1. 澄清不一致之處及詞彙,說明讓審查者了解
2. 博士班學生的層級:what à how à why。要說明為何要如此寫的理由。
3. 了解item mean為何要相等之理由。
原先之寫作
Reviewer's comments: In terms of scaling assumptions, is it appropriate for the mean of subscale to be equivalent? If the scale is administered to certain diagnostic group which the subjects may present better score on certain aspect of QOL, will you still expect to see equivalent mean for each scale? How this can be applied to the results of your study?
Reply:
Under traditional summated rating (Likert) scaling criteria, item means should be equivalent within a scale. The item means compare within its own scale, not with other scales. The SF-36 has 8 scales, such as the PF, RP, BP, GH, VT, SF, RE, and MH scales. Each scale has its own items. For example, the RP scale has 4 items, and these 4 items should have equivalent means.
Yes, it is appropriate for the equivalent of item means, because the design of the SF-36 was under traditional summated rating (Likert) scaling criteria.
If the scale was administered to a certain diagnostic group, the subjects may present better score on certain aspects of QOL. We would expect to see the equivalent of item means within its scale, not the equivalent means between the scales.
In our study, we found that the item means within its scale were approximately equivalent for the PF, RP, GH, RE scales in heroin abusers. The item means of the BP and SF scales were different within their scales.
修改後之寫作
Reply:
In our study, we use the terms ” scale and item”, not “scale and subscale”. The SF-36 has 8 scales and each scale has its own items.
Under summated rating (Likert) scaling criteria, item means and standard deviations should be roughly equivalent within its scale. The equivalent of item means and standard deviations indicates that the distributions of the items are the same. The same item distributions mean that these items reflect the same construct. Because reflecting the same construct, it is meaningful for summating the items. For example, the 4 items of the RP scale has similar item means and standard deviations, so it is meaningful for summating the scores of the 4 items to reflect the RP construct.
Yes, it is appropriate to examine the equivalent of item means and standard deviations, because the design of the SF-36 was under summated rating (Likert) scaling criteria.
If the scale was administered to a certain diagnostic group, the subjects may present better score on certain aspects of QOL. We would expect to see the equivalent of item means and standard deviations within its scale.
Based on assumption of the equivalent of item means and standard deviations, we find the roughly equivalent of item means and standard deviations within its scale, such as the PF, RP, GH, RE scales. Therefore, we may state that the items reflect the same construct for its scale. We have revised the content: “The item means and standard deviations were roughly equivalent with its scale, except those of the BP and SF scales. In practice, a possible reason for item means and standard deviations not being equivalent is heterogeneity of item content. ”
沒有留言:
張貼留言