Analyzing Financial Behavior in Undergraduate Students in Economics, Administration and Accounting Sciences

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

1. The title states graduate students, but further along the text and in the methods, the entire study is described as performed on undergraduate students (bachelor's degree).
2. The authors write that the behavior has been "widely studied," but "significant gaps remain," but they do not articulate which gap their work closes, what is the novelty of the work.
3. The literature review is of a general nature and is not shown by the research gap.
4. The authors did not establish a connection with previous validations of FB-13. There is no brief comparison with the early results (where a single-factor structure was reported).: It does not show which items and loads coincide/diverge, how key quality metrics are comparable, and why a different (three-factor) structure appeared in your sample.
5. There are inconsistencies in the Methods section: the target population is indicated differently (graduate vs undergraduate), EFA and CFA, judging by the description, were performed on the same sample. The interpretation of fit indexes is incorrect in some places, there are no checks of multigroup invariance (gender/course) and extended reliability metrics. A description of the selection is required.
6. The authors chose a three-factor model, but did not compare it with the promised one-factor model, so it is not clear why it is better. Some of the compliance indicators are called "good", although this is an average level in terms of values, plus there are inconsistencies in the numbers and χ2 is applied where the expected frequencies are low.

Author Response

Dear Reviewer, we have taken your valuable comments into account to improve our manuscript. We have highlighted all the changes requested by the team of three reviewers in green.

C1) The title states graduate students, but further along the text and in the methods, the entire study is described as performed on undergraduate students (bachelor's degree).
R1) We apologize for the mistake; the title has been corrected. The entire study refers to undergraduate students.

C2) The authors write that the behavior has been "widely studied," but "significant gaps remain," but they do not articulate which gap their work closes, what is the novelty of the work.
R2) We have reworded the introduction to provide the required clarifications.

C3) The literature review is of a general nature and is not shown by the research gap.
R3) We have reworded the introduction to identify the research gaps covered by our manuscript.

C4) The authors did not establish a connection with previous validations of FB-13. There is no brief comparison with the early results (where a single-factor structure was reported).: It does not show which items and loads coincide/diverge, how key quality metrics are comparable, and why a different (three-factor) structure appeared in your sample.
R4) We have provided further information regarding the previous study by Méndez-Prado et al., 2022, see Table 1 (factor loadings).

C5) There are inconsistencies in the Methods section: the target population is indicated differently (graduate vs undergraduate), EFA and CFA, judging by the description, were performed on the same sample. The interpretation of fit indexes is incorrect in some places, there are no checks of multigroup invariance (gender/course) and extended reliability metrics. A description of the selection is required.
R5) There are not two groups of students; all are undergraduate students. We have reported that the FACTOR software began its CFA analysis with a multigroup search (2 samples 357 (=714/2)), obtaining a communality index (S = 0.9962) as a result. Regarding a possible difference in distribution by gender, the Mann-Whitney U test did not show any differences by gender for any of the 13 items in FB-13.

C6) The authors chose a three-factor model, but did not compare it with the promised one-factor model, so it is not clear why it is better. Some of the compliance indicators are called "good", although this is an average level in terms of values,
R6) We have compared our results (with 3 factors) with the CFA for a single factor with 13 items, reporting the values in Table 8. This demonstrates that the proposed model is better than the original contrast model (with a single factor).

C7) plus there are inconsistencies in the numbers and χ2 is applied where the expected frequencies are low.
R7) We are very grateful for your valuable technical input. We have supplemented the Pearson Chi-Square Tests in Table 10 with the Freeman–Halton (F-H) extension of Fisher’s Exact Tests.

Reviewer 2 Report

Comments and Suggestions for Authors

Some corrections sought 

The research is intricately woven to examine the relationship between financial behaviour and the sociodemographics parameters using the very broader sample. When reading through the paper, I found the theme to be more appropriate in the current context but fails to point out the same across the introduction chapter. specifically; A theoritical framework and empirical evidence on the validity of the FB–13 instrument can be elaborated and literature review can be strengthened.  The examination of previous such studies with respect to EFA and CFA can be done in the beginning. There seems to be no prior theortical models discussed in the study which is major concern reading the first three chapters. Does having very diverse group for the study based on age and experience delimit the results obtained. Better to put this into perspective in the beginning. The literature reviews suggested in the paper are very specific in choice of the group. Kindly justify the advantage and limitations for having a unbalanced datasets with respect to many socio-demographic variables. The three factors identified in the study can be defined appropriately in the beginning of the CFA to provide better meaning when applied for result analysis. 

