Comparing main effect model and interaction model.

[Sung]

Hi,

Variable "age" was significant in multiple hierarchical regression for main effect model. So, I thought age was important for posttraumatic stress disorder (PTSD).

However, age was significant in multiple hierarchical regression for interaction effect model.
I mean
  1st step: traumatic experience --> significant
  2nd step: age --> significant
  3rd step: traumatic experience * age --> significant

In this case, how can I interpret this result of age? Does the existence of interaction effect suggest that the data are more supportive of the interaction model?

In contrast, if age was not significant in main effect model, but significant in 2nd step at interaction effect model, how can I interpret this discrepant results?

Thanks,