Startups in a Remote Region: Policy Implications from the Israeli Experience

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The manuscript addresses a relevant and timely question on whether (and for which types of ventures) peripheral regions can successfully attract startups. The paper states a clear objective (2.4) and two testable hypotheses (3), and it implements a coherent comparative design (Tel-Aviv center vs. Beer-Sheva periphery; n=202) with appropriate statistics (χ², Mann-Whitney, logistic regression with odds ratios). Tables synthesize the findings clearly.

What stands out (strengths):

• Clear research framing (explicit contribution and objective; H1-H2) and a transparent set of ecosystem factors.
• Consistent comparative evidence: e.g., government financial support shows higher availability in the periphery and a strong association with peripheral location in the logits.

How to sharpen the contribution (actionable edits)

• Make the policy message “fit-based” and explicit.
  Your own ORs allow a clear triage: peripheral incentives most effectively “pull” R&D-based (OR≈4.30), open-innovation (~2.22), and international-link startups (~2.27). Say this up front in the Conclusions/Abstract and propose prioritization of these types for peripheral instruments (e.g., matching grants, consortia projects, soft-landing for R&D sites).
• State the boundary condition of agglomeration forces.
  The collaboration with suppliers/consumers factor systematically pushes firms toward the center (OR<1, significant in multiple cuts). Name this as a hard constraint that policy must acknowledge (and can only partially relax through ICT/transport).
• Elevate the counter-intuitive finding on talent.
  Median access to skilled workers is similar in the center and periphery (often Mdn=4 for both), which runs against a common stereotype. Flag this clearly and discuss the Israeli distance context you note in the text. This is a publishable insight of its own.
• Tighten causal language around “market failure.”
  You currently write that the pattern “indicates the existence of regional market failure.” Given the observational design, consider reframing to “policy-addressable availability gap” or “policy-sensitive dependence”; preserve the strong result without over-claiming causality.
• Increase sampling transparency and interviewer controls.
  Specify the sampling frame and the randomization procedure behind the “202 randomly selected startups,” report the response rate, and briefly describe how the 49 interviewers were trained and monitored to limit interviewer bias. This will substantiate your “random” claim.
• Add robustness and interpretability:

• • Include controls for age/size in the logits (you already show balance tests; adding them as covariates reduces omitted-variable concerns).
  • Report marginal effects alongside ORs (improves managerial readability).
  • Consider a pooled model with interactions (availability × startup type), clustered SEs, and (if feasible) a reduced-dimension variant to address multicollinearity (PCA or regularization).
• Add a one-figure 2×2 policy map:
  X-axis: reliance on proximity (suppliers/consumers); Y-axis: reliance on government support. Place your typologies into quadrants (e.g., R&D/open/international in high-Y; incremental/open in high-X). This will make your policy guidance immediately actionable. (Based on your Tables 2, 3, and discussion.)
• Minor editorial fixes for clarity:
  • Correct “Table 31” → “Table 3.”
  • Add a brief reading guide to Table 3 (what OR>1 vs. <1 means here).
• Ethics/data statements.
  You already include a Data Availability Statement; add a one-sentence note on informed consent/anonymity for respondents.

Author Response

Comments 1:

Make the policy message “fit-based” and explicit.

Your own ORs allow a clear triage: peripheral incentives most effectively “pull” R&D-based (OR≈4.30), open-innovation (~2.22), and international-link startups (~2.27). Say this up front in the Conclusions/Abstract and propose prioritization of these types for peripheral instruments (e.g., matching grants, consortia projects, soft-landing for R&D sites).

Response 1:

Thank you for this helpful suggestion. We have revised the manuscript accordingly and made the policy message more explicit. The triage of R&D-based, open-innovation, and internationally oriented startups is now clearly presented in the Discussion and emphasized at the beginning of the Conclusions, along with examples of relevant policy instruments.

Comments 2:

State the boundary condition of agglomeration forces.

The collaboration with suppliers/consumers factor systematically pushes firms toward the center (OR<1, significant in multiple cuts). Name this as a hard constraint that policy must acknowledge (and can only partially relax through ICT/transport).

