implications. First, we clarified the structure of the paper and emphasized that the econometric, machine learning, and clustering analyses are complementary components of a single empirical framework. Second, we added an integrated summary of the main findings, showing that visible crime exposure, fear of crime, and institutional/social trust are the most consistent determinants of perceived crime risk. Third, we substantially revised the policy section to make the implications more concrete and implementable. The revised manuscript now links each empirical result to specific policy actions, including targeted prevention in transport hubs and tourist areas, reassurance campaigns, stronger citizen–institution interaction, urban regeneration, and social inclusion policies. These revisions were intended to make the manuscript easier to follow, strengthen the connection between evidence and practice, and avoid broad or sweeping policy recommendations. In addition, we have expanded the interpretation of potentially counterintuitive empirical findings, particularly the positive coefficient of Life Satisfaction (LISA) and the different signs of Social Participation (SPAR) across static and dynamic panel specifications.

 

Q4. The authors utilize the ISTAT-BES system to examine their research questions. However, this “system” is not identified or described in the Abstract.


A4. Thank you for this comment. We revised the Abstract to identify and briefly describe the ISTAT-BES system at its first occurrence. The Abstract now specifies that the data come from the Benessere Equo e Sostenibile (BES, Equitable and Sustainable Well-being) framework developed by the Italian National Institute of Statistics (ISTAT). This clarification was added to make the data source immediately understandable to readers.

 

Q5. In the Abstract, the last sentence is 8 lines long. This is one example of the manuscript being confusing and difficult to follow.

 

A5. Thank you for this observation. We revised the Abstract to improve readability and reduce sentence length. In particular, the original concluding sentence was divided into several shorter sentences that separately present the main findings, the interpretation of the results, and the policy implications. These changes were made to improve clarity and make the Abstract easier to follow for readers.



Q6. Throughout the manuscript, “filler words” are constantly used that make the manuscript “wordy” and difficult to follow. Examples include “In this sense” (to begin the entire manuscript), the use of “This” to begin a sentence (what, or who is “this”?)

 

A6. Thank you for this comment. We carefully revised the manuscript to reduce unnecessary filler expressions and improve readability throughout the text. In particular, we removed or substantially reduced repetitive formulations such as “in this investigation”, “in this sense”, and similar transitional expressions that did not add substantive meaning. We also revised numerous sentences beginning with generic references such as “This” or “These” by replacing them, where appropriate, with more explicit subjects (e.g., “the evidence”, “the results”, “the empirical findings”, or “the analysis”). In addition, we simplified sentence structure across the manuscript, shortened several overly long passages, and reduced redundancy in the presentation of results and interpretations. These revisions were intended to improve clarity, readability, and the overall flow of the manuscript while preserving the original meaning of the analysis. We believe that the revised version is substantially easier to follow and more concise than the original submission.



Q7. Citations are needed after the sentence “A significant body of literature has demonstrated the gap between objective…”

 

A7. Thank you for this observation. During the revision, the sentence mentioned by the reviewer was removed as part of the broader restructuring and copyediting of the Introduction. The same idea is now expressed in revised form and supported by appropriate citations in the surrounding discussion on the gap between objective crime rates and subjective perceptions of insecurity.



Q8. The first three pages of the manuscript is one paragraph. There are no transitions from the different topics of thought, and this section reads as if “everything but the kitchen sink” has been put into this section.

A8. Thank you for this comment. We substantially revised the Introduction to improve its organization and readability. The original long section was restructured into multiple thematic paragraphs, each addressing a specific aspect of the research problem, including the relevance of perceived insecurity, the gap between objective crime and subjective perceptions, the Italian context, the ISTAT-BES framework, the main strands of the literature, the existing research gaps, the research questions, and the study’s contributions. We also added clearer transitions between topics and reduced the amount of material presented within individual paragraphs. These revisions were intended to improve the logical flow of the Introduction and make the manuscript easier to follow.



Q9. In the Introduction section, please number the research questions to make them stand out from the rest of the section.

A9. Thank you for this suggestion. We revised the Introduction to present the research questions as explicitly numbered research questions (RQ1 and RQ2). This change was made to improve readability and to make the study objectives more visible and easier to identify within the Introduction.



