Any time you re-specify or modify your model, you are implicitly changing its meaning in some fundamental way. In many instances, a change in model specification results in a trivial or unimportant corresponding alteration of the model’s substantive meaning, but in other cases model modification can foreshadow a strong shift in the model’s meaning from a theoretical standpoint. Therefore, it is crucially important to think through each proposed model modification and ask yourself if making the modification is theoretically consistent with your research goals.
A second consideration to take into account when you modify a model is that you are relying on the empirical data rather than theory to help you specify the model. The more empirically-based modifications you incorporate into your final model, the less likely the model is to replicate in new samples of data. For these reasons, you should modify your models based upon theory as well as the empirical results provided by the modification indices.
As a practical consideration, it is also worth noting that AMOS provides modification index output only when complete data are input into the program. In other words, you cannot obtain modification index information when you use missing data with AMOS.
扫码加好友，拉您进群
全部版块
我的主页
收藏