Response 2:

Thank you for this important comment. We have revised the manuscript accordingly and explicitly addressed the role of collaboration with suppliers and consumers as a boundary condition of agglomeration forces. This point is now clearly emphasized in the Conclusions section, noting that while policy measures such as ICT improvements and enhanced transport infrastructure may partially mitigate the constraint, they cannot fully eliminate it.

Comments 3:

Elevate the counter-intuitive finding on talent.

Median access to skilled workers is similar in the center and periphery (often Mdn=4 for both), which runs against a common stereotype. Flag this clearly and discuss the Israeli distance context you note in the text. This is a publishable insight of its own.

Response 3:

Thank you for this valuable observation. We have revised the text to emphasize the similarity in access to skilled workers between the center and periphery and framed it in the Israeli context, where short commuting distances and high mobility help explain this pattern. We also added a reference to recent work on labor mobility in Israel [20] to support this point.

Comments 4:

Tighten causal language around “market failure.”

You currently write that the pattern “indicates the existence of regional market failure.” Given the observational design, consider reframing to “policy-addressable availability gap” or “policy-sensitive dependence”; preserve the strong result without over-claiming causality.

Response 4:

Thank you for this important clarification. We have revised the language accordingly and now describe the finding as “a policy-sensitive dependence on government support” rather than referring to it as a “market failure”.

Comments 5:

Increase sampling transparency and interviewer controls.

Specify the sampling frame and the randomization procedure behind the “202 randomly selected startups,” report the response rate, and briefly describe how the 49 interviewers were trained and monitored to limit interviewer bias. This will substantiate your “random” claim.

Response 5:

Thank you for this important suggestion. We have revised the Methods section to specify the sampling frame and procedure: the 202 startups were randomly selected by the student interviewers from a list of firms provided by the supervising professors. We also added a description of how the 49 interviewers were trained and closely supervised by the professors throughout the entire data collection process to ensure consistency and limit interviewer bias.

Comments 6:

Add robustness and interpretability:

Include controls for age/size in the logits (you already show balance tests; adding them as covariates reduces omitted-variable concerns).

Report marginal effects alongside ORs (improves managerial readability).

Consider a pooled model with interactions (availability × startup type), clustered SEs, and (if feasible) a reduced-dimension variant to address multicollinearity (PCA or regularization).

Response 6:

We thank the reviewer for this constructive suggestion. As recommended, we have included control variables for firm age and size in the logistic regressions, and the results are reported in Table 3. To improve interpretability, we also provide the average marginal effects (AME) in parentheses alongside the odds ratios (OR). The text in both the methodology and the results sections has been revised accordingly.

Unfortunately, it is not feasible to estimate a pooled model with interactions between availability and startup type given the limited number of observations relative to the large number of interaction terms required. Moreover, our objective is not to test whether specific cutoffs drive the results, but rather to examine how different ecosystem factors influence the likelihood of startups locating in the periphery within each state. For this reason, we focus on estimating and interpreting the separate effects of the factors for each type of startup.

Comments 7:

Add a one-figure 2×2 policy map:

X-axis: reliance on proximity (suppliers/consumers); Y-axis: reliance on government support. Place your typologies into quadrants (e.g., R&D/open/international in high-Y; incremental/open in high-X). This will make your policy guidance immediately actionable. (Based on your Tables 2, 3, and discussion.)

Response 7:

We thank the reviewer for this excellent suggestion. In line with your recommendation, we have added a 2×2 policy map (now Figure 2) that positions the startup typologies according to their reliance on proximity and government support. This figure has been inserted and described at the end of the Discussion section, where it serves to visually synthesize the main findings and clarify the policy implications.

Comments 8:

Minor editorial fixes for clarity:

Correct “Table 31” → “Table 3.”

Response 8:

We appreciate the reviewer’s careful reading. The reference has been corrected as suggested.

Add a brief reading guide to Table 3 (what OR>1 vs. <1 means here).

Thank you for the suggestion. We added a brief reading guide when presenting Table 3 (in the Results), clarifying how to interpret the OR (>1 periphery; <1 center) and AME.

Comments 9:

Ethics/data statements.

You already include a Data Availability Statement; add a one-sentence note on informed consent/anonymity for respondents.