Q10. In the Introduction section, please move the text related to the “three main contributions” of the study into a new “Current Study” section before the Methodology. This text can be revised into three sentences for the Introduction section, but the bulk of the text about contributions is more appropriate in a separate “Current Study” section.

 

A10. Thank you for this suggestion. We revised the manuscript by shortening the discussion of the study’s contributions in the Introduction to three concise sentences. We then moved the fuller discussion of the study’s contributions into a new “Current Study” section placed before the data and methodology section. This new section clarifies the study’s multidimensional framework, longitudinal regional perspective, and complementary methodological strategy.

Q11. In the Introduction section, please remove the paragraph about how “the article continues as follows”. Readers know the layout of academic articles, they do not need to be told the ordering of traditional manuscript sections.

 

A11. Thank you for this suggestion. We removed the paragraph describing the organization of the manuscript from the Introduction. The revised version now moves directly from the study objectives and contributions to the substantive sections of the paper, improving conciseness and readability.



Q12. Unfortunately, the Literature Review reads as “grocery listing” the findings after a "theme" is presented. For example, a paragraph begins with a theme, and then various studies are described in the same fashion, versus synthesizing the literature.

A12. Thank you for this comment. We substantially revised the Literature Review to avoid presenting previous studies as a list of individual findings. The revised section now synthesizes the literature around broader conceptual dimensions, including the gap between objective crime exposure and perceived insecurity, environmental conditions, social capital and institutional trust, emotional responses, and subjective well-being. We also reduced repetitive study-by-study descriptions and clarified how each strand of literature informs the empirical framework of the present study.



Q13. The formatting of the tables are not clear nor concise. There is text such as “See Table 1” without any description of what the tables are, and the text does not refer to the statistics in the tables (and why so many statistics are included) in each of the 14 tables.

A13. Thank you for this comment. We substantially revised the presentation of the tables throughout the manuscript to improve clarity and readability. First, we removed generic references such as “See Table X” and replaced them with concise explanatory text introducing each table and clarifying its purpose within the analysis. The revised manuscript now explains what each table contains, why it is included, and how it contributes to the interpretation of the results. Second, we expanded the discussion of the tabulated results by explicitly referring to the most relevant statistics and findings reported in each table. Rather than simply presenting tables, the revised text now highlights the key patterns, relationships, and comparisons that are important for understanding the determinants of perceived crime risk. Third, we revised table notes and captions to make them more informative and self-contained. The revised notes explain the meaning of the reported indicators, the rationale for including specific statistics, and the analytical contribution of each table to the overall argument of the paper. Overall, these revisions were intended to improve the integration between the narrative discussion and the tabulated evidence, making the empirical results easier to follow and interpret.

 

 

 

Q14. In Section 3 (Data Sources, Variables, and Descriptive Analyses), please describe how each of the variables are measured. For example, how is the dependent variable (Perceived Risk of Crime) exactly measured? Through a survey item? For example, Table 2 says that the “fear of crime” variable “captures the emotional response to crime, measuring the extent to which individuals feel afraid of becoming victims”, but it is not clear the exact survey item, or how this item is worded, to help measure this variable. Therefore, due to this lack of clarity in the variables, there is lack of clarity in the analyses, and then a lack of clarity in the results and conclusions presented.

A14. Thank you for this important observation. We substantially revised Section 4 to improve the transparency and operational clarity of all variables used in the analysis. First, we expanded the description of the dependent variable and the explanatory variables in the Data Sources and Variables section. The revised text now clarifies not only the conceptual meaning of each indicator but also its operational definition within the ISTAT-BES framework. Second, we redesigned Table 2 to provide a more detailed description of the variables. In addition to the variable name and acronym, the revised table now reports the conceptual meaning, the exact operational measurement, and the original data source for each indicator. For example, the dependent variable (Perceived Risk of Crime) is now explicitly defined as the percentage of households reporting a high or moderate risk of criminality in the area where they live. Similarly, the operational definitions of social participation, generalized trust, trust in the judicial system, pickpocketing, fear of crime, life satisfaction, negative future outlook, and dissatisfaction with the local landscape are now reported in detail. Third, the revised table specifies the measurement scale used for each variable, including percentages, rates per 1,000 inhabitants, and mean scores on a 0–10 scale where applicable. The corresponding ISTAT survey or administrative source is also reported for each indicator. These revisions were intended to improve methodological transparency, facilitate the interpretation of the empirical results, and provide readers with a clearer understanding of how the variables were constructed and measured within the ISTAT-BES framework.

