A 42% favorable rating does not translate to or imply a 58% unfavorable rating. The two numbers won't add up to 100% due to "no opinion/don't know". For example, in May, 1992, when Clinton's favorable rating was 42%, his unfavorable was 48% -- not 58%.
Clinton trailed Bush in this race at this time by 30+ points as I recall.
My entire point is the polls this far out mean nothing and we all know how of late how completely inaccurate they have been. They were way off the mark in the 2010 midterms and much worse in the 2014 Midterms showing the Dems in much tighter races than was the outcome.
It is a combination of bias and the public’s avoiding these phone calls because of the inherent mistrust in the media in general.