I am doing a hierarchical multiple regression with 70 observations. Is it okey in block one to add item 1 item 2 and item 3 separately from a survey, or does it count as three predictors? In block two I have 2 total scores of 2 surveys. The reason there is not total score on the first survey is because item 1 and 2 are yes/no, while the third is (once a week, twice a week, once a month etc). Because if it counts as three predictors, I'll have five predictors in total and then it would be too many for number of observations.

I also have problem where the collinearity tolerance is under 1.

Plus there is strong negative correlations found.