 

 

 

 

Q15. Relatedly, how are the data collected? What does “380 regional observations” mean? Does this mean that 380 individuals were surveyed? How was the survey conducted? How were these participants chosen? What are their characteristics? There is no information about the characteristics of the “regional observations”.

 

A15. We thank the Reviewer for this important observation. We have substantially revised and expanded Section 4 to address all the points raised. First, we clarify that "380 regional observations" does not refer to 380 individual respondents. The dataset consists of repeated regional observations for the twenty Italian regions over the period 2004–2022. In a fully balanced panel, twenty regions observed over nineteen years generate 380 region-year observations. The unit of analysis is therefore the region-year, not the individual. Second, we have added a dedicated paragraph describing the survey instruments through which ISTAT collects the underlying individual-level data. The Multipurpose Survey on Aspects of Daily Life (AVQ) is a large annual sample survey interviewing approximately 20,000 households and 50,000 individuals each year. Since 2018, it has been conducted using a sequential CAWI/PAPI mixed-mode technique. The survey employs a two-stage stratified sampling design: in self-representing municipalities, households are selected directly without replacement and with equal probabilities; in non-self-representing municipalities, municipalities are first selected as primary units without replacement and with variable probabilities, followed by household selection. The Survey on Citizens' Safety (Indagine sulla Sicurezza dei Cittadini) — used for crime-related indicators such as pickpocketing and fear of crime — was conducted on a sample of 29,317 individuals aged 14 and over, interviewed both by telephone and in person. Third, regarding the characteristics of the observations, we explain that the present study does not use individual-level microdata. Individual responses collected by ISTAT are aggregated into regional-level indicators according to ISTAT's official statistical methodologies. Consequently, individual respondent characteristics such as age, gender, education, or income distributions cannot be reported at the study level, as the observational unit is the Italian region over time, not the individual survey respondent. This approach is standard practice in studies relying on the ISTAT-BES framework.



Q16. There is a great amount of detail concerning the mode and median of the variables, when the mean, range, and standard deviations would suffice.

 

A16. We thank the Reviewer for this observation. We agree that the original descriptive statistics section contained excessive detail that went beyond what is necessary for the reader to understand the distributional properties of the variables. Accordingly, we have substantially revised and streamlined the descriptive statistics section in Section 4. Specifically, we have removed all references to modes, medians, quartiles, and kurtosis values from the narrative text. The revised text now reports only the mean, standard deviation, and range for each variable, as suggested by the Reviewer, along with the Shapiro-Wilk test results to justify the choice of robust estimation strategies. A brief note has also been added to direct the reader to Table A3 in Appendix A, where the full distributional statistics — including medians, quartiles, kurtosis, and skewness — remain available for consultation.The revised descriptive statistics section can be found in Section 4 of the manuscript.



Q17.There are multiple one-sentence paragraphs throughout the Results section.

 

A17. We thank the Reviewer for this observation. We agree that the presence of isolated single-sentence paragraphs disrupts the flow of the Results section and reduces its readability. We have therefore carefully revised the manuscript to address this issue. Specifically, all single-sentence paragraphs identified throughout the Results section have been eliminated by merging them with adjacent paragraphs or integrating them into the surrounding narrative. In Section 6, the fragmented description of Figure 2 has been rewritten as a single, coherent paragraph. In Section 8, the isolated sentence on the persistence of perceived insecurity over time has been incorporated into the econometric results paragraph, and the isolated sentence on variable importance and dropout loss has been merged into the machine learning paragraph. As a result, each paragraph in the Results section now develops a complete and self-contained argument, improving the overall coherence and readability of the manuscript. The revised text can be found in Sections 5, 6, 7, and 8 of the manuscript.



Q18. Unfortunately, there are so many different types of analyses and presentation of the Results that the reader gets lost at the overall purpose of the research questions and how the analyses specifically relate to each question. Please consider streamlining the analyses and directly connect each analysis to a research question to make it easier for the reader to follow.