Apart from these points, few key points to consider are as follows;

1. Line 29 and Line 34 are too broad to make the conclusions. They do not reflect the results correctly. It can be rephrased to line with the objectives.
2. Future research should be in continuation to the results set in the study. They seem too diverge to draw the lines of future study. The scope of research is more on methodology rather than the results. Consider rephrasing the abstract.  
3.  The context of financial behaviour and its meaning can be elaborated to provide a context. The lines 43 to 62, can be seemingly examined from the point of view of various demographic and socio-economic perspectives. The sentences are more causal in nature and not seem to be descriptive. 
4. The study by patrisia et.al (2023) can be revisited to examine the results. The lines 65 to 67 use the word religious financial education which can be looked into. 
5.  Kindly refer to the paper based on which the lines 78 to 80 are opinioned. 
6. The lines 88 to 94 can be elaborated to include the key results of the countries, relationship between socio-demographic variables and the financial behaviour. Does the context of financial behaviour vary from country to country is the primary questions which is lacking in all the paragraphs. Better to introduce this theme in the beginning. 
7. The meaning across the various paragraphs fail to elaborate on the central theme of the study in the beginning. 

The context in the materials and method seem to be more apt in building the context in the introductory chapter rather than the second chapter. It is just a suggestion. 

  8. In the data analysis and processing procedures

  The null hypothesis is H01. Then alternative can be H11 instead of H1. 

  The null hypothesis for second is H02. Then alternative can be H12. 

9. The second hypothesis revolves around the correlations. When by default, the presence of identity matrix is defined by the bartlett test, how relevant would it be to include a broader hypothesis on the correlations. 

10. The results of the EFA can be eloborated to signify the results the factors identifed and the number of factors and naming of the factors to provide a context next for the CFA. Is there are any model which can be revisited as a context to build the model? Better to elaborate the context before going for results and analysis. 

 

Author Response

Dear Reviewer, we have taken your valuable comments into account to improve our manuscript. We have highlighted all the changes requested by the team of three reviewers in green.

Apart from these points, few key points to consider are as follows;

C1. Line 29 and Line 34 are too broad to make the conclusions. They do not reflect the results correctly. It can be rephrased to line with the objectives.
R1. We have improved the presentation of results in the abstract, highlighting the changes.

C2. Future research should be in continuation to the results set in the study. They seem too diverge to draw the lines of future study. The scope of research is more on methodology rather than the results. Consider rephrasing the abstract.
R2. We have focused future research on both the conclusion and the abstract.

C3. The context of financial behaviour and its meaning can be elaborated to provide a context. The lines 43 to 62, can be seemingly examined from the point of view of various demographic and socio-economic perspectives. The sentences are more causal in nature and not seem to be descriptive.
R3. We have improved the initial paragraphs of the introduction, clearly defining the construct of financial behavior, naming related variables and the contextual dimensions that affect it (Lines 43 to 79).
 
C4. The study by patrisia et.al (2023) can be revisited to examine the results. The lines 65 to 67 use the word religious financial education which can be looked into.
R4. We are very grateful for this comment. We have reviewed the article you mention in greater detail, which has helped us improve the introduction between lines 80 and 101.

C5. Kindly refer to the paper based on which the lines 78 to 80 are opinioned.
R5. We have reconsidered these lines, being more explicit in the cases of Brazil and Colombia. (Dos Santos & Barros, 2011; Abril-Teatin et al., 2022).

C6. The lines 88 to 94 can be elaborated to include the key results of the countries, relationship between socio-demographic variables and the financial behaviour. Does the context of financial behaviour vary from country to country is the primary questions which is lacking in all the paragraphs. Better to introduce this theme in the beginning.
R6. We have provided detailed national examples from Brazil, Colombia, Indonesia, Japan, Malaysia, Poland, the USA, and Vietnam (lines 72 to 138). In addition, we have left the big question suggested in the first paragraph of this introduction, preceding the various national examples.

C7. The meaning across the various paragraphs fail to elaborate on the central theme of the study in the beginning.
R7. In the first paragraph of the introduction, we highlighted the question of variation between countries, and in the last paragraph of the introduction, we connected the objective and the working hypotheses.

The context in the materials and method seem to be more apt in building the context in the introductory chapter rather than the second chapter. It is just a suggestion. 

C8. In the data analysis and processing procedures:
- The null hypothesis is H01. Then alternative can be H11 instead of H1. 
- The null hypothesis for second is H02. Then alternative can be H12.
R8. We have improved the numbering of the hypotheses. 

C9. The second hypothesis revolves around the correlations. When by default, the presence of identity matrix is defined by the bartlett test, how relevant would it be to include a broader hypothesis on the correlations.
R9. We have modified the hypotheses regarding expanded correlations, achieving more robust hypotheses.

C10. The results of the EFA can be elaborated to signify the results the factors identified and the number of factors and naming of the factors to provide a context next for the CFA. Is there are any model which can be revisited as a context to build the model? Better to elaborate the context before going for results and analysis.
R10. We have compared our results (with three factors) with the CFA for a single factor with 13 items, reporting the values in Table 8. This demonstrates that the proposed model has better results than the original contrast model (with a single factor). Additionally, we have provided further information regarding the previous study by Méndez-Prado et al., 2022 (original model), see Table 1 (factor loadings).

