There are some two-way interaction of factors (two) with three levels each and 10 continuous variables. What shall I do if few two-way interactions are not significant at one levels and significant in another level? Do I need to ignore those? Any feedback would be appreciated.