 

A18. We thank the Reviewer for this important observation. We agree that the original manuscript did not make sufficiently explicit the connection between each analytical component and the two research questions guiding the study. We have revised the manuscript accordingly. Specifically, each results section now opens with an explicit statement linking the analysis to the relevant research question. Section 5 opens by stating that the panel econometric results directly address RQ1 — what are the main determinants of perceived crime risk in Italian regions? Section 6 opens by stating that the clustering analysis contributes to addressing RQ2 — how do crime exposure, emotional reactions, institutional trust, social participation, and well-being conditions interact in shaping perceived insecurity? — and this connection is reinforced at two further points within the section where the cluster profiles are interpreted in light of RQ2. Section 7 opens by clarifying that the machine learning analysis serves as a robustness check on the findings related to both RQ1 and RQ2, rather than as an independent analytical objective. Section 8 synthesizes the evidence from all three approaches and provides an integrated answer to both research questions. These revisions make the analytical structure of the paper more transparent and allow the reader to follow the logical progression from research questions to empirical findings across all sections of the manuscript. The revised text can be found in Sections 5, 6, 7, and 8 of the manuscript.



Q19. Throughout the Results section, the authors discuss how the findings support or do not support prior research (with accompanying citations). This type of information would be more suitable in the Discussion section (not intertwined within the results). For example, the second to last paragraph on page 12 would be more appropriate in the Discussion section (along with the prior paragraphs before it with reference to past research), and the last paragraph on page 12 would be an appropriate conclusion paragraph to the entire manuscript. The same situation occurs on page 18 for the paragraph that begins with “Overall, the findings of the cluster analysis…”

A19. We thank the Reviewer for this important structural observation. We agree that the original manuscript did not maintain a clear separation between the presentation of results and their interpretation in light of prior research. We have revised the manuscript accordingly. Specifically, we have introduced a new dedicated Discussion section (Section 10), placed between the Integrated Evidence section and the Limitations section. The following material has been removed from the Results sections and reorganized into the Discussion: the paragraphs comparing the trust and social participation findings with prior research (originally in Section 5), the paragraph on the remaining well-being and environmental variables and their consistency with the literature (originally in Section 5), and the paragraph on the component profiles and their interpretive implications in light of prior studies on collective efficacy, social capital, and fear of crime (originally in Section 6). The Discussion section now brings together the interpretation of all empirical findings in relation to the existing literature, covering the econometric results, the clustering profiles, and the machine learning robustness checks within a unified interpretive narrative. As a result, the Results sections now report empirical findings only — coefficients, directions, significance levels, and cluster profiles — while the Discussion section provides the theoretical interpretation and connection to prior research. The revised structure can be found in Sections 5 through 10 of the manuscript.



Q20. On page 20, the text says “Table X” where it looks like the authors forgot to insert what table number they are referencing.

 

A20. We thank the Reviewer for pointing out this oversight. We have carefully reviewed the entire manuscript and corrected the incomplete table reference. The placeholder "Table X" has been replaced with the correct table number. We have also conducted a full check of all table and figure references throughout the manuscript to ensure that no other incomplete or inconsistent cross-references remain. All references to tables and figures are now complete and correctly numbered throughout the text.



Q21. It may be helpful for the figures to focus on patterns versus color, as not all articles may be able to print/be shown online in color.

 

A21. We thank the Reviewer for this suggestion. The figures in this manuscript were produced using JASP, which does not currently support the addition of hatching patterns or greyscale-compatible fills as an export option. We note that the journal Social Sciences publishes articles in open-access format with full color reproduction available online at no additional cost to authors or readers, which means that color differentiation in the figures remains fully accessible to all readers in the published version. We have nonetheless ensured that, where possible, cluster assignments and group distinctions are also conveyed through differences in shape and symbol type, in addition to color, to improve readability across different display contexts.

 

Q22. Throughout the Results section, there is a great deal of time spent on focusing on defending the correct type or best fitting model versus focusing on the specific findings of the analyses. This information detracts from the Results section and makes this section difficult to follow.

