Includes five additional polls released between 10/25 and 10/28.
146 Likely Voter Polls from Pollster (7/1-10/28) Dependent variable: Obama Lead over Romney Ordinary Least Squares coefficient std. error t-ratio p-value ------------------------------------------------------------ const 5.72235 0.579939 9.867 1.54e-17 *** Debate1 −5.47401 1.05759 −5.176 8.24e-07 *** Trends Pre-Debate −0.0514766 0.00937824 −5.489 2.00e-07 *** Post-Debate −0.0124666 0.0424105 −0.2940 0.7693 House Effects Rasmussen −2.35773 0.437616 −5.388 3.18e-07 *** Gallup −4.34481 1.18503 −3.666 0.0004 *** YouGov/Econ 1.80215 1.02690 1.755 0.0816 * ARG −1.93337 0.844044 −2.291 0.0236 ** DemCorps 2.15593 1.03252 2.088 0.0387 ** UTech/NJ −3.81771 2.00889 −1.900 0.0596 * NPR 2.74136 1.44450 1.898 0.0599 * Allstate/NJ 3.74853 2.00311 1.871 0.0635 * Reason/Rupe 3.85149 2.00263 1.923 0.0566 * JZAnalytics 2.65474 0.790120 3.360 0.0010 *** Mean dependent var 1.075342 S.D. dependent var 3.055996 Sum squared resid 519.2214 S.E. of regression 1.983304 R-squared 0.616576 Adjusted R-squared 0.578815 F(13, 132) 16.32820 P-value(F) 1.13e-21
For models like this with Rasmussen included, I created separate pre- and post-debate trends and tested whether each significantly differs from zero. This model shows that the trend before the Denver debate was significant favorable to the President, but since the debate the race has stagnated.