Response 9:

Thank you for the comment. We added an ethics statement under the Data Availability Statement: “All participants provided informed consent prior to the survey, and responses were collected and analyzed anonymously.”

Reviewer 2 Report

Comments and Suggestions for Authors

This manuscript presents a timely and empirically grounded investigation into the locational dynamics of startups in Israel, contrasting the core (Tel Aviv) with the periphery (Beer-Sheva).The minor revisions suggested above primarily concern clarifications, emphasis on limitations, and improved presentation of complex results.

Specific comments:

1. Sample Size Imbalance: The authors correctly identify the small sample size from the periphery (n=45) as a limitation. While the statistical methods are robust, some sub-group analyses in Table 3 have very low observations. This should be more strongly emphasized as a limitation, suggesting that the findings, particularly for the less common startup types, should be seen as indicative and warrant further validation with a larger sample.

2. Clarification of Startup Typologies: The definitions and operationalization of the six startup typologies could be briefly expanded upon in the methodology section (Section 4). For instance, how was "open innovation" defined and identified in the survey? Was it based on formal partnerships, outsourcing of R&D, etc.? A sentence or two for each type would enhance reproducibility.

3. Discussion of Unexpected Findings: Two surprising results deserve more discussion: 1) High Skill Availability in Periphery: The explanation (short distances in Israel) is plausible but deserves more nuance. Is this specific to Beer-Sheva, which hosts a major university (BGU) and has good transport links to Tel Aviv? This might not hold in larger or more isolated countries, a point worth noting to contextualize the generalizability of this particular finding. 2) Low University-Industry Collaboration: The discussion in the results (Page 11) is excellent, but it should be echoed in the main discussion section. This finding is counter-intuitive given Israel's high ranking and could be a standalone point about the difference between national potential and localized reality.
4. Table 3 Formatting and Readability: The header says "Table 31". Please correct to "Table 3". Also, check the N and significance values for consistency (e.g., the model for "Disruptive -> No" has an N of 131 but a Chi-Sq of 33.78, which seems high and significant, but the reported significance is (.00) while the pseudo R2 is 0.35. The row for "No" under "Disruptive" seems to have the values for "Yes" and vice versa? Please double-check the data placement in all columns).

Comments on the Quality of English Language

The English could be improved to more clearly express the research.

Author Response

Comments 1:

Sample Size Imbalance: The authors correctly identify the small sample size from the periphery (n=45) as a limitation. While the statistical methods are robust, some sub-group analyses in Table 3 have very low observations. This should be more strongly emphasized as a limitation, suggesting that the findings, particularly for the less common startup types, should be seen as indicative and warrant further validation with a larger sample.

Response 1:

We thank the reviewer for this important comment. We have revised the Limitations paragraph to explicitly acknowledge the sample size imbalance and to note that some subgroup analyses rely on low counts. We now emphasize that these findings should be regarded as indicative and require further validation with larger, more representative samples.

Comments 2:

Clarification of Startup Typologies: The definitions and operationalization of the six startup typologies could be briefly expanded upon in the methodology section (Section 4). For instance, how was "open innovation" defined and identified in the survey? Was it based on formal partnerships, outsourcing of R&D, etc.? A sentence or two for each type would enhance reproducibility.

Response 2:

Thank you for the suggestion. We have added a concise paragraph in Section 4 (immediately after introducing the six properties) that briefly defines and operationalizes the six startup typologies. This clarifies measurement and improves reproducibility.

Comments 3:

Discussion of Unexpected Findings: Two surprising results deserve more discussion: 1) High Skill Availability in Periphery: The explanation (short distances in Israel) is plausible but deserves more nuance. Is this specific to Beer-Sheva, which hosts a major university (BGU) and has good transport links to Tel Aviv? This might not hold in larger or more isolated countries, a point worth noting to contextualize the generalizability of this particular finding. 2) Low University-Industry Collaboration: The discussion in the results (Page 11) is excellent, but it should be echoed in the main discussion section. This finding is counter-intuitive given Israel's high ranking and could be a standalone point about the difference between national potential and localized reality.