 

A22. We thank the Reviewer for this observation. We agree that the original Results section contained an excessive amount of text dedicated to justifying and comparing model specifications, which diverted attention from the substantive findings. We have revised the manuscript accordingly by removing from the Results section all paragraphs focused on model selection and algorithm comparison. Specifically, the paragraph listing and comparing the four panel data estimators has been removed from Section 5, as this information is already covered in the methodological description in Section 3. The paragraph comparing the performance of clustering algorithms — including Fuzzy C-Means, K-Means, Random Forest, and hierarchical clustering — has been removed from Section 6, as the relevant validation statistics are already reported in Table A4 in Appendix A. In Section 7, the paragraphs describing the evaluation metrics and comparing the performance of the individual machine learning algorithms have been removed, as this information is fully available in Table 5. Each results section now opens directly with the substantive findings, keeping the focus on what the analyses reveal about the determinants of perceived crime risk rather than on technical model selection procedures.

 

Q23. In the Discussion section, the authors make broad and sweeping statements about policy and program implications that are already occurring in practice. Examples such as “primary, secondary, and prevention programs” are already occurring in practice. Greater specifics are needed that can actually be implemented. For example, how can crime prevent strategies focus on social conditions that shape citizens’ perceptions of insecurity? What reform efforts can improve the efficacy of the judicial system, reduce the delay of legal proceedings, increase the visibility of fair and effective law enforcement, etc.? What community programs can be implemented to promote collective responsibility for safety? Currently, the Discussion section uses general terminology (and programs/policies that are already occurring, such as community policing) as suggestions for implications based on the study’s findings. Unfortunately, this section reads rather “fluffy” versus engaging in critical thinking.

 

A23. We thank the Reviewer for this substantive critique. We fully agree that the original policy section relied on generic frameworks and broad terminology that lacked the specificity and critical thinking that empirical findings of this kind should generate. We have substantially revised both the Policy Implications section and the Conclusions in response. The Policy Implications section has been completely rewritten. Each implication is now directly anchored to a specific empirical finding from the econometric or cluster analysis, accompanied by a concrete implementation mechanism and a measurable expected outcome. Generic references to community policing, primary/secondary/tertiary prevention frameworks, and broad civic participation initiatives have been removed entirely. Specifically, the revised section addresses five targeted areas. First, on visible predatory crime, we propose evidence-based hotspot policing strategies concentrated in metro stations, tourist districts, and open-air markets — the specific locations where pickpocketing rates are systematically higher in Italy — integrated with real-time citizen reporting platforms. Second, on institutional trust, we propose concrete judicial efficiency reforms tied to Italy's PNRR agenda, including case management reforms to reduce trial duration, expansion of alternative dispute resolution for minor offenses, and regional transparency dashboards for court performance. Third, on collective efficacy, we propose neighborhood safety councils with formal decision-making power over local security budgets and co-production agreements between municipal police and resident associations in high-PRC regions — explicitly distinguishing these from generic civic participation programs. Fourth, on environmental dissatisfaction, we propose the structured application of CPTED principles tailored to the regional cluster profiles identified in the study, with rapid-response maintenance programs in high-LDIS regions. Fifth, on socio-economic pessimism, we propose place-based economic activation programs specifically designed for Southern Italian regions where pessimism, low life satisfaction, and high PRC co-occur. Table 7 has been revised to reflect these specific interventions, replacing the previous generic policy matrix with a six-row table that links each finding to a targeted intervention, a concrete implementation mechanism, and an expected outcome. The reference to primary/secondary/tertiary prevention has also been removed from the Conclusions, which have been rewritten to present three specific policy priorities grounded in the study's findings.



Q24. Currently, only two study limitations are briefly described in the manuscript. In the Discussion section, please identify additional study limitations and elaborate on how these limitations can impact the results.

 

A24. We thank the Reviewer for this important observation. We agree that the original Limitations section was insufficiently developed, covering only three closely related limitations without adequately elaborating on their implications for the results. We have substantially expanded the section, which now identifies seven distinct limitations and discusses how each one may affect the reliability, generalizability, and causal interpretation of the findings. The revised section addresses the following limitations. First, the ecological fallacy risk inherent in regional-level analysis is discussed with a concrete example showing how within-region heterogeneity in trust can mask subgroup-level dynamics. Second, within-region heterogeneity is elaborated with reference to the internal diversity of Italian regions such as Lombardy. Third, the absence of individual-level microdata is discussed in terms of its implications for omitted variable bias, noting that the inability to control for age, gender, education, income, and victimization history may confound the regional-level associations. Fourth, the absence of causal identification is addressed explicitly, noting that the panel models identify associations rather than causal effects and that the relationship between fear of crime and perceived risk may be subject to simultaneity bias, potentially leading to overestimation of the fear of crime coefficient. Fifth, the mildly unbalanced panel structure is discussed in terms of potential selection bias if data availability is correlated with perceived insecurity levels. Sixth, the country-specific focus on Italy is identified as a limitation for external validity, noting that findings related to judicial trust and pickpocketing may not generalize to other institutional contexts. Seventh, the omission of media exposure, political discourse, and online information environments is discussed as a potential source of omitted variable bias in the fear of crime and negative future outlook coefficients. The revised Limitations section can be found in Section 11 of the manuscript.



