Validation of Korean Version of the Hardiness Resilience Gauge
Round 1
Reviewer 1 Report
Participants
Add the number of total respondents and the number of incomplete and excluded respondents (response rate).
Add range for age of participants.
Results
CFA fit indices aren’t good also for the modified model. Correlations between factors are very high. Have you test a model with only one factor? Fit indices of three-factor model are better than one-factor model? Add the analysis and discuss the comparison between the two models.
Discuss the three parameters of error covariance allowed to vary in modified model. What do these items share?
Several standardized regression weights (SRWs) are low (<.50), especially in Control factor. Add a discussion about it.
What are the values in column CR in Table 3? Are they Construct Reliability (CR) values? Why are they above 1? Please report correct CR values and also Average Variance Extracted (AVE) values and discuss the results.
Table 4 is unclear, it’s better divide the table in two separate tables: one for descriptive statistics of HRG and one for correlations
Discussion
Change the term “satisfactory” with “acceptable” at line 228.
Remove the sentence "This result indicated that Korean HRG would be more useful in measuring the robustness of individuals with three factors and relatively few elements" because results don't indicate it. In fact, total score shows higher correlation than the three factors.
Author Response
Thank you very much for your comments to improve the quality of this articles.
The revised parts were marked in red, and we included the page and line of the revised part.
Response to Reviewer 1 Comments
Point 1. Add the number of total respondents and the number of incomplete and excluded respondents (response rate).
Response 1: Unfortunately, there was an oversight because we wrote what we had habitually described about the participants. There were no incomplete questionnaires because online the questionnaire (the response) was submitted only when all questions were answered. But one respondent reported her age as 15, and so was excluded from the analysis. According to your advice, the following explanation is included. (Line 76-78)
Participants were recruited from the SNS where adults participate and from the university student Internet community. One respondent who reported an age of 15 was excluded from the analysis.
Point 2. Add range for age of participants.
Response 2: Thank you for your comment. The age range of respondents is described as below. (Line 80-81)
The participants’ ages ranged from 18 to 73, and with a mean of 42.31 ± 12.73 years.
Point 3. CFA fit indices aren’t good also for the modified model. Correlations between factors are very high. Have you test a model with only one factor? Fit indices of three-factor model are better than one-factor model? Add the analysis and discuss the comparison between the two models.
Response 3: Thank you very much for your good comment on this part. The purpose of this study was to verify the model of the three factors. And it was impossible that researchers could choose the number of factors because it was validated and standardized scales with copyright in Multi-Health Systems. Even so, we tried a confirmatory factor analysis with one factor, and this model showed poor model fit. We added these results to Table 2 and this section as follows: (Line 162-164)
We first tested a one-factor model, which showed poor model fit (TLI=0.794, CFI=0.810, SRMR=0.064, and RMSEA=0.076). Next, we assessed the hypothesized three factor model.
In the CFA, especially the model with three error covariances showed very high correlation coefficients between subscales. In the correlation analysis, the correlation coefficient between the three subscales was not very high. So, we presented the results of the correlation analysis. These results were described and explained as follows. (Line 206-210)
The correlational analysis revealed that the correlation between challenge and commitment subscales was relatively high (r = 0.73, p < 0.01). The control subscale was also strongly correlated with the challenge (r = 0.61, p < 0.01) and the commitment subscales (r = 0.60, p < 0.01). Even so, the confirmatory factor analysis model with one factor showed poor model fit.
Point 4. Discuss the three parameters of error covariance allowed to vary in modified model. What do these items share?
Response 4: Thank you for your valuable comment. As you pointed out, we added some discussion as below: (Line 246-251)
Because item 4, 10, and 12 belong to the control subscale and are based on the belief that one's life depends on one's own actions, it is presumed that item 4 and 10 share substantial variance with item 12. However, it is necessary to confirm in future studies that there may be much variance shared between item 5, related to preference for changes in daily routine in the challenge subscale and item 6, related to enjoying daily activities in the commitment subscale
And the following contents were included as limitations of the study. (Line 307-309)
Finally, the covariance levels of items 5 and 6, items 4 and 12, and items 10 and 12 were high. Because there was no theoretical basis for modified model allowing three error covariances, it is necessary to confirm this phenomenon in further studies.
