Precision in the computation of polynomials

I’d had a debate about the importance of using single vs. double precision in calculations based on polynomials. I decided to do a little math on it. I doubt this is anything groundbreaking, but some people might find the math interesting.