Response 3:

We thank the reviewer for these valuable suggestions. In the Discussion section, we have added two new paragraphs. The first expands on the surprising finding regarding skilled labor availability, situating it in the Israeli context of short commuting distances and high mobility, and noting its limited generalizability to larger or more isolated countries. To strengthen this point, we also referred readers to recent work on labor mobility in Israel [20]. The second echoes the low level of university–industry collaboration observed in the Results, emphasizing the gap between Israel’s high national ranking and the more limited regional reality, and drawing out the policy implications of this discrepancy.

 

Comments 4:

Table 3 Formatting and Readability: The header says "Table 31". Please correct to "Table 3". Also, check the N and significance values for consistency (e.g., the model for "Disruptive -> No" has an N of 131 but a Chi-Sq of 33.78, which seems high and significant, but the reported significance is (.00) while the pseudo R2 is 0.35. The row for "No" under "Disruptive" seems to have the values for "Yes" and vice versa? Please double-check the data placement in all columns).

Response 4:

We thank the reviewer for this careful reading. The typo in the table header has been corrected. In addition, we re-ran the models and thoroughly revised Table 3. As a result, all numbers, including the N and significance values, have been updated and are now correct.

Reviewer 3 Report

Comments and Suggestions for Authors

This is a promising paper that explores the spatial distribution of startups in peripheral regions and analyzes the impact of ecosystem factors on location decisions of different types of startups based on empirical data from Israel. Here are some comments that I hope will help improve the quality of the article.

 

The authors' introduction is too brief to resemble an academic paper. After reading the introduction, readers should understand what problem the article seeks to address.

 

The manuscript provides valuable empirical evidence, but the theoretical innovation is relatively limited, primarily validating existing theories rather than proposing new frameworks. I suggest the authors further elaborate on their theoretical contributions, particularly regarding the dependency relationship between startups and government support, which represents a valuable theoretical contribution. The authors could compare their findings with the following study to enrich understanding of startup behavioral patterns: Wąsowska, A., Obłój, K., & Kopiński, D. (2024). Motives, challenges, and strategies of CEE companies in (very) distant markets. The case of Polish firms in Sub-Saharan Africa. JEEMS Journal of East European Management Studies, 29(4), 723-737. This research emphasizes how resource-constrained enterprises rely on "relational" and "entrepreneurial" competitive advantages to overcome geographical distance challenges, contrasting with this paper's finding that peripheral startups depend more heavily on government support.

 

The sample size is imbalanced, with 157 companies in the center and 45 in the periphery, which may lead to insufficient statistical power and bias. While the paper acknowledges this limitation, it does not adequately discuss its implications. The smaller peripheral sample may not represent all types of startups, leading to insufficient statistical significance in certain subgroup analyses (such as startups within science parks).

 

The research design is a cross-sectional survey, making it difficult to establish causal relationships between variables. Although associations were found, it cannot be determined whether ecosystem factors influenced location choice or whether location choice influenced perceptions of ecosystem factors. This issue requires authors' consideration and discussion in the text.

 

Table 3 is marked as "Table 31," which is a typographical error.

Author Response

Comments 1: 

The authors' introduction is too brief to resemble an academic paper. After reading the introduction, readers should understand what problem the article seeks to address.

Response 1:

 We thank the reviewer for this constructive feedback. We agree with the assessment and have revised the introduction to more clearly articulate the specific problem the paper addresses. A new paragraph has been added after the first to explicitly frame the policy challenge our research seeks to resolve, thereby establishing a clearer rationale for the study. We believe this revision fully addresses the reviewer's concern and we are grateful for the suggestion.

Comments 2: 

The manuscript provides valuable empirical evidence, but the theoretical innovation is relatively limited, primarily validating existing theories rather than proposing new frameworks. I suggest the authors further elaborate on their theoretical contributions, particularly regarding the dependency relationship between startups and government support, which represents a valuable theoretical contribution. The authors could compare their findings with the following study to enrich understanding of startup behavioral patterns: Wąsowska, A., Obłój, K., & Kopiński, D. (2024). Motives, challenges, and strategies of CEE companies in (very) distant markets. The case of Polish firms in Sub-Saharan Africa. JEEMS Journal of East European Management Studies, 29(4), 723-737. This research emphasizes how resource-constrained enterprises rely on "relational" and "entrepreneurial" competitive advantages to overcome geographical distance challenges, contrasting with this paper's finding that peripheral startups depend more heavily on government support.

