To test the moderating hypotheses (Hypothesis3–6), we added the interaction terms in Models M1* and M2* (see Tables 2 and 3). To deal with multicollinearity between the interaction terms, we mean centered each scale that constitutes an interaction term and created the interaction terms by multiplying the relevant mean-centered scales.


能看下这句话吗？ 为了解决交互项之间的多重共线性？这里 交互项不就和多重共线性有关联了吗？  另外，谁能翻一下这句话，不太懂这句话什么意思

啥跟啥啊

在计量模型中，引入交互项会产生很强的多重共线性。要处理这种多重共线性，需要中心化变化。即7楼提到的：
To test the moderating hypotheses (Hypothesis3–6), we added the interaction terms in Models M1* and M2* (see Tables 2 and 3). To deal with multicollinearity between the interaction terms, we mean centered each scale that constitutes an interaction term and created the interaction terms by multiplying the relevant mean-centered scales.
为了测试放松假设（假设3-6），我们将交互项引入模型m1和m2（看表2和表3）。为了处理交互项的多重共线性问题，我们中心化均值化量纲，这样就形成了交互项，通过乘以相对中心均值化量纲形成了新的交互项。

[quote]zhoufan010 发表于 2014-7-15 17:59 https://bbs.pinggu.org/thread-4714507-1-1.html