check the articles on "formative" measures

for example, think of organizational constraints which consists of various components such as supervisory behaviors, coworkers, company polices, and etc. and they are not highly correlated

Hi, i'm not sure if this will help your case, but a while ago, I read a student's thesis that might be similar with yours. She is basing her case on work-life balance PoV from Rantanen et al.

Anyway, if you decided to make those three as separate variables, then generally speaking, you should first use mean of each three to see the conflict/facilitation dynamic between each facets (WL, LS, SW). If after that you want to simply see the level of C/F, then yes, you can just average those 3 to see the level of C/F.

And regarding forming the variable. I think you can. From my understanding, your WSL variable would have two major dimensions (C/F) and three factors for each dimension (WS, SL, LW). I don't think items from a factor should correlate strongly with other factors though. E.g items for WS should be exclusive to WS alone, and SL for SL alone. (Hope you understand my description). If an item correlates with two or more factors, that item would be unreliable as it has two or more interpretations. E.g it can say A about a factor and B about the other factor.

Anyhow, i hope this helps. I do not sufficiently understand WLC theory in general, so i'm sorry i cannot give you a more elaborate and practical example. Goodluck for the thesis!!

Ps: i mentioned Rantanen et al bc i think it might be relevant with your study. Sorry i forgot to elaborate that.

Tae-Yeol Kim, Bonita Lee-McFadden

Thank you both for your answers! I think I know what I need to do now!