I want to measure whether three predictors predict a fourth variable. The data was collected on a daily basis for 7 days. The problem is correlation was high among the daily measures, and thus multicollinearity must be considered. Is it ok to still run an HLM test? or can i use a path analysis? or are there any other solutions?

Noted that the three predictors can't be combined into one variable.