Point 5. Several standardized regression weights (SRWs) are low (<.50), especially in Control factor. Add a discussion about it.
Response 5: Thank you very much for your comment. If it's developing a scale, I can exclude some items with low SRW, but we’d appreciate it if you could understand that we can't exclude items of HRG that are standardized and copyrighted by MHS. It would have been nice if the SRW of all items was .50 or higher, but it was not, so I described it as a limitation of the study and suggested some further study with other larger Korean samples. (Line 304-307)
Third, all subscales included an item with relatively low SRW, leading also to somewhat low AVEs especially for the control subscale. And, subscales were rather strongly intercorrelated. Thus, further studies should assess the HRG Korean with other, larger Korean samples.
Point 6. What are the values in column CR in Table 3? Are they Construct Reliability (CR) values? Why are they above 1? Please report correct CR values and also Average Variance Extracted (AVE) values and discuss the results.
Response 6: To my shame, my research assistant input Critical Ratio. Therefore, we calculated CR and AVE and included them in Table 3. And we described it in the method and result part, and I add it in the discussion part as below. (Line 157-159) (Line 193-196) (Line 304-307)
And, we also calculated the composite reliability (CR) and average variance extracted (AVE) to identify convergent validity. A CR larger than 0.70 and an AVE larger than 0.50 suggest a good convergent validity [28].
And, CRs of all subscales were satisfactory because they were above 0.70 (0.85 for challenge, 0.71 for control, and 0.86 for commitment). However, AVE was low, so all subscales had AVEs less than 0.50, because it became quite low if an item with SRW lower than 0.50 was included.
Third, all subscales included an item with relatively low SRW, leading also to somewhat low AVEs especially for the control subscale. And, subscales were rather strongly intercorrelated. Thus, further studies should assess the HRG Korean with other, larger Korean samples.
Point 7. Table 4 is unclear, it’s better divide the table in two separate tables: one for descriptive statistics of HRG and one for correlations.
Response 7: Thank you for your comment. We input descriptive statistics of HRG Korean in a new table of correlation matrix for subscales of HRG Korean (Table 4). Therefore, there will be no confusion in understanding the table. (Line 211-213)
Point 8. Change the term “satisfactory” with “acceptable” at line 228.
Response 8: As you advised, we changed the term “satisfactory” to “acceptable” as below. (Line 243)
Even so, RMSEA, the absolute goodness-of-fit indices, of HRG Korean was acceptable.
Point 9. Remove the sentence "This result indicated that Korean HRG would be more useful in measuring the robustness of individuals with three factors and relatively few elements" because results don't indicate it. In fact, total score shows higher correlation than the three factors.
Response 4: Thank you for your comment. Wouldn't it be okay to leave that sentence of the discussion as it is if we mentioned in response to your point 3 are included as follows? It's an important part of this study. It will be highly appreciate if you understand that.
In the CFA model, especially in a model with three error covariance showed very high correlation coefficients between subscales. In the correlation analysis, the correlation coefficient between the three subscales was not very high. So, we presented the results of the correlation analysis. The results were described and explained as follows. (Line 206-210)
The correlational analysis revealed that the correlation between challenge and commitment subscales was relatively high (r = 0.73, p < 0.01). The control subscale was also strongly correlated with the challenge (r = 0.61, p < 0.01) and the commitment subscales (r = 0.60, p < 0.01). Even so, the confirmatory factor analysis model with one factor showed poor model fit.