Reviewer 3 Report

Comments and Suggestions for Authors

The aim of the study was to analyze the factor structure and socio-demographic correlations of the FB-13 scale among undergraduate students (in Economics, Business, Administration, and Accounting) in Tegucigalpa, Honduras. A total of 714 undergraduate stu-129 dents from the National Autonomous University of Honduras (UNAH) participated (37% 130 men and 63% women), aged between 16 and 45, residing in rural (19%) and urban (81%) 131 areas. The research aim  has been accompanied by two types of hypotheses. Statistical analyses allowed to test the proposed hypotheses. In the study, statistically significant associations were  identified between the FB–13 and certain sociodemographic variables, specifically parental status and work experience. Conversely, no statistically significant relationships were  observed between the FB–13 and the variables gender, age, residential area, marital status, employment status, or undergraduate level.

I think, the study is interesting, and the paper is generally well prepared, with relevant data. However, I have following observation and remark.

The authors state that the research sample consisted of 37% men and 63% women. In  my opinion, this selection is highly surprising.  It may affect the obtained results. Therefore, the authors should explain why the proportion of women is so high. The remaining characteristics are not as controversial, as one can easily explain why the sample consisted of 85% singles or 81% urban residents. Nevertheless, this should be justified in the sample selection too.

This needs to be corrected in the paper. I have no other comments, and once the aforementioned issues are addressed, the paper can be published.

Author Response

Dear Reviewer, we have taken your valuable comments into account to improve our manuscript. We have highlighted all the changes requested by the team of three reviewers in green.

C1) The authors state that the research sample consisted of 37% men and 63% women. In  my opinion, this selection is highly surprising.  It may affect the obtained results. Therefore, the authors should explain why the proportion of women is so high. The remaining characteristics are not as controversial, as one can easily explain why the sample consisted of 85% singles or 81% urban residents. Nevertheless, this should be justified in the sample selection too.
R1) The population of undergraduate students in Economics, Administration, and Accounting Sciences surveyed at the Faculty of Sciences in the University City of Tegucigalpa is 54.89% female (10,049 students) and 45.11% male (8,260 students). Regarding a possible difference in distribution by gender, the Mann-Whitney U test did not show any differences by gender for any of the 13 items in FB-13. Additionally, 85% of respondents were single because the study focused on undergraduate students (66% between 16 and 24 years old. See table 3), and 81% were urban residents because the study was conducted in the country's capital, one of the three most developed areas (see section 2.2 Context).

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

1. The methods are described in detail and this is good. However, it is not explained why the unweighted least squares was chosen for EFA?
2. In Figure 1 (Scree plot), please sign the axes, visually mark the "elbow" point and the clipping line.
3. Figure 2 is not clear enough. It is recommended to increase the resolution and font on the axes, add panel designations (a-d) and axial captions so that the drawing is visible and readable without magnification.
4. Figure 3 is not clear enough too. It is recommended to increase the font size for better readability.

Author Response

Dear Reviewer, we have taken your valuable comments into account to improve our manuscript. We have highlighted all the changes requested in green.

C1. The methods are described in detail and this is good. However, it is not explained why the unweighted least squares was chosen for EFA?
R1. The choice of the unweighted least squares (ULS) method for exploratory factor analysis is based on methodological and empirical criteria widely supported by the specialized literature. Unlike methods such as maximum likelihood (ML), ULS does not require strict assumptions of multivariate normality or continuity of the observed variables, making it particularly suitable for ordinal data derived from Likert-type scales. Simulation studies have shown that ULS offers more stable and accurate estimates in contexts with asymmetric distributions, moderate sample sizes, and complex factorial structures (Krijnen, 1996; Ximénez & García, 2005; Li, 2016; Morata-Ramírez et al., 2015). Furthermore, its computational implementation ensures efficient convergence. Taken together, this evidence justifies the use of ULS as a robust and methodologically relevant alternative for the present study. (See lines 245-253, add four additional references).

C2. In Figure 1 (Scree plot), please sign the axes, visually mark the "elbow" point and the clipping line.
R2. We have improved Figure 1 (Scree plot) with 300 dpi, enlarged the axes, marked the “ elbow” point and the clipping line (eigenvalue = 1). See lines 345-346 and Figure 1.

C3. Figure 2 is not clear enough. It is recommended to increase the resolution and font on the axes, add panel designations (a-d) and axial captions so that the drawing is visible and readable without magnification.
R3. We have improved Figure 2 with 300 dpi. We increased the resolution and font size on the axes, and added details to the callouts. See lines 402-403 and the figure panel 2.

C4. Figure 3 is not clear enough too. It is recommended to increase the font size for better readability.
We have improved Figure 3 with 300 dpi. We increased the resolution and size of the axis and legend fonts. See the figure panel 3.

Reviewer 2 Report

Comments and Suggestions for Authors

Thank you on making the appropriate corrections. There is no correction sought anymore. 

Author Response

C1. We are very grateful for your response and your agreement to our corrections to the first round of revisions.

R1. We are very grateful for your response and your agreement to our corrections to the first round of revisions.