Response 2:

 Thank you for this constructive comment. We clarified our theoretical contribution by explicitly framing the startup–government support relationship as a contingency mechanism and by adding a two-dimensional policy map (Figure 2) that positions startup typologies by their relative reliance on government support versus proximity-based ecosystem factors. The figure provides a concise synthesis of the analytical results and serves as the basis for our policy implications. We also added a brief citation in the Conclusions (end of the first paragraph) to situate our interpretation within the broader literature—stating that our findings are consistent with related evidence on firms’ adaptation to distance (Wąsowska, Obłój, & Kopiński, 2024)—and updated the reference list accordingly.

Comments 3:

The sample size is imbalanced, with 157 companies in the center and 45 in the periphery, which may lead to insufficient statistical power and bias. While the paper acknowledges this limitation, it does not adequately discuss its implications. The smaller peripheral sample may not represent all types of startups, leading to insufficient statistical significance in certain subgroup analyses (such as startups within science parks).

Response 3:

 We thank the reviewer for this important comment. We have revised the Limitations paragraph to explicitly acknowledge the center–periphery sample-size imbalance and to note that some subgroup analyses rely on low counts. We now emphasize that these findings should be regarded as indicative and require further validation with larger, more representative samples.

Comments 4:

The research design is a cross-sectional survey, making it difficult to establish causal relationships between variables. Although associations were found, it cannot be determined whether ecosystem factors influenced location choice or whether location choice influenced perceptions of ecosystem factors. This issue requires authors' consideration and discussion in the text.

Response 4:

 We thank the reviewer for raising this important methodological point. We agree that the cross-sectional nature of our study limits our ability to make definitive causal claims. To address this, we have added a new paragraph to the limitations section of our manuscript. This new text explicitly acknowledges that our study establishes associations rather than causality and discusses the possibility of reverse causality. We also suggest that future longitudinal research would be valuable for tracking location decisions over time to untangle these complex relationships. We believe this addition makes the boundaries of our research clearer and strengthens the paper.

Comments 5: 

Table 3 is marked as "Table 31," which is a typographical error.

Response 5:

We thank the reviewer for this careful reading. The typo in the table header has been corrected.

Reviewer 4 Report

Comments and Suggestions for Authors

1. Background and Significance of the Study

 This paper accurately captures the core issue of "spatial imbalance in the distribution of startups", a contextualization that is both universal and specific, given the global regional commonality of the problem and Israel's unique innovation policies (e.g., the Capital Investment Incentives Law). This paper fills the gap in the existing literature that focuses on the "overall characteristics of regional innovation" or the "behavior of a single type of startup", and for the first time systematically analyzes the "interaction between different types of startups and regional ecosystem factors", and in particular, identifies the role of R&D in the development of regional innovation. This paper is the first to systematically analyze "the interaction mechanism between different types of startups and regional ecosystem factors", and in particular, it clarifies the differentiated needs of R&D-intensive, open and innovative startups and other segmented startups for ecological factors, which enriches the theory of regional innovation ecology and entrepreneurial geography. Based on Israel, this paper reveals the impact of key factors such as support and supply chain collaboration on startups' location choice, and provides an empirical reference for innovation policy in remote regions around the globe - an actionable policy framework for mitigating the "core-periphery" innovation imbalance.

1. Research Methods

 This paper adopts a comparative research design for sample selection, selecting 202 random samples (157 in the central area and 45 in the remote area), controlling for the "intermediate area interference effect", and verifying that the samples do not differ significantly in the age of the startups by chi-square test with sample matching. However, the sample in remote areas is only 45 (22%), and the sample size of some segmentation types (e.g., "startups with international links") may be insufficient or affect the statistical validity; it is not stated whether the weighting method is used to balance the sample bias. Attention should also be paid to data control. It is mentioned that "49 graduate students implemented the face-to-face questionnaire", but the results of questionnaire reliability and validity tests (e.g., Cronbach's alpha coefficient, KMO test) are not reported, and it is not stated how to control interviewer bias (e.g., standardized interview guidelines, cross-checking), and the data The reliability of the data needs to be further supported.