Q25. In the Discussion section, please describe suggestions for future research based on the current study’s findings. Currently, no suggestions are provided.

 

A25. We thank the Reviewer for this observation. We agree that the original Discussion section did not include directions for future research. We have added a dedicated paragraph at the end of Section 10 that identifies six specific directions for future research, each directly grounded in the empirical findings or methodological limitations of the current study. The first direction proposes replicating the analytical framework using individual-level microdata — specifically the European Social Survey or ISTAT multipurpose microdata files — combined with multilevel modeling, to disentangle contextual from individual-level mechanisms and identify the demographic subgroups most affected by each determinant of PRC. The second direction calls for research examining the specific types of social participation that most effectively build collective efficacy over time, building on the dual role of social participation identified in this study. The third direction proposes investigating the mechanisms through which visible predatory crime shapes perceptions beyond direct victimization, including indirect victimization and social information transmission, using sub-regional spatial crime data. The fourth direction suggests longitudinal cluster analysis to examine the temporal stability of the eight regional profiles identified here and their responses to specific shocks such as the COVID-19 pandemic or judicial reforms. The fifth direction recommends cross-national replication using comparable regional datasets from other European countries, to assess whether the relationship between institutional trust and perceived insecurity generalizes beyond the Italian context. The sixth direction proposes incorporating media exposure, political discourse, and social media sentiment indicators as additional explanatory variables, to disentangle media-amplified narratives from direct community experiences in shaping fear of crime and negative future outlook. The revised text can be found at the end of Section 10 of the manuscript.



Q26. Please provide a concluding paragraph to the entire manuscript, versus ending the manuscript with a basic sentence that has been repeated over and over again throughout the manuscript. Currently, the manuscript abruptly ends.

 

A26. We thank the Reviewer for this observation. We agree that the original manuscript ended abruptly with a sentence that summarized findings already stated multiple times throughout the paper. We have added a substantive concluding paragraph at the end of Section 12 that provides a genuine closing to the manuscript. The new paragraph steps back from the specific empirical findings to reflect on the broader significance of the study. It argues that perceived insecurity is one of the most consequential yet misunderstood dimensions of social well-being, and that the persistent focus of policy on crime rates and enforcement represents a fundamental mismatch with what the evidence shows. Drawing on the study's findings, it reframes the gap between objective crime and subjective insecurity not as a cognitive distortion but as a social phenomenon shaped by institutional quality, community bonds, environmental conditions, and collective expectations about the future. The paragraph concludes by positioning the study's key empirical finding — that institutional trust, collective efficacy, and environmental quality are among the most powerful levers for reducing perceived insecurity — as both a research contribution and a call for a fundamentally different approach to community safety policy. The revised Conclusions section can be found in Section 12 of the manuscript.



Q27. In addition, please review the entire manuscript to ensure that statements are not overgeneralized or overstating conclusions that cannot be made based on the study’s findings.

 

