I have a dataset with 16 questions on a Likert scale (1-5) measuring the barriers to computer use. The dependent variable is the mean barrier index for which the 16 questions are the independent variables. They all have a byte datatype. There is one grouping variable with two categories. One of the independent variables, however, messes up the output in multiple linear regression. In the second group, the regression output is meaningless with all coefficients as 1, and all confidence intervals also as 1-1.
I'm not sure I completely understand the question, but if the DV is some sort of mean of the IVs then what you are doing is incorrect. Since you already know how mean_barrier_index is computed, you don't need a regression. And, if it is computed from the IVs, then you will get collinearity and other problems.