Author Response File: 
Author Response.docx
Reviewer 2 Report
Dear authors,
Your paper - Validation of Korean version of the Hardiness Resilience Gauge presents a great contribution to validate the Korean version of the Hardiness Resilience Gauge (HRG Korean). You suggest that the HRG is a valid instrument for measuring Koreans’ hardiness and resilience to promote their well-being and sustainability. I only have some suggestions regarding the introduction and conclusion part.
- I think it would be a great improvement to add a literary review part section and also improve the references list;
- The conclusion is poorly written and I will suggest improvements to that part.
Thank you and good luck.
Kind regards
Author Response
Thank you very much for your comments to improve the quality of this articles.
The revised parts were marked in red, and we included the page and line of the revised part.
Response to Reviewer 2 Comments
Point 1. I think it would be a great improvement to add a literary review part section and also improve the references list.
Response 1: Thank you very much for your comment. As shown below, sentences that make it possible to better understand the reviews in the introduction are included. And we included a few more references. It will be highly appreciate if you understand that we can’t write the introduction in a lengthy manner because the introduction is concisely described these days compared to the past and it is a scale validation study. (Line 31-43)
Multiple studies confirm that psychological hardiness and resilience are personal traits that buffer the negative effects of various stressors, maintain individuals’ quality of life, and further improve their well-being and sustainability [1−4]. In 1979, Kobasa introduced the concept of hardiness as a personality factor that plays an important role in individuals’ psychological resilience, and which can help them recover from the negative consequences of stress [5]. It has long been empirically shown that the hardy personality type is resistant to various kinds of stressors [6−8]. For example, hardiness has a health psychological effect that protects individuals from cardiovascular disease, the most common disease caused by stress [9−10]. Thus, hardiness is a psychological variable that many researchers are interested in not only psychology but also nursing and medical fields.
Hardiness was originally conceived as a personality trait or disposition that develops early in one’s life and is relatively stable over time and across situations, although it may change under certain conditions [11−12]. Therefore, Bartone refers to hardiness as dispositional resilience [13]. In Korea, there are more studies on resilience than hardiness [14−15]; however because hardiness is a devotional trait that makes it possible for individuals to recover from severe stress or traumatic experiences, measuring hardiness allows us to predict their resilient abilities more accurately during stressful events. In other words, it is possible to predict whether an individual can resiliently recover from the negative effects of stresses by measuring hardiness.
12 Bartone, P.T. Resilience under military operational stress: Can leaders influence hardiness? Mil. Psychol. 2006, 18, S131-S148. doi.org/10.1207/s15327876mp1803s_10
17. Ouellette, S C. Inquiries into hardiness. In L. Goldberger, & S. Breznitz (eds.), Handbook of Stress: Theoretical and Clinical Aspects (2nd ed., pp. 77–100). Free Press, New York, NY, USA, 1993.
28. Anderson, J.C.; Gerbing, D.W. Structural equation modeling in practice: A review and recommended two-step approach. Psychol. Bul. 1988, 103, 411–423.doi.org/10.1037/0033-2909.103.3.411
Point 2. The conclusion is poorly written and I will suggest improvements to that part.
Response 2: As you advised, we have revised and improved the conclusion, while striving to keep it concise, as follows: (Line 312-321)
This study has demonstrated that the HRG Korean can be a useful tool for measuring hardiness in Koreans. Although the factorial structure of the HRG Korean was less satisfactory than expected, the three factor model (with three error covariances) showed a good factorial structure and an acceptable model fit. Reliability of the HRG Korean estimated with the internal consistency was also high. Despite these advantages, additional studies with the HRG Korean should be conducted with larger samples of Koreans.
The HRG Korean has shown excellent concurrent and predictive validity in this study, lending credence to its utility as a comprehensive measure of the hardiness construct. As such, it should prove a useful tool that can be applied by academicians and clinicians to measure hardiness and resilience in the Korean context.
Author Response File: Author Response.docx
Round 2
Reviewer 2 Report
Dear authors,
Thank you for your improvements.
Kind regards