• Data results and discussion

The discussion in this paper is closely aligned with the literature, such as echoing "the critical role of support" with Rodríguez-Pose et al.'s (2015) theory of "influencing regional innovation", and "the central advantage of supply chain collaboration " with Krugman's (1991) "economic theory of agglomeration", which effectively reconciles the contradiction between "innovation in remote areas is dependent on central resources" and "autonomous innovation is possible" in previous studies. The contradiction between "innovation in remote areas is dependent on central resources" and "autonomous innovation is possible" in previous studies is effectively reconciled. However, the explanation for the unintended consequence - "sufficient skilled workers in remote areas" is attributed only to "short distances" - is inadequate, and does not take into account Israel's specific talent policies (e.g., tax incentives for remote talent) or commuting data to persuade the public to take a closer look at the results of the study. The contradiction of "low level of university-enterprise collaboration (median < 3 in both regions) but high Israeli GII ranking" is only mentioned in the context of "more startup collaboration in science parks" without analyzing in depth the difference between national and regional collaboration networks, and the depth of the discussion needs to be deepened. It does not analyze the difference between the national collaboration network and the regional collaboration, and the depth of the discussion needs to be improved.

1. Research Conclusions

 The conclusion closely echoes the research objective of "analyzing the differential response of different types of startups to ecosystem factors", and clearly puts forward that "startup location is not a single issue, but is jointly determined by the type of innovation, the degree of ecological factors, and the availability of regional factors". However, the conclusion part does not clearly distinguish between "statistical findings" and "practical inferences", for example, "support is important for remote startups" is directly inferred to "support for R&D startups should be prioritized", and "support for R&D startups" is directly inferred to "support for R&D startups should be prioritized". For example, "support is important for remote startups" is directly inferred as "priority should be given to supporting R&D startups". Although there are data to support this, the logic of derivation from "statistical significance" to "policy prioritization" needs to be more clearly explained.

1. Policy Recommendations

The policy recommendations are closely related to the research results and have strong relevance, but they do not specify the specific design of the policy tools, such as the "R&D Startup Priority Support", which does not specify the percentage of subsidy and the application conditions, and does not propose an evaluation system of the policy effect, such as how to measure the "ecosystem building effectiveness "(the survival rate of startups, the growth rate of R&D inputs, etc.), which is not conducive to the dynamics of the policy after implementation.

1. Research Limitations and Future Prospects

The authors propose "expanding the sample of remote areas and ecological factor variables", which is a feasible direction; further additions can be made - ① conducting cross-country comparative studies (e.g., comparing Finland, Canada, and other countries with the same regional innovation gap) to enhance the generalizability of the conclusions; ② using panel data to track the dynamics of startups' site selection and enhance the ability of causal analysis; ③ Analyze in-depth the "Enabling Mechanisms of Science Parks for Remote Startups" to provide more specific guidance for the construction of the parks.

• Overall Evaluation

 The topic of this paper has both theoretical and practical value, accurately responds to the global problem of "spatial distribution imbalance of startups", and the uniqueness of the Israeli case enhances the reference value of the study, providing a clear empirical basis for innovation policy in remote areas. However, the paper needs to be improved in the following aspects:

• Add the sample size of remote areas
• Add the results of reliability tests (e.g., Cronbach's alpha coefficient, KMO test) for the "face-to-face questionnaire administered to 49 graduate students", and explain how to control interviewer bias (e.g., standardized interview guidelines, cross-checking).
• Is the explanation of "sufficient number of skilled workers in remote areas" attributed only to "short distance" convincing enough?
• For the contradiction of "low level of school-enterprise collaboration (median < 3 in both regions) but high GII ranking in Israel", only "more startup collaboration in science parks" is mentioned, should we analyze more deeply the difference between national and regional collaboration networks? Is it necessary to analyze more deeply the differences between national and regional collaboration networks?
• The direct inference from "support is important for remote startups" to "support for R&D startups should be prioritized" is supported by the data, but the "statistical significance" to "policy prioritization" needs to be more clearly explained. Although there is data support, the logic from "statistical significance" to "policy prioritization" needs to be clarified.
• Add clarification on the design of policy tools, e.g. "Priority support for R&D start-ups" does not specify the percentage of subsidy and the application conditions, and add a system to evaluate the effect of the policy.

