I think Limbaugh’s strong coverage made the difference in the margin, if not the victory itself. He was on the race virtually solid from 1-2 PM. Local coverage and commentary notes that turn-out picked up unexpectedly, at least to Castle and regular Republican observers, in the late afternoon , so much so it spooked them and gave them reason to think turnout had gone beyond what they wanted and supposedly, that’s when castle began to feel “there was a problem, “ and he could lose. That is in line with thier having said all along that their “internal polling” looked good. Furthermore the more conservative House candidate, Urquardt, by PPP poll that first suggested O’Donnell would win, was running ahead of O’Donnell’s numbers in his race ( against a more establishment Republican woman, Rollins). However, Limbaugh, of course , did not mention him. So the voters who were drawn in by Limbaugh’s commentary, and being naive to Urquardt/ Rollins, may just have marked off the alphabetically favored Rollins ( candidates list in alphabetical order on ballot).
To finish the story, Urquardt did win, but narrowly, and finished behind O’Donnell in margin over his opponent, contrary to the PPP polling.