Let's say I run a cross-cultural study with message framing as the IV, self-care behavioral intentions as the DV. I found support for metric invariance w behavioral intentions but no support for scalar invariance. While the difference in behavioral intentions between regions cannot be meaningfully interpreted due to lack of scalar invariance, is it still acceptable to run a regression with Region*Framing as an interaction term for self-care behavioral intentions as the DV? I am unfamiliar w measurement invariance so this may be a basic question. Thank you!