Author Response

Comments 1: 

Add the sample size of remote areas

Response 1:

We thank the reviewer for this point. The sample size for the remote peripheral region (n=45) is now explicitly stated in the Abstract (page 1), the Materials and Methods section (page 6), and is also noted in Table 1 (page 8). In addition, we now reiterate this number in the Conclusions section (page 16) to ensure clarity and consistency throughout the manuscript.

Comments 2: 

Add the results of reliability tests (e.g., Cronbach's alpha coefficient, KMO test) for the "face-to-face questionnaire administered to 49 graduate students", and explain how to control interviewer bias (e.g., standardized interview guidelines, cross-checking).

Response 2:

We thank the reviewer for this important comment. Since our survey consisted primarily of single-item factual questions (e.g., startup type, location, availability of specific ecosystem factors), Cronbach’s alpha is not applicable. To assess the robustness of the variables used in the regression models, we employed variance inflation factor (VIF) tests and confirmed that multicollinearity is not a concern; we now state this explicitly in the paper. In addition, we expanded the Materials and Methods section to clarify how we ensured reliability and minimized interviewer bias: the 49 student interviewers received structured training sessions and were accompanied and supervised by the professors throughout the entire data collection process.

Comments 3-4: 

• Is the explanation of "sufficient number of skilled workers in remote areas" attributed only to "short distance" convincing enough?
• For the contradiction of "low level of school-enterprise collaboration (median < 3 in both regions) but high GII ranking in Israel", only "more startup collaboration in science parks" is mentioned, should we analyze more deeply the difference between national and regional collaboration networks? Is it necessary to analyze more deeply the differences between national and regional collaboration networks?

Response 3-4:

We thank the reviewer for both of these valuable comments. In the Discussion section, we have added two new paragraphs. The first expands on the surprising finding regarding skilled labor availability, situating it in the Israeli context of short commuting distances and high mobility, and noting its limited generalizability to larger or more isolated countries. To strengthen this point, we also referred readers to recent work on labor mobility in Israel [20]. The second echoes the low level of university–industry collaboration observed in the Results, emphasizing the gap between Israel’s high national ranking and the more limited regional reality, and drawing out the policy implications of this discrepancy.

Comments 5: 

The direct inference from "support is important for remote startups" to "support for R&D startups should be prioritized" is supported by the data, but the "statistical significance" to "policy prioritization" needs to be more clearly explained. Although there is data support, the logic from "statistical significance" to "policy prioritization" needs to be clarified.

Response 5:

To strengthen the connection between our quantitative results and their policy implications, we have added a sentence in the Discussion section. This new sentence explicitly states that the statistical strength of our findings—specifically the large effect sizes and high sensitivity to government support for certain startups—provides the direct justification for our policy recommendation to prioritize these groups. We believe this makes the logical step from our empirical findings to our policy recommendations clearer and more direct.

Comments 6: 

Add clarification on the design of policy tools, e.g. "Priority support for R&D start-ups" does not specify the percentage of subsidy and the application conditions, and add a system to evaluate the effect of the policy.

Response 6:

We thank the reviewer for this constructive suggestion. To strengthen the policy relevance of our analysis, we added a two-dimensional policy map (Figure 2) that positions startup typologies by their reliance on government support and proximity-based ecosystem factors. In addition, in the Conclusions section we included a sentence noting that such support should be accompanied by a monitoring and evaluation framework (e.g., tracking startup survival, innovation output, and collaboration rates) to assess its effectiveness and allow for policy adjustments.

Round 2

Reviewer 3 Report

Comments and Suggestions for Authors

The author's revisions were timely and addressed my concerns. I agree to publish the current version.

Reviewer 4 Report

Comments and Suggestions for Authors

The author has revised the document very well, and I think it meets the requirements for publication.