I have a non-significant 3-way interaction (A X B X C mixed design, 2 binary (A/B: 0,1) and 1 continuous:C), alongside with two significant 2-way interaction(A X C and B X C), and my questions are,
should I interpret the results from the 3-way interaction output, or should I run new analyses for the two significant 2-way interactions separately and interpret these instead?
my instinct about the first question is to interpret the original 3-way output because it estimates the two 2-way together, even if the 3-way is not significant, if this is the case, e.g., when decomposing A X C by the /EMMEANS command, should i specify the estimation at which level of B or simply ignore it (see the pic below)? If variable B were continuous, could I ask for estimating at mean level of B (the last line in the pic)?