A few methodological points jumped out at me that might be worth thinking about. Nothing fatal, but stuff that makes me pause.
First off, that wild gender skew (329 women vs. 83 men) is a big one. We know there are gender differences in personality expression and coping mechanisms. Do you think the findings, especially the prevalence and characteristics of the Undercontroller profile, would hold up in a more balanced sample? It kinda makes me wonder if the profiles are capturing something specific to this predominantly female undergrad experience.
Also, using hypothetical vignettes for measuring coping is always tricky. People are notoriously bad at predicting their own behavior in stressful situations. Any thoughts on how social desirability might have played a role here? I’m imagining participants reading the scenario and thinking, “Well, of course I’d use problem-solving,” even if in the moment they’d just panic and eat a whole pizza.
I was also a bit stuck on the LPA model selection. The VLMR test wasn’t significant for the 3-class model (p = 0.205), which traditionally suggests the 2-class model is better, but you all went with three based on theory and other fit indices. Was this a tough call? I’m totally on board with not slavishly following a single statistic, but it does make me curious about how those two profiles would have looked.
Finally, not controlling for the perceived stressfulness of the vignettes in the coping ANOVAs seems like a potential confound. If Resilients found the scenarios less stressful to begin with, it’s easier for them to choose adaptive strategies. Was there a reason you decided against using that stress rating as a covariate?