Why do different versions of Amos give different values for CFI?

Prior to version 4.02, when a model included means and intercepts as explicit model parameters, Amos used a different baseline model than most other SEM programs used in computing fit measures like NFI, NNFI and CFI. Amos's baseline model required each observed variable to have a mean of zero. By contrast, most other SEM programs allowed the means to be unconstrained in the baseline model. Because Amos's baseline model typically fit extremely badly, fit measures like NFI, NNFI and CFI took on larger values in Amos than in most SEM programs. In other words, Amos's baseline model was so bad it made all your models look good by comparison.

Amos's old baseline model (used prior to version 4.02) was not wrong. In fact, Amos 5 and later still allow that old baseline model as an option when you perform specification searches. However, the difference between Amos's baseline model and the one used by most other SEM programs was causing confusion, and so the decision was made in 4.02 to allow means to be unconstrained in Amos's standard baseline model. So in version 4.02, Amos fell into line with other SEM programs. The result is that reported CFI (as well as NFI and NNFI) became smaller in version 4.02 for models that include means and intercepts. (Models that do not include means and intercepts were not affected by the change in version 4.02.)

Beginning with Amos 5, you have a choice of four baseline models when doing a specification search. Means can be either fixed at zero or unconstrained, and correlations can be either fixed at zero or constrained to be equal.