With the introduction of the new crosstabs showing the Daily Kos/SEIU/PPP weekly poll split into Red States, Blue States, and Swing States, I thought it would be informative to look back at the spring and summer numbers to see how these numbers behaved during a period of relative stability in the Presidential race.
Because these subsamples have a smaller number of respondents, we expect more bounciness than the toplines. Note that only changes in the vote share of about 8 points or more are statistically significant. Here's what we see:
In two cases, however, there was significant movement in the numbers coinciding with an Obama announcement.
On May 9th, Obama gave his historic interview supporting marriage equality. In the poll completed May 10-13, Obama's numbers jumped by 8 points in the Blue States while Romney fell by 7; the next week, the numbers went right back, with Obama falling 9 in the Blue States and Romney gaining 9. Obama also fell and Romney gained in Swing States, but these changes were not large enough to be statistically significant.
Obama announced his new immigration policy on June 15. In the poll completed June 14-17, Obama's numbers fell by 11 points in the Red States while Romney gained 14; the next week, Obama gained back 10 and Romney fell 10. (In the Blue States during the same time period, Obama initially gained 7 and Romney fell 5; subsequently Obama fell 7 and Romney gained 6.)
Both instances show a significant, but temporary, change in vote preferences in Blue or Red States, but nothing significant in the Swing States. Perhaps non-competitve states have a larger share of the voting public with weakly held preferences that are easily swayed by current news topics, while campaign activities create more hardened voter preferences in swing states?