A27. We thank the Reviewer for this important methodological observation. We have conducted a thorough review of the entire manuscript and identified and corrected six categories of overgeneralized or epistemically imprecise statements. First, causal language implying that the study demonstrates rather than identifies associations has been systematically replaced throughout. The phrase "demonstrating that institutional legitimacy and social capital are central protective factors" has been revised to "suggesting that institutional legitimacy and social capital may act as protective factors," correctly reflecting that the study identifies regional-level associations rather than causal mechanisms. Second, deterministic policy language has been softened to reflect the inferential nature of the policy implications. Phrases such as "can directly lower," "can strengthen," and "can reduce" have been replaced with "may help reduce," "may strengthen," and "may reduce," acknowledging that the policy implications are informed inferences from observational data rather than guaranteed effects. Third, absolute superlatives have been qualified. The phrase "the most powerful levers available" has been revised to "among the most promising levers available, based on the evidence presented here," correctly limiting the claim to the scope of this study rather than implying a universal ranking of all possible interventions. Fourth, causal framing in the policy section has been corrected. The phrase "perceived insecurity is driven primarily by" has been replaced with "perceived insecurity is most strongly associated with," which accurately reflects the associational nature of the findings. Fifth, language implying independent causal effects in Table 7 has been revised. "Independently increases PRC" has been replaced with "is independently associated with higher PRC" for both the landscape dissatisfaction and negative future outlook variables. Sixth, the phrase "is the strongest determinant of perceived crime risk" in Section 8 has been qualified to "emerges as the strongest determinant of perceived crime risk in this study," limiting the claim to the specific dataset and analytical framework used here. All revised statements now accurately reflect the observational and regional nature of the evidence base.



Q28. In the Discussion section, for this study to make a clear contribution that is translatable for both practitioners and policymakers, it would be helpful if the authors provided some of the main “takeaway” points of the entire study in the Discussion section. Overall, this section is an opportunity for the authors to think critically about the findings and provide a detailed discussion, and more depth is needed across this section.

 

A28. We thank the Reviewer for this substantive and constructive observation. We agree that the original Discussion section was insufficiently developed and did not provide the critical depth or practitioner-oriented takeaways that a study of this kind should offer. We have substantially revised and expanded Section 10 in response.

The revised Discussion now integrates four concrete takeaways for practitioners and policymakers directly within the thematic paragraphs, rather than treating them as a separate addendum. Each takeaway is derived from a specific empirical finding and is framed in terms of what it means for those responsible for designing and implementing community safety policy.

The first takeaway, embedded in the discussion of pickpocketing and fear of crime, is that measuring crime rates alone is an insufficient basis for evaluating public safety. Perceived insecurity is a distinct outcome that requires its own indicators and intervention logic, and the ISTAT-BES framework already provides the tools to monitor it at the regional level.

The second takeaway, embedded in the discussion of institutional trust, is that how institutions behave matters as much as how much crime occurs. A fair and responsive police force and a judicial system that resolves cases within a reasonable time are measurable contributors to citizens' sense of safety — and investments in procedural justice and civic accountability are therefore central to crime prevention, not peripheral to it. The discussion of the dual role of social participation also highlights a critical distinction for policymakers: short-term civic engagement programs that do not build sustained collective ownership of safety may inadvertently amplify perceived risk rather than reduce it.

The third takeaway, embedded in the discussion of environmental dissatisfaction and socio-economic pessimism, is that perceived insecurity is partly a symptom of how citizens feel about their collective future. Investing in visible neighborhood renewal and credible narratives of local progress is therefore a public safety intervention, not merely a quality-of-life measure.

The fourth takeaway, embedded in the discussion of the cluster analysis, is that there is no single Italian model of perceived insecurity. The eight regional profiles show that the same level of PRC can arise from very different combinations of risk factors, and that effective policy responses must be locally calibrated rather than nationally uniform.

In addition, the synthesis paragraph has been strengthened to draw out the critical implication that perceived insecurity requires a correspondingly multidimensional institutional response, and the directions for future research have been revised to flow naturally from the empirical findings rather than appearing as a mechanical list. The revised Discussion can be found in Section 10 of the manuscript.

 

 

Author Response File: Author Response.docx

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

Social Sciences R&R Article Review

Revised Title: Crime, Trust, and Quality of Life: Determinants of Perceived Insecurity across Italian Regions

Original Title: Integrated Evidence on the Determinants of Perceived Crime Risk and Implications for Preventive Community Safety Policies

The authors of this revised article examine various variables that may predict the Perceived Risk of Crime (PRC) measure. Using data from the ISTAT-BES system, results from panel regression models and supplemental analyses indicated that pickpocketing crimes and fear of crime are significant predictors of a perceived risk of crime.

