Computers in Human Behavior (CHB) 저널로부터 의뢰를 받아서 작성한 저의 reviewer’s comment입니다.
타당하지 않은 연구방법론과 타당하지 않은 literature review, 참신성의 부족 등의 이유로 completely reject 판정을 내렸던 manuscript입니다.
——
While this is an interesting topic and well-written for the most part, but the manuscript has a number of drawbacks. Most importantly, the six items (factors) of Internet services used here were not well defined, with problematic factor analysis. Furthermore, overall contribution of this study to the existing literature seems limited or unclear. Lastly, the theory (particularly in the part of Literature Review) needs more in-depth explanations and needs to be more tightly linked to the selection of this study variables & hypotheses to be examined. I therefore recommend that it should be completely rejected. The author(s) should consider in-depth reviewing on relevant literature, elaborating on the design of survey, and data analysis to substantiate the author(s)’s claims for future submission.
Minor issues:
1. A battery (usually in HCI studies) means a metric or test kit for assessing operator performance. For readers, the author(s) needs to specify what battery exactly stands for in this study (for example, ‘battery of revised Internet services’ in the section of Introduction).
2. Grammar and referencing needs to be checked and corrected throughout (i.e., “The Big Five traits… psychologists believe provide a full…” (pp.5).
3. “These traits are empirically supported.” (pp.5): the author(s) needs to articulate this statement, if the author(s) intended to mention that the utility and applicability of the Big Five have been validated in a wide variety of settings (although the examples are appeared in the following).
4. Whilst I do not have an issue with the literature reviewing on the Big Five in the contexts from the career choice of students, leadership, the technology acceptance model, etc. (pp.5-6), this whole line of review actually has nothing to do with your main research topic.
5. I partially agree with the author’s opinion that people who are neurotic are more likely to engage in the Internet activities more often than less neurotic people. You may want to check more references of Amichai-Hamburger, 2002; Amichai-Hamburger et al., 2002; Amichai-Hamburger & Ben-Artzi, 2003; Ehrenberg et al., 2008; McElroy et al., 2007; Tuten & Bosnjak, 2001.
6. The literature review on the Big Five on the Internet Service (pp.6-7) does not seem to be sufficient to relate your research.
7. “This paper generally adopted the practice by [2] to develop an update list of Internet services” (pp.7): I do not clearly understand how the author(s) conducted an updating list of current Internet services. More importantly, how did the author(s) recruit the twenty Internet ‘frequent users’? What’s the rationale behind this? Are they all experts in your study domain, for example? Furthermore, I don’t think the Internet websites identified by 20 ‘frequent users’ are not really representative of frequently used Internet websites.
8. The author(s) mentioned that this study made 23 lists of Internet service. In Table 3, for example, there are 4 types of blogging service – write blogs, read blogs, write micro-blogs, and read micro-blogs. Do we have to separate the terms of blogs and micro-blogs? I cannot even recognize the difference between them.
9. The six-point scaling system is not a frequently used one in behavioral research. Related reference(s) needs to be provided.
10. The survey design includes four demographic factors of gender, marital status, age, and employment status to be controlled for in the multiple regressions, but the author(s) does not mention any logic or theory regarding why these factors should be considered for this study in the section of literature review.
11. I understand that the author utilized a non-random convenience sampling, but how did the author(s) recruit the respondents participated in this survey? Did the author(s) send an invitation to potential respondents by email or phone, for example? Were they all voluntarily participated in? what’s the net response rate?
12. To my knowledge, each dimension of the Big Five is totally exclusive to each other. For example, those with high level of neuroticism can possess a high level of agreeableness trait. But in the section of 3.2, the author(s) reported that the score on agreeableness is highest while the score on neuroticism is the lowest. Is this reporting necessarily? What was the intention to do?
13. The author(s) used a term of ‘rho” but it should be “p” if intended to represent the significance level.
14. The author conducted a series of ANOVA for identifying any differences in dependent measures. But, this study employed a convenient sampling with the nature of non-random sampling for collecting sample data. What was the rationale behind using ANOVA in this study?
15. In Table 3, the author named the sixth component extracted as a factor of ‘games & online friends’ but this is totally confounded. These two cannot be considered as a single factor. Also, there are several items (e-banking (.693, this may be not an issue and acceptable though), online radio (.536), instant messaging (.664), and finding new friends online (.655)) that have factor loadings lower than .7 which is considered as satisfactory cut-off loading (unless the items are newly developed), these were not eliminated for obtaining a better-fitted model. What was the rationale or logic for including these items for data analysis?
16. In pp.9, the Cronbach’s alpha below the value of 0.70 is not usually accepted as a satisfactory threshold. The author(s) needs to justify the rationale for it with proper referencing.
17. There are several statistical & methodological issues missing prior to conducting multiple hierarchical regressions in this study, including (1) common method bias and (2) outliers check (i.e., Mahalanobis distance and Mardia’s multivariate skewness and kurtosis tests)? If not, the author(s) never have to proceed (multiple) regressions. I am aware that the author(s) checked the multicollinearity, but the values ranged from MIN to MAX should be reported in the manuscript with the minimum cut-off used. Also, the tolerance is practically equal to 1/VIF, thus the representation of ‘the tolerance rates and VIF’ should be avoided (pp.9).
18. It may be useful if the measurement instrument (i.e., survey items) used in this study is provided in the section of Appendix.
19. “the influence of personality of Internet services may change over time as the services evolve” (pp.11): this may be true, but your study was not designed to address this issue.
21. Yes, your results of data indicates that “their difference in personality drives them to use Internet in different ways”, but you never have to speak of ‘why’ with your current survey design.
22. The part of Discussion is fairly well written (basis upon an assumption that your survey methodology has not a serious concern), but there are several illogical jumping too.
——
