I wrote on July 19th that we would not know till we got some new preference data (particularly from Ohio and Indiana) whether John McCain has gained (conversely, Barack Obama has slipped) in the preference choice for presidency since last month i.e. June. Please read my posting, “Has there been a July slippage in Barack Obama’s support?”
Now we have lots and lots of new data since the 19th. So what’s the verdict? The overall state of the race is about where it was in mid-June. We have lots of data to support the postulate that the state remains about where it was, and some (but conflicted) data to show that may be Obama’s support has eroded a little bit. Here is the data –
A study of the state-by-state poll preference measures before the month of July and those from the month of July from Michigan (EPIC/MRA), Pennysylvania (Rasmussen) New Hampshire (University of New Hampshire, and Rasmussen), Colorado (Rasmussen), Virginia (PPP and Rasmussen), and Florida (Rasmussen) and the national polls (NBC/Wall Street Journal, Fox News) show the preferences to be about where they were in April-June — there is no systematic and obvious evidence of degradation of support for Obama.
However, the Quinnipiac preference measures in four major states — Colorado, Michigan, Wisconsin and Minnesota — and Rasumussen’s Ohio state preference measure do show degradation in Obama’s support in these states since June.
But then, Gall Up reported on July 23rd, “Since Barack Obama clinched the Democratic nomination and moved into a front-running position for the general presidential election in early June, he has seen his standing versus John McCain improve among voters in red states, blue states, and competitive (or purple) states. Obama has gained at least 3 points in the Obama-McCain gap in all three state groupings compared with voter sentiments in March through May.” And in Ohio, the PPP preference measures show stability.
Finally, the Pew Research Survey shows that Hispanics prefer Senator Obama (66 percent) to Senator McCain (23 percent) making many traditionally Republican western and south-western states (e.g., Colorado, Nevada) within plausible reach of Obama.
So what’s the verdict? The preferences are about where they were a month or two back — Obama is ahead marginally to modestly. May be Obama has lost 1-2 percent support at the margin from his peak measures. Any other inference would not have face and convergent validity when you account for model specification and sampling errors and differences.
In any case, it is premature to predict a victor in the November presidential elections.
Tags: Barack Obama, Convergent validity, Face validity, Fox News, Gall Up, Hispanics, John McCain, NBC, Pew Research, Polls, Public Policy Polling, Quinnipiac poll, Rasmussen Reports, The Wall Street Journal, U.S. Presidential Elections