After reading the revised manuscript and the authors’ response to the reviewer feedback, it is apparent that the authors considered the comments I made and revised their manuscript accordingly. For example, the Introduction section was revised to provide transitions across thoughts, arguments, and topics, along with formally introducing the research questions. The Literature Review was revised to not “grocery list” the findings and now provides a synthesis of the prior literature. The revised manuscript now also includes a “Current Study” section based on my suggestions, and implemented suggested details to the Methodology, streamlined (to some degree) the Results section, and included additional study limitations. Revisions were also made to make specific policy implications based on the study findings, and minor concerns have been fixed.

Moving forward, I believe that some of the tables and figures are still not needed. For example, Table 6 and Table 7 are “summary tables”, but is redundant from the presentation of the results and summary paragraphs. The same can also be said for Table 1 and Figure 4 (especially since the font for Figure 4 is so small). In addition, Section 4 is one large paragraph (over 2-3 pages). Please divide and revise into multiple paragraphs to ease readability. This issue also occurs in Section 6, Section 8, Section 9, and other sections. Please revise the entire manuscript to not have sections that are single paragraphs.

Author Response

Point to Point Answers to Reviewer 2

 

Q1. Revised Title: Crime, Trust, and Quality of Life: Determinants of Perceived Insecurity across Italian Regions. Original Title: Integrated Evidence on the Determinants of Perceived Crime Risk and Implications for Preventive Community Safety Policies.

A1.  Dear editor during the review process we decided to change the title from “Crime, Trust, and Quality of Life: Determinants of Perceived Insecurity across Italian Regions” to “Integrated Evidence on the Determinants of Perceived Crime Risk and Implications for Preventive Community Safety Policies”.


Q2. The authors of this revised article examine various variables that may predict the Perceived Risk of Crime (PRC) measure. Using data from the ISTAT-BES system, results from panel regression models and supplemental analyses indicated that pickpocketing crimes and fear of crime are significant predictors of a perceived risk of crime. After reading the revised manuscript and the authors’ response to the reviewer feedback, it is apparent that the authors considered the comments I made and revised their manuscript accordingly. For example, the Introduction section was revised to provide transitions across thoughts, arguments, and topics, along with formally introducing the research questions. The Literature Review was revised to not “grocery list” the findings and now provides a synthesis of the prior literature. The revised manuscript now also includes a “Current Study” section based on my suggestions, and implemented suggested details to the Methodology, streamlined (to some degree) the Results section, and included additional study limitations. Revisions were also made to make specific policy implications based on the study findings, and minor concerns have been fixed.

A2. We sincerely thank the reviewer for the positive assessment of our revisions and for the careful and constructive feedback provided throughout the review process. We are very glad that the changes made to the Introduction, the Literature Review, the new "Current Study" section, the Methodology, the Results, the study limitations, and the policy implications are considered appropriate and responsive to the previous comments. Your guidance has substantially strengthened the manuscript, and we have continued to build on it in addressing the remaining points raised in this round.


Q3. Moving forward, I believe that some of the tables and figures are still not needed. For example, Table 6 and Table 7 are “summary tables”, but is redundant from the presentation of the results and summary paragraphs. The same can also be said for Table 1 and Figure 4 (especially since the font for Figure 4 is so small). In addition, Section 4 is one large paragraph (over 2-3 pages). Please divide and revise into multiple paragraphs to ease readability. This issue also occurs in Section 6, Section 8, Section 9, and other sections. Please revise the entire manuscript to not have sections that are single paragraphs.

A3. We thank the reviewer for these helpful suggestions, which have improved the conciseness and readability of the manuscript. Regarding the redundant tables and figures, we have removed Table 6 and Table 7, since the integrated summary of results and the policy implications are already covered in the corresponding paragraphs. For the same reason, we removed Table 1 (literature mapping), as the synthesis of the literature is now fully developed in the Literature Review narrative. Figure 4 was also removed, both for its overlap with the text and for the legibility issues noted. All remaining tables and figures have been renumbered consecutively, and every in-text cross-reference updated accordingly. Regarding the single-paragraph sections, we have divided Section 4 into multiple paragraphs organized by logical unit, and applied the same revision to Sections 6, 8, and 9. As requested, we also reviewed the entire manuscript to eliminate any remaining single-paragraph sections, dividing Section 2 (Literature Review) and Section 3 (Current Study) accordingly. These changes substantially improve the paper's readability.

 

Author Response